Talk:Consciousness and the Different Types of Simulation

From Simulism

I would welcome discussion on this. One of the reasons for suggesting this split, is the discussion about the need to create minds which are 'simulations in themselves' in the case of relativistic simulations, or (which has not yet been fully explored) simulations in which time travel were allowed. The Wikipedia sub-type 'solipsistic simulation' offers a possible methodology, but, I believe that this would present intractable procedural problems in attempting to implement such a solution on a 'brain-in-a'vat'; in essence, it can be argued that if this were to happen, free will for the simulee would be abolished.

--TonyFleet 10:38, 18 February 2007 (CET)

---

Interesting essay. The reason I use 3 degrees is that I not only take into account the 'realness of the mind' but also the 'realness of the body'. If I were consequent, this would need a 4th possibility, where the body is real but the mind is simulated (comparable to how Agent Smith, being a computer program, takes control over a real body in The Matrix Reloaded. This however is not really a 'simulation' in the way we describe it here.

Looking purely at the mind, I think your extrinsic/intrinsic distinction is a good one. --Ivo 14:00, 18 February 2007 (CET)

---

The reason I need to do this, is that I think that there are important arguments which limit the kind of simulation we could be in. I believe that evidence such as the absence of bugs AND the difficulties in programming relativistic events rule out 'reality' being an extrinsic simulation; furthermore, the difficulties of programming both relativity and quantum mechanics would lead me to believe the only possible simulation would be solipsistic. You will need to wait for further articles for the evidence for this, but the ideas are already scattered around the discussion pages.

--TonyFleet 15:31, 18 February 2007 (CET)

[edit] Anyone want to help with Some Definitions?

(from Dainton, B. (2002), 'Innocence Lost: Simulation Scenarios - Prospect & Consequences.)

Dainton defines some different types of simulation, and they all make sense, but I am struggling to get my head around the different combinations.

The definitions come in pairs:

Complete Simulation = every part and aspect of the simulation generated by artificial means.

Partial Simulation = some parts or aspects of experience are generated artificially (e.g. where a person retains their individual psychology).

Hard Simulation = simulations reulting from tampering with the neural hardware of the brain.

Soft Simulation = streams of consciousness running on computers; no external hardware support.

Active Simulation = simulants have free will and actions not dictated or confined to particular courses of actions other than the constraints of the simulation.

Passive Simulation = simulant has pre-programmed set of experiences.

Original Psychology Simulation = simulant has an external existence outside the simulation and retains original psychology.

Replacement Psychology Simulation = simulant has external existence, but none of the original psychology is retained, only consciousness is tranferred.

Communal Simulation = many simulants share the same experience.

Individual Simulation = one simulant only has the simulation experience.

There are potentially 32 separate combinations of these, but not all of the combinations exist.

I am attempting to clarify whether any of the combinations are, in theory, impossible.

Conclusion 1: If a partial simulation, then we require an external entity, therefore this is only possible in the H-simulation case. A Partial S-Simulation is therefore impossible.

Conclusion 2: If an S-simulation, then no external hardware support, so this only occurs in the Complete simulation case.

Conclusion 3:This appears to mean that there are three distinct combinations: Partial/Hard, Complete/Hard and Complete/Soft However, this now impinges on the Psychology category.

Partial/Hard means that there is an external entity who retains their Original Psychology. Complete/Hard means that there is an external identity who has a replacement psychology, and the consciousness only is transferred. Complete/Soft means that there is no external entity, but that the Psychology is entirely artificial, and is constructed. Unfortunately in this case, Dainton's categories do not work.

Conclusion 4: We would require therefore a new category in the case of an S-Simulation, which is a Constructed Psychology, as there is no original, and therefore none to replace.

Conclusion 5: If an H-simulation, it is possible to have either Replacement or Original psychology. If an S-Simulation, we can have neither, and must have a Constructed Psychology.

Dainton goes on to argue that a communal passive simulation is not feasible, as effectively if all that a group of minds is doing is witnessing the same situation, and there is no interaction, there is no community and therefore it cannot be regarded as communal.

Conclusion 6: We can only have the following: Communal/Active; Individual/Active, Individual/Passive.

This effectively reduces the number of possibilities to 9 distinct types. However, this disagrees with Dainton's analysis in the paper, as he has 12 types.

In addition:

The Individual Passive simulation is also problematic. If we have an artifical consciousness subjected to this , then it can only experience, it cannot affect the situation. In fact it is trapped in the simulation with no way out. The only reason that I could see for doing this would be to 'test' artifical consciousnesses to see how they would react under different situations. Even then we would need to 'fool' it into thinking that it has free will, otherwise, if the entity realised that none of its actions could affect its environment then the exercise would be pretty pointless. This is just about feasible, but for an H-Simulation, a replacement psychology to be subjected to a passive Individual simulation is a completely pointless exercise, unless this was a research project to see whether the replacement psychology would actually work. As an expereince for the simulant, this would have no value whatsoever; it would simply mean a transfer of consciousness, but with no memory, just to experience.

This analysis now leaves the possibilities as follows:

		H-Simulation	H-Simulation	S-Simulation	S-Simulation
		COMMUNAL	INDIVIDUAL	COMMUNAL	INDIVIDUAL
ACTIVE	ORIGINAL	yes	yes	no	no
ACTIVE	REPLACEMENT	yes	yes	no	no
ACTIVE	CONSTRUCTED	no	no	yes	yes
PASSIVE	ORIGINAL	no	yes	no	no
PASSIVE	REPLACEMENT	no	unlikely	no	no
PASSIVE	CONSTRUCTED	no	no	no	unlikely

Apart from experimental situations to see whether or not these things could, in principle be created, we are elft with effectively 7 workable types.

Questions:

1. Do these cover all simulation types? I do not believe so, as there is no distinction between physical, virtual and mixed simulations. Unfortunately this will mean unpicking the H-simulation and S-Simulation as Dainton's anlaysis assumes that the simulation is running on a computer and the brain is hard-wired or downloaded.

2. Are now these all distinct and possible, and are there examples out there in SF, Film etc. where these have all been used? If not, what would these look like?

Original --TonyFleet 20:49, 9 March 2007 (CET) Updated--TonyFleet 09:05, 16 March 2007 (CET)

[edit] Tables of Possibilities

Virtual simulations

The simulation is entirely virtual, an no physical part of the simulant enters into the simulation.

		H-Simulation	H-Simulation	S-Simulation	S-Simulation
		COMMUNAL	INDIVIDUAL	COMMUNAL	INDIVIDUAL
ACTIVE	ORIGINAL	Neo in Matrix	Vanilla Sky	no	no
ACTIVE	REPLACEMENT	yes	yes	no	no
ACTIVE	CONSTRUCTED	no	no	Agents in Matrix	Moriarty in Star Trek
PASSIVE	ORIGINAL	no	yes	no	no
PASSIVE	REPLACEMENT	no	unlikely	no	no
PASSIVE	CONSTRUCTED	no	no	no	unlikely

Physical/Mixed Simulations

Simulations which involve the physical presence of the simulant in the simuation, either through their embedding in the simulation, or through their physical movements being replicated in the simulation. Clearly in this case, we cannot have an all S-simulation, as the simulant must have a physical presence.

		H-Simulation	H-Simulation	S-Simulation	S-Simulation
		COMMUNAL	INDIVIDUAL	COMMUNAL	INDIVIDUAL
ACTIVE	ORIGINAL	Star Trek Holodeck	Truman Show?	no	no
ACTIVE	REPLACEMENT	yes	yes	no	no
ACTIVE	CONSTRUCTED	no	no	no	no
PASSIVE	ORIGINAL	no	yes	no	no
PASSIVE	REPLACEMENT	no	unlikely	no	no
PASSIVE	CONSTRUCTED	no	no	no	no

Does anyone want to suggest any more, or argue with my categorisations?

--TonyFleet 10:09, 16 March 2007 (CET)