Virtually Indistinguishable
NOTE: This is a “soft publish.” It’s been a long time since I practiced my philosophical reasoning and I’ve had this just sitting for a long time. This is a challenge to myself—get back into philosophical reasoning, get back to thinking about this…until I do, everyone can read this draft!
Robert Nozick’s “experience machine” is an important thought experiment within philosophy, purporting to show we have an intuition against simple hedonism and desire satisfaction theories of well-being. Because we would not get into the machine, we must think there is something more to well-being than just being happy or having our desires satisfied (though technically an oversimplification, I will just say “hedonism” from here on). Otherwise, we would be okay getting into the machine even though we were not really doing the things it made us experience.
I suspect our intuition is actually against virtual things, or virtuality. I will not be arguing that your intuition is against virtuality; I will be arguing that we do not have a basis on which to assume that virtuality is lesser than reality, all else being equal. This will require adapting the experience machine to make sure our intuition was really about hedonism, not about virtuality.
Mach Bands
The image to the left illustrates an illusion known as “Mach Bands,” a visual phenomenon discovered by Ernst Mach (the same man for whom the speed of sound is named) Mach Bands are the areas in the image where the contrast between a dark band of color with a light band of color results in an area that appears darker in the dark band of color and an area that appears lighter in the light band of color. For me, it is easier to note that the image appears to have depth — the light bars seem to be casting a shadow on the darker ones
I use this illusion because it is an instance where a transition between two objects or concepts heightens the contrast between the two objects or concepts.
A similar example can be observed by taking three pots of water, one of which has ice in it and the other two which have room temperature water in them. Place one of the room temperature pots in front of you, with the ice water to the left of it and the other room temperature pot to the right of the first pot. Put your left hand in the left pot and your right hand in the right pot. There is an obvious difference on each hand.
Now, taking your right hand out of the room temperature water, place it in the other pot of room temperature water. You should notice no change. But, when you remove your left hand from the ice water and put it in the room temperature water, the water will feel warm, despite not having changed. The transition from one state (surrounded by ice water) to another state (surrounded by room temperature water) causes a heightened perception of the middle pot’s warmth. You have found a physical analogue to a Mach Band. From here on, I’ll use “Mach Band” to refer to any experience where a transition between two things heightens the contrast between them.
There is a Mach Band between being out-of-the-machine and being in-the-machine — it is the process of getting into the machine. If the machine produces a perferct simulation, then our non-virtual experience and virtual experience are differentiated by the transition of stepping from out of the machine (where our experiences, Nozick has told us, are non-virtual) into the machine (where our experiences, Nozick has told us, will be virtual).
What is Virtual?
Mach Bands, and in particular this Mach Band, are important because it not only demarcates virtuality and reality, but because it seems to heighten the contrast between the two.
The intuition the experience machine is supposed to get at is that hedonism alone is not enough to live well — we want to really have done the things that bring us pleasure. By stepping in the machine, we would only think we would have done the things that bring us pleasure, and that’s just not the same. If the pleasure were all that mattered, then we should excitedly get into the machine.
It is easy to think of virtuality and reality as binary values, especially polar ones. Nozick, accidentally, helps this assumption. He clearly differentiates that there is a virtual world in-the-machine and the real world out-of-the-machine.
The experience machine does not simply provide a view of the objects surrounding you; it is not a cave wall that requires a copy of each object to be represented or a puppet stage requiring miniatures. Since it does not require a cat in order to show you a cat, it can be surmised that the machine is either able to bypass your senses’ normal processes or is able to project representations in such a way as to make them appear as not-representations to your senses’ normal processes.
The former method would allow a machine to show you a cat merely by presenting information to your optic nerves as something besides photons, to your auditory nerves as something besides vibrations of a medium, etc. This is like what you see in the film series The Matrix, where the user does not use their eyes, ears, or other sensory organs when they are in the Matrix.
The latter method would allow a machine to show you a cat by presenting the information to your optic nerves as photons, to your auditory nerves as vibrations of a medium, etc, but to store the information in another format. This is like what you see in the Star Trek franchise’s “holodeck,” where stored information can be used to create mass so the user’s eyes, ears, or other sensory organs work “normally” when they are in the holodeck.
While it’s not important to perform a textual analysis to figure out what Nozick meant by virtuality, I will distinguish when an argument only applies to Matrix-style virtuality or Star Trek-style virtuality. What is common between these different styles, however, is that an object need not be present to cause the user to have an experience of that object.
We Experience Virtuality In Our Everydayness
In an auditory illusion called the Franssen Effect 2, an observer is positioned in a room with two speakers in front of her at different locations. Speaker 1 begins playing a particular sound that quickly fades in, then remains constant. After some time, the sound is faded over to Speaker 2. However, if asked, the observer will still locate the source of the sound as Speaker 1, even if the experimenter unplugs it while they are watching. This “mislocalization” occurs because of how we locate sources of sounds after we have started to (unconsciously) perceive their echoes. When we cannot locate any source it is assumed the source is at the last known position, even to the extent that we will hear what is not there. In cases where we cannot localize the sound, our auditory system provides a virtual source — that is, a perceived source need not be the source in order for us to experience it as such. [2]
In a tactile illusion called the Thermal Grill Illusion, alternating bars of a warm and cold material are placed in a grill-like pattern. To touch the warm material produces a warm sensation; to touch the cold material produces a cold sensation. To touch both materials simultaneously produces a sensation described as unpleasant or painful, or more colorfully characterized as just weird and akin to the burning sensation of sticking one’s hand in a snowbank. [3] Preliminary research indicates this may be an “attention-orienting” feature of our nervous system when faced with new stimuli. In such a case we can say that the weird sensation is virtual — that is, the thermal grill, or other object, need not be a “weird” object (the conjunct sensations of the thermal grill are quite normal) for us to experience it as such. [4]
More Virtuality We Already Experience
Examples of virtuality based on illusions are numerous, but I think the experience of virtuality runs through our everyday experience even more deeply.
As beings with binocular vision, our eyes record two distinct perspectives, which partially overlap. We have the experience of seeing a single perspective because these two perspectives are merged by the optic centers of our nervous system in a process called “binocular fusion.” This unified perspective is a virtual one.
This unified perspective is virtual because our visual system does not record this perspective. The objects recorded are present, but are perspectives like objects? Can they even be virtual per our working definition? Unlike objects, they are not transferable between persons or multiply observable. You and I can each see the coasters lying on the coffee table, but you and I cannot see each other’s perspective. We may not even be able to see from each other’s perspective simultaneously, metaphors aside.
If perspectives are not objects, and are not like objects, does the definition even apply? Yes. In a Matrix-style virtuality our optic system could fuse an image from binocular inputs generated by the experience machine. In such a case the perspective would be a virtual one, because the unified perspective need not be present for us to experience it. This failure of our definition is not a failure to apply to perspectives but a failure to include the species’ of perspectives under the genus of “object” in our ontology.
Perhaps a perspective is virtual only insofar as the objects within it are virtual? This definition fails on at least two counts. First, it would mean that whenever you go to a movie theater, if you sat in the front row and were to hold your head such that the only thing in your field of vision was the projected images of the film, your perspective would be a virtual one. Second, it would mean that you could have a hybrid perspective (one partially virtual and partially real), which is subject to a sorities argument as to just what would then constitute a virtual perspective or a real perspective. In addition, both face intuitive difficulties.
Thus, a virtual perspective is one that we experience though it does not correspond exactly to the records or inputs from which it is generated. Binocular fusion is a process of extrapolating the conjunction of the two initial records to form a unified perspective. This resulting perspective is a virtual one insofar as our optic system does not actually receive that perspective as its input but generates it. What we see is virtuality, not the reality each eye records from a distinct perspective.
Should We Value Our Everydayness Less Because of its Virtuality?
This poses something of a problem, for if what we see is a virtual perspective of the objects around us, then what we are seeing is not really what is around us according to Nozick. If an experience in virtuality is less valuable than an experience in reality, does this imply our experience of seeing is less valuable than we have thought?
Despite the interlacing of virtuality and reality, I think our value judgments should not change. The value in our lives is not negatively affected by the discovery of the virtuality of experiences.
Implicitly, you probably value your binocular fusion. You may never have thought about it before, but so long as you do not genuinely wish you constantly experienced the world as two separate images, you implicitly are placing a value on your binocular fusion, despite the resulting virtuality.
If you read the other examples, similar could be said for your ability to locate the sources of sounds, or have an unconscious alert to pay attention to unfamiliar stimuli. Merely because the world includes inputs or objects that appear (or do not) in your experience does not add or strip away the value your life already has. This is supported by two related observations.
First, reality is full of inputs which do not appear in our experience. If the discovery of the infrared portion of the light spectrum stripped away value already in our lives that would pose an argument against the discovery of new information in general. More likely, Nozick’s own conclusion would be that any value our lives already had was not negatively affected by the discovery of a possible input which we had not and do not experience, and that if it were negatively affected by the discovery then our initial valuation was reached in error because inputs or objects that do not appear in our experience do not add or strip away any value our lives already have.
Second, reality includes inputs which appear for some persons but not for others. If the value our lives already have were increased by the appearing of inputs in our experiences, then our lives’ values would vary with our varying abilities to experience inputs. Ie, the life of someone with 20/20 vision would be more valuable than the life of someone with 20/40 vision, because of the appearing of more visual inputs in the former’s experiences. Likewise, the life of someone with 20/20 vision when they were young and with 20/40 vision when they were older would be more valuable when they were young than when they were older, because of the appearing of more visual inputs in their youthful experiences. If we are unwilling to say that variations in sensory ability (and thus variations in resulting experiences) cause variations in the value of our lives, then we should also be unwilling to say that variations between inputs received and experiences cause variations in the value of our lives.
Thus, the fact the inputs we receive are different or can be different than what we experience is not enough to say that an experience is more or less valuable. This extends to virtuality. That something is virtual, that is, that the inputs we receive are different than what we experience, is not enough to say that an experience is more or less valuable.
The analogue with the experience machine should be easy to see. If you were to step into the experience machine, become famous for your philosophical insight and retired happy, when the simulation ended, would that happiness become less valuable simply because you discovered it was added to your experience of the world by virtuality?
The Equal Value for Equal Work Objection
The Equal Value for Equal Work Objection (EQO) is as follows:
Given that our conscious experience is already laced with virtuality, and we do not devalue that experience because of this, we should assign equal value for equal work done by the virtuality and reality we experience.
If I were an Objective-List Theorist with regards to well-being, but I limited my list to the experience of hearing a scale , the EQO entails that I should value the experience of hearing the Shepard’s tone at the end of “I am the Walrus” just as much as I value hearing any scale.
The EQO’s application to hedonism is even more straightforward. It would entail that we should value the pleasure we experience if it is caused by virtuality to the same degree we value the pleasure we experience caused by reality.
An argument against the EQO comes to hand easily. I’ll call it the Correspondence Objection. I suspect it is the first-line of defense Nozick himself would rely on. The Correspondence Objection goes something like this:
The EQO fails because our intuition is not based on our experience of objects but on the fact that the objects are really there.
First, some self-promotion: even if this cripples the EQO, I think a second counter-argument succeeds, the External World Objection (EXO). You can read it here.
I recognize that the Correspondence Objection seems superficial in its setup, but the concern (that we are not being deceived about the world around us) attached to the Correspondence Objection is not to be underestimated.
Correspondence Objection v1: Reality is Public, Virtuality is Private
So far as I’ve talked about virtuality, I’ve discussed experiences that were not tradable between persons or multiply observable. When you hear the mislocated sound in the Franssen Effect 2, I do not experience your mislocating the sound, even if I similarly mislocate it. The advantage, so far, of reality seems to be that it is public while virtuality is private. Privacy of experiences is not necessary for virtuality, however. In fact, it seems necessary that Star Trek-virtuality is public; Matrix-virtuality could be private or public.
If the experiences I have identified as virtual are private because they are of a different species than public virtuality, then the EQO is in trouble. Although the experiences I have identified as virtual are private, this is not because the virtuality we experience in our everydayness is of a different species than public virtuality would be.
The objects needed for binocular fusion, for the thermal grill, etc, are publicly available. Our experiences of (private) virtuality result because of how these objects interact with the statistically normal constitution of human beings.
As such, our (private) experience of virtuality resulting from binocular fusion is no different from the (private) experience of reality for someone with diplopia (a condition where one sees their two recorded perspectives superimposed on the same space). Though diplopia is a private experience, we do not discount it as less real than the public objects one experiences as superimposed. By extension, though the virtuality we experience in our everydayness may be private, we should not discount it as lesser than virtuality. Thus, whatever value we give public virtuality, private virtuality is not lesser simply because it is private.
Correspondence Objection 2: Public and Private Virtuality are Different, so it Makes Sense to Value Them Differently
The Correspondence Objection could instead differentiate private experiences of virtuality from public experiences of virtuality by their sources. A simulator is sufficient for private experiences of virtuality, but it is necessary for public experiences of virtuality.
This objection will also fail.
One might try and draw this distinction because one source is natural and the other is artificial, or human-made. Such a vague boundary immediately means that the pleasure you get from Shakespeare’s plays is necessarily different than the pleasure you get from a sunset.
A better way to try and draw this distinction would be to differentiate these because one source is naturally occurring and the other is a digital system. However, drawing the distinction this way would mean that we could simply wait until virtuality could be simulated by a non-digital system. A sufficiently complicated mechanical computer might suffice, or more likely, a quantum computer that didn’t use digital states. The redrawing of boundaries that would occur to further exclude systems that are not naturally occurring is ad hoc. More importantly, that we experience the world via the sensory integration of received photons, the vibrations of a medium, chemical compounds with gustatory properties, the non-quantum characteristics of masses, etc, is arbitrary. That these arbitrary sources of empirical experiences should count but not a digital source should count is hard to believe.
Correspondence Objection v3: Virtuality and Reality Interact with Our Senses Differently, and that Matters
Another attempt the Correspondence Objection could make to distinguish between virtuality and reality by source is to note where in our nervous system the source of the experience interfaces. In the case of reality it is at our sense organ. With virtuality it occurs upstream, either between the sense organ and the brain or at the brain itself. Notice that this does not disqualify Star Trek-virtuality, which would interface with our sense organ.
Our having the set of sense organs that we do is ultimately arbitrary, so to claim that reality is what is interfaces at these sense organs requires either that we just happen to be attuned to interface with reality or, more plausibly, that reality surrounds us and that our sense organs are incidentally of the sort that allow us to interact with it. Although I think it is trivial whether we experience, for example, the Franssen Effect 2 because it vibrated a medium which our eardrum translated into the electric signals used by the nervous system or because the nervous system directly received the appropriate electric signals, if one were intent on pushing this upstream/downstream argument, I would refer them to the EXO argument.
The EXO shows that our evidences for thinking our experiences are sourced in reality now and our evidences in the machine for thinking those experiences were sourced in reality would be indistinguishable. As such, the only discernible difference is their taking place out-of-the-machine and in-the-machine, which is to say their location or order. Although an experience’s location or order can warrant consideration, it is not enough to justify the intuition that we should not get into the machine when they would be otherwise identical.
This is a problem for the Correspondence Objection because it murkies the claim that we should value what interacts with our sense organs, as it is reality, but that we should value less what interacts upstream of our sense organs, as it is virtuality. The EXO shows any inputs we experience cannot assumed to be reality, prima facie.
Conclusion: An Experience (in The Machine) Should Not Count for Less Just Because It is Virtual
The conclusion from the EQO is that when virtuality does equal work, with regards to our experiencing it, then we should give it equal value as reality. I’ve supported this claim by arguing that we have already experienced virtuality in our everydayness, and that upon discovering that we have, our past experiences do not become less valuable.
If this is the case, then achieving fame in the experience machine should be considered on par in value as achieving fame outside of the experience machine. Should the EQO, or that parallel attack on the experience machine the EXO, succeed, then the thought experiment demands rewriting to ensure that the Mach Band that heightens the contrast between the virtuality inside the experience machine and the reality outside the machine does not bias our intuition. One possible setup is below:
Your scientist friend excitedly tells you about the “reality simulator” she has been building and testing. It now works, and perfectly simulates reality, she claims. The best part — her reality simulator allows an outside operator to input what the user should experience. Through this process, the world can be made perfect, experiences that should be impossible felt, etc. The workday ends and you both return to your homes. You go through your routine like normal and fall asleep easily.
The next day, you wake up. You have been working on a philosophy article for quite some time, and receive notice that it has accepted for publication in the field’s most prestigious journal.
Is your pleasure, or the satisfaction of your desire that this should happen, enough to say that your life is going well? [5]
Wait, you might say, am I in the reality simulator or not? Did this really happen?
That question is exactly the point!
Did it really happen? So far as you can tell, yes; so far as everyone you meet tells you, yes.
Without a Mach Band to distinguish between virtuality and reality, we are left in an uncomfortably familiar position — we must decide if the evidence we have is enough to prove that something has “really” happened. At which point we must further decide if pleasure alone is enough to give our lives value.
As of 2014, the state of technology is such that the virtual and nonvirtual have quite distinctive Mach Bands by which we can distinguish what is virtual and what is nonvirtual: graphics are imperfect, computing power is limited, handheld controllers that convert unconventional motions into analogous motions are necessary, some kind of projector is required to “immerse” ourself, computer intelligence is limited, etc. This differentiator is necessary to our being able to draw a conclusion from the experience machine.