Tuesday, June 2, 2009

Welcome to the Machine, Part 5: Simulation taxonomy

Previously in series: The Ethics of Simulated Beings, Descartes's Malicious Demon, The Simulation Argument, Kurzweil's Nano Neural Nets.

As shocking as the Simulation Argument is, it's (arguably) a revelation that's no less shocking than previous existential paradigm shifts. While undoubtedly disturbing to the people alive at the time, previous civilizations have come to grips with the knowledge that they do not live on a flat Earth nor at the center of the Universe.

Like the simulation argument, these previous scientific epiphanies assaulted humanity's sense of itself and its cosmic importance within the Universe. But just as it no longer troubles us to know that we don't live at the center of the Universe, it shouldn't bother us to know that we don't reside in the deepest reality. While it's tempting to diminish the "realness" or the validity of a virtual world, so long as certain attributes of existence exist, there's no good reason to value one realm over another.

This being said, there are a number of unanswered questions about the type of simulation we could be living in—answers to which could have a profound impact on our self-conception.

We do not have the means yet to determine whether or not we live in a simulation, let alone the means to determine its potential type and nature. But this hasn't prevented serious speculation; we may be able to describe and categorize the possible simulation types and varieties of virtual life:

Hard and soft simulations

The possibility exists, for example, for what philosopher Barry Dainton describes as hard and soft simulations. Hard simulations result from directly tampering with the neural hardware ordinarily responsible for producing experience whereas people running in a soft simulation have no corporeal source—they are exclusive streams of consciousness generated by computers running the appropriate software; there is no external hardware support.

The inhabitants of The Matrix had bodies that existed outside of the simulation, thus qualifying it as a hard simulation. Sensory experience could be directly machine-controlled through the stimulation of the appropriate areas of the sensory cortex and the movements of the simulated body would be under the control of the source mind, but there would be no need for the source body to actually move. As Morpheus noted, "What is real? How do you define real? If you're talking about what you can feel, what you can smell, what you can taste and see...then real is simply...electrical signals interpreted by your brain."

Complete and partial simulations

There's also the possibility for complete and partial simulations. In a complete simulation, every element of the experience is generated by artificial means (e.g. the complete suppression of all psychological characteristics (including memory) in favor of novel ones) .

But in a partial simulation only some parts or aspects of the experience are generated artificially (e.g. the person retains their individual psychology).

Active and passive simulants

Dainton also describes active and passive simulants. Actives are completely immersed in virtual environments, but they are in all other respects free agents—or, as Dainton concedes, free as any agent can be. Their actions are not dictated by the program, but instead flow from their own psychologies, even if these are machine-implemented.

Passive subjects, however, have a completely preprogrammed course of experiences. "The subjects may have the impression that they are autonomous individuals making free choices," writes Dainton, "but they are deluded." All their conscious decisions are determined by the program. They have apparent psychologies, and are conscious, feeling agents, he notes, but their real psychologies are entirely suppressed or nonexistent.

Original and replacement psychologies

Other varieties of simulated life include subjects who have either retained their original psychologies or are given entirely new ones. In an original psychology simulation, a simulant has an external existence outside the simulation and retains their original psychology -- again, The Matrix provides a good example. But in a replacement psychology situation, the simulant has external existence, but none of the original psychology is retained, only consciousness is transferred.

Communal and individual simulations

Simulation experiences could also be communal or individual.

Communal simulations have a virtual environment that is shared by a number of different subjects, each with individual and autonomous psychological systems.

In an individual simulation, however, there is only one real subject with an autonomous psychology; the other "inhabitants" of the simulation are merely automatons, parts of the machine-generated virtual environment. Communal and individual simulations could also be combined, where 'real' psychologies are intermixed with automatons. This scenario is (somewhat) explored in the 1999 film, The Thirteenth Floor.


Which leads to the next level of complexity, the idea that these simulation types could be mixed and matched. Indeed, if powerful simulation technology were to be commonplace it is by no means inconceivable that these simulations, particularly those of the hard variety, would be generated in sufficient numbers.

One thinker who has thought of the various different combinations is Tony Fleet. While there are as many as 32 different combinations, he argues that only 9 of them are viable and/or logically consistent. For example, in a partial simulation scenario, an external entity it required -- therefore this is only possible in the hard simulation case; a partial soft simulation is therefore impossible.

Fleet speculates that the only viable combinations can involve the communal/active; individual/active, individual/passive simulation types (be sure to check out his tables). That said, he does not believe that we've covered all simulation types. For example, there is no distinction between physical, virtual and mixed simulations. Some more work clearly needs to be done to create a complete simulation taxonomy along with all logically consistent combinations.

Posthuman vacations

This opens the door to some remarkable possibilities. How might these simulations and virtual reality experiences be utilized by our descendants, or even our future selves?

It's conceivable that people might take virtual reality 'trips' to the past quite frequently. They would also likely be used on an occasional basis during history lessons for those with a particular interest in experiencing what it was like to live during certain periods of the past (Bostrom's Ancestor Simulations come to mind).

But such trips might also be taken for entertainment purposes. A future activity in a posthuman world might very well involve regular immersive and interactive journeys into simulated realities. And in order to increase the authenticity of such adventures, it's quite possible that posthumans may choose to temporarily suppress their psychologies and memories. Of course, they would recall the entire experiencing after re-awakening in their genuine reality as their authentic selves.

Which means that you might actually be an autonomous simulant with a replacement psychology living in a hard simulation.

And if that's the case, now what? How are you supposed to live?

A topic I'll return to in my next post.

No comments:

Post a Comment