Jared Edward Reser

Abstract

The generic simulation hypothesis states that an observer and the apparent world surrounding that observer are implemented by a computational process in a more fundamental reality. The hypothesis does not, by itself, specify how many of the apparent inhabitants of the simulated world are conscious. Nick Bostrom’s simulation argument ordinarily concerns ancestor simulations populated by many conscious people, but Bostrom also briefly considered selective simulations containing only a small group or a single fully simulated individual, surrounded by “shadow-people” modeled only well enough to avoid detection. Grace Helton has more recently argued that ethical or computational considerations might lead simulators to instantiate some sentient beings while rendering others as convincing but mindless automata. David Chalmers accepts the coherence of such solipsistic simulations but argues that their anthropic importance may be limited because each contains only one conscious observer.

This paper develops a broader framework for analyzing the distribution of consciousness within simulated worlds. It introduces consciousness density, the proportion of apparent agents who are conscious, and consciousness occupancy, the proportion of apparent agent-time that is accompanied by experience. These variables distinguish population-complete simulations from consciousness-sparse, singleton, and intermittently conscious simulations. The paper also develops the concept of an observer-centered world, in which environmental and social fidelity are maintained primarily relative to one focal observer or a small focal group. Such a world need not contain strict philosophical zombies that duplicate human functioning under every possible condition. Its nonconscious agents need only remain behaviorally adequate within the interactions and investigations accessible to the focal observer.

These distinctions yield the solipsistic burden of simulation belief. Assigning substantial credence to an unspecified simulation hypothesis requires assigning some conditional credence to sparse-conscious and singleton architectures unless they can be independently excluded. The argument does not establish that singleton simulations are probable. Their probability depends on simulator motives, the relationship between behavior and consciousness, environmental sharing, and the scaling of computational cost with conscious population. A cost-adjusted observer model shows that Chalmers’s numerical objection succeeds when densely conscious simulations generate more conscious observer-moments per unit of computation, but it can weaken or reverse when observer-centered singleton worlds are substantially cheaper per conscious observer. The paper concludes that popular simulation belief often tacitly assumes a socially complete world in which friends, relatives, and strangers retain independent inner lives. Once that assumption is separated from the generic simulation hypothesis, simulation belief acquires a neglected uncertainty about social reality itself.

Keywords: simulation hypothesis, consciousness, solipsism, other minds, observer-centered simulation, philosophical zombies, artificial agents, anthropic reasoning, meaningful life, consciousness density

1. Introduction

Many discussions of the simulation hypothesis treat the hypothesis as a change in the substrate of reality. Trees, houses, planets, bodies, and brains remain much as they appear, but their physical organization is ultimately implemented by computation in another universe. Under this interpretation, discovering that reality is simulated would revise the metaphysical account of what objects are made of while leaving most ordinary relationships intact. Other people would still be conscious, their emotions would still be genuine, and social life would remain a shared encounter among minds.

That comforting interpretation contains an additional assumption. It assumes that the apparent population of the world corresponds to its conscious population. The generic claim that a world is simulated does not guarantee this correspondence. A simulator may implement every apparent person as a continuously conscious agent, but it may instead instantiate only some of them, only one of them, or different agents at different times.

Suppose an observer looks toward a tree, a street sign, and a friend. Under the simulation hypothesis, the tree and sign may be computationally modeled structures rather than base-level physical objects. The same is true of the friend’s visible body and behavior. The simulator could be implementing a conscious mind behind that behavior, but it could also be producing the behavior through a system that lacks subjective experience. From the focal observer’s perspective, both architectures could generate the same conversation, remembered history, affectionate expression, and apparent social continuity.

This possibility does not mean that someone who believes in the simulation hypothesis must believe that everyone else is unconscious. It means that the generic hypothesis leaves the issue unresolved. A simulation believer who assumes that all apparent people are conscious has adopted a particular simulation architecture rather than the simulation hypothesis in its most general form.

Nick Bostrom’s original argument concentrated on ancestor simulations containing large populations of conscious humanlike beings. Bostrom nevertheless explicitly considered more selective simulations containing a small group or one individual, with the rest of the apparent population implemented as zombies or “shadow-people.” He also noted the possibility of abridging portions of simulated mental lives and supplying false memories for the omitted intervals. Bostrom concluded that one-person simulations would have to be extraordinarily numerous to dominate observer counts, but the conceptual possibility was present from the beginning. (Simulation Argument)

Grace Helton has recently developed the epistemic and axiological consequences of this possibility. She argues that simulators might avoid creating billions of sentient beings because of computational expense or concern about the suffering those beings would experience. They could instead instantiate some conscious creatures and render the rest as convincing but mindless automata. Helton’s central concern is that preserving knowledge of tables, shrubs, and other nonsocial structures would not preserve knowledge of other minds, even though social knowledge contributes centrally to the meaningfulness of many lives. (Grace Helton)

David Chalmers accepts that a solipsistic simulation is coherent and grants that simulators may prefer such worlds for reasons of ethics or efficiency. He nevertheless argues that the number of conscious observers creates an important obstacle to concluding that a simulated observer is probably alone. A million simulations containing one conscious person each still contain fewer conscious observers than a single simulation containing one billion conscious people. Chalmers calls this observer weighting a “solipsistic sim blocker.” (Consciousness Network)

The present paper builds on these precedents but shifts the focus from the binary question of solipsism to the general distribution of consciousness inside simulated worlds. A simulation can contain every apparent person as a conscious being, one focal consciousness surrounded by nonconscious agents, or any intermediate configuration. It can also alter that distribution over time, maintaining some minds continuously while instantiating others only during selected periods.

The central thesis is:

The generic simulation hypothesis contains a hidden population parameter: the number, distribution, continuity, and degree of consciousness instantiated among the apparent inhabitants of the simulated world.

Recognizing this parameter produces several consequences. The apparent size of a simulated universe does not determine how many conscious minds it contains. Believable social behavior does not necessarily require the complete simulation of a human brain under every possible condition. Observer counting must incorporate the cost of producing worlds with different conscious populations. Finally, assigning substantial credence to an unspecified simulation hypothesis transfers some credence to consciousness-sparse worlds unless those worlds can be independently ruled out.

This conclusion will be called the solipsistic burden of the simulation hypothesis. The burden is a requirement of credal accounting rather than a proof that solipsism is likely. Someone can rationally conclude that a singleton simulation is extremely improbable, but that conclusion requires assumptions about simulation architecture, simulator motives, artificial consciousness, behavioral modeling, and computational economies of scale. It does not follow from the bare statement that reality is simulated.

2. From Ancestor Simulations to Selective Simulations

2.1 Bostrom’s population-complete starting point

Bostrom’s argument begins by imagining technologically mature civilizations with enough computational capacity to run detailed simulations of their evolutionary predecessors. If the simulated people are conscious, and if posthuman civilizations create vast numbers of them, simulated humanlike observers could greatly outnumber biological originals. The argument then uses observer weighting to suggest that an individual with human-type experiences should assign substantial probability to being among the simulated majority. (Simulation Argument)

This reasoning presupposes that the simulated population contains conscious observers. Bostrom adopts a qualified substrate-independence assumption according to which consciousness could arise in a nonbiological system if the relevant computational organization of the brain were reproduced in sufficiently fine-grained detail. He does not claim that behavioral imitation alone would suffice. His assumption is that structurally reproducing the consciousness-relevant processes of the brain could produce subjective experience. (Simulation Argument)

The familiar ancestor simulation is therefore population-complete with respect to its intended human subjects. The simulated people whose histories are being reconstructed possess their own experiences. Their relationships are relationships among conscious beings, and their social histories are jointly inhabited from multiple first-person perspectives.

The ancestor-simulation framework can encourage the impression that any large simulation would have this property. Yet Bostrom himself did not restrict all possible simulations to population-complete worlds. He considered selective simulations containing only a small number of fully simulated people or a single person. The remaining human representations would be modeled only at the level necessary to prevent the target observers from noticing irregularities. (Simulation Argument)

2.2 Selective and singleton worlds

A selective simulation instantiates only a subset of the apparent population with the full organization required for consciousness. A singleton simulation instantiates one focal observer while presenting an apparent social world containing many other people.

These terms concern the simulated world rather than reality as a whole. A singleton simulation would still have creators, computers, and perhaps entire civilizations outside it. It therefore would not imply global metaphysical solipsism, according to which only one mind exists anywhere. It would imply local phenomenal solipsism: only one consciousness exists among the apparent inhabitants of the focal world.

The singleton architecture can take several forms. A biological brain in base reality might receive simulated sensory inputs. An artificial consciousness might inhabit a virtual environment. A completely simulated brain could be implemented alongside a behaviorally generated social world. A small group of conscious participants might share a world whose remaining population consists of artificial characters.

These possibilities reveal that the term simulation does not determine its unit of simulation. The unit could be an entire cosmological history, a planet, a city, a social experiment, a laboratory, one lifetime, one decision process, or one conscious stream. The apparent universe experienced by a focal observer may be much larger than the dynamically implemented portion of the world.

2.3 The Helton-Chalmers exchange

Helton argues that simulation scenarios create distinctive difficulties for knowledge of other minds. Structural or causal-role accounts may vindicate the reality of simulated physical objects while leaving the consciousness of apparent people unresolved. A simulator could have ethical reasons to avoid creating minds destined to suffer, or practical reasons to avoid the computational burden of simulating billions of psychologies. It might instantiate some sentient beings while representing the rest through convincing automata. (Grace Helton)

Chalmers grants that a world of this kind is coherent. He distinguishes strict zombies, which duplicate a person’s physical or functional organization without consciousness, from the less demanding possibility of behaviorally convincing “faux-folk.” He doubts that a simple lookup-table system could efficiently duplicate unrestricted human behavior, but he acknowledges that advanced behaviorally equivalent systems are difficult to exclude. (Consciousness Network)

Chalmers also sharpens the value issue. The principal loss in a solipsistic world may not be merely a loss of knowledge that other minds exist. The deeper loss is the absence of those minds themselves. Affection, friendship, understanding, and shared accomplishment would be deprived of their reciprocal subjective dimension even if the focal observer falsely believed the relationships were mutual. (Consciousness Network)

This exchange establishes the immediate philosophical background for the present analysis. The remaining task is to treat consciousness distribution as a variable, connect it to simulation architecture, and determine how it should affect belief in generic simulation hypotheses.

3. The Hidden Consciousness-Population Parameter

3.1 The generic simulation hypothesis

Let (S) denote the generic simulation hypothesis:

[
S:
\text{The focal observer and apparent world are implemented by a computational process in a host reality.}
]

This definition deliberately leaves several matters unspecified. It does not state whether the entire universe is continuously represented, whether only the observer’s causal surroundings are generated, whether the simulation is historical or experimental, or whether every apparent person possesses consciousness.

Let (\mathcal{A}) denote the set of simulation architectures compatible with (S). These architectures can include:

• (A_F): a population-complete simulation;
• (A_P): a partially conscious or consciousness-sparse simulation;
• (A_G): a small-group simulation;
• (A_1): a singleton simulation;
• (A_I): an intermittent simulation in which conscious minds are instantiated only during selected intervals;
• (A_E): an emergent world simulation in which consciousness arises through bottom-up physical evolution.

A rational credence in the generic hypothesis must be distributed across this architecture space:

[
P(S\mid E)

\sum_{a\in\mathcal{A}}
P(S,a\mid E),
\tag{1}
]

where (E) represents the observer’s evidence.

Equation 1 exposes the hidden assumption in many informal endorsements of simulation theory. A person may report:

[
P(S\mid E)=0.8
]

while imagining almost exclusively (A_F), a complete shared world containing a conscious counterpart for every apparent human being. That person has not assigned 80 percent probability to the generic hypothesis alone. The actual object of belief is closer to:

[
P(S,A_F\mid E)=0.8.
]

This is a much narrower proposition. The distinction matters because other architectures have different epistemic, moral, social, and anthropic implications.

3.2 A taxonomy of consciousness distribution

The simulation design space can be organized according to its conscious population:

Architecture

Conscious inhabitants

Apparent inhabitants

Social structure

Population-complete

Most or all apparent people

Full apparent population

Reciprocal conscious society

Consciousness-sparse

A minority

Full apparent population

Mixed conscious and nonconscious agents

Group-centered

A selected group

Larger apparent population

Shared world for focal participants

Singleton

One focal observer

Potentially billions

Apparent society without other local consciousness

Intermittent

Variable across time

Full apparent population

Minds activated, suspended, or abridged

Emergent world

Determined by simulated dynamics

Determined by simulated dynamics

Consciousness arises wherever sufficient organization develops

These categories are not mutually exclusive in every respect. A group-centered simulation could be intermittent, and an emergent simulation could later be selectively modified. The taxonomy serves to separate population assumptions that are usually bundled together.

The table also shows why simulation size and conscious population are independent. A small virtual room could contain ten conscious minds. An apparent cosmos containing billions of galaxies could be organized around one consciousness. Apparent spatial extent is therefore not a reliable proxy for the number of subjects experiencing the world.

4. Consciousness Density and Consciousness Occupancy

4.1 Instantaneous consciousness density

Let (A(t)) be the set of apparent person-like agents represented at time (t). Let (C(t)\subseteq A(t)) be the subset that instantiate consciousness at that time.

The instantaneous consciousness density is:

[
\delta(t)

\frac{|C(t)|}{|A(t)|}.
\tag{2}
]

In a population-complete simulation:

[
\delta(t)\approx 1.
]

In a singleton simulation containing one conscious observer and eight billion apparent people:

[
\delta(t)
\approx
\frac{1}{8\times10^9}.
]

This definition is a population ratio rather than a physical concentration of consciousness in space. It asks how densely subjective perspectives populate the apparent social world.

The binary distinction between conscious and nonconscious agents is an initial simplification. A more general model could assign each apparent agent a consciousness measure (q_i(t)), representing degree, richness, probability, or theory-relative consciousness status:

[
\delta_q(t)

\frac{\sum_{i\in A(t)}q_i(t)}
{|A(t)|}.
\tag{3}
]

Equation 3 accommodates minimally conscious, partially integrated, or uncertain artificial agents. It does not require agreement that consciousness is fundamentally scalar. The (q_i) values can instead summarize a multidimensional assessment for a particular analytical purpose.

4.2 Consciousness occupancy through time

Population density at one moment does not capture intermittent architectures. A simulator could instantiate an apparent person consciously during conversations with the focal observer, suspend or simplify the person afterward, and later reactivate the process with memories describing the intervening period.

To represent this possibility, define consciousness occupancy over interval (T):

[
\Omega_T

\frac{
\int_0^T |C(t)|,dt
}{
\int_0^T |A(t)|,dt
}.
\tag{4}
]

(\Omega_T) measures the fraction of apparent agent-time accompanied by consciousness. A world can have a high instantaneous density during crowded interactions while maintaining a low long-run occupancy because most apparent private lives are not consciously experienced.

Bostrom anticipated a related possibility when he considered abridging portions of simulated mental lives and supplying false memories for omitted experiences. The idea can be extended from the focal observer to surrounding agents. A person’s apparent biography could include years of remembered activity even if only selected periods were instantiated as conscious episodes. (Simulation Argument)

A simulation could therefore attenuate consciousness along at least three population dimensions:

[
\text{number of conscious agents},
\qquad
\text{duration of their consciousness},
\qquad
\text{richness of each conscious stream}.
]

These dimensions can vary independently. A simulator might preserve one continuous and phenomenologically rich observer while reducing the number and duration of other conscious streams.

4.3 Social attenuation without focal attenuation

This possibility motivates the concept of social attenuation without focal phenomenological attenuation:

A simulation preserves the richness, continuity, and causal organization of a focal consciousness while reducing the number, duration, autonomy, or conscious status of the agents surrounding it.

The focal observer’s experience can remain vivid and socially elaborate. The attenuation occurs in the unobserved subjective side of the apparent relationships. From the focal perspective, a friend may continue to speak intelligently, remember prior conversations, provide emotional support, and respond to new circumstances. What is missing is the friend’s independent first-person experience.

This extends the distinction between resource capacity and actual conscious allocation. A simulator that cannot or does not wish to instantiate billions of conscious streams need not diminish the focal observer. It can preserve one stream at full fidelity while economizing on the rest of the social world.

5. Observer-Centered Worlds

5.1 World-centered and observer-centered simulation

A world-centered simulation attempts to maintain a self-consistent environment independently of any particular inhabitant. Its objects and agents continue evolving whether or not they interact with a designated observer. A bottom-up cosmological simulation and a full ancestor simulation are world-centered in this sense.

An observer-centered simulation allocates representational and computational resources primarily according to the causal, perceptual, and investigative requirements of one focal observer or a small focal group. Distant regions can remain compressed, unseen histories can be stored as summaries, and apparent agents can be represented at the level required for interactions with the focal set.

Observer-centered does not mean that the world must visually appear to revolve around the observer. The observer can encounter surprises, resistance, complex institutions, unfamiliar people, and events that seem to have independent causes. The relevant dependence exists at the implementation level rather than the narrative level.

Bostrom already described an observer-sensitive physical simulation in which distant astronomical objects are compressed and microscopic details are supplied when inhabitants examine them. He also considered correcting anomalies by revising brain states or rerunning short intervals. The same general architecture can be extended from physical observation to social interaction. (Simulation Argument)

5.2 The focal causal envelope

Let (F) denote the focal observer or focal group. Define (\mathcal{I}_F) as the interaction envelope containing the observations, conversations, experiments, interventions, and social relationships accessible to (F) during the simulated interval.

A surrounding agent (j) need not duplicate an ordinary human across every logically or physically possible interaction. It needs to satisfy an observer-relative adequacy condition:

[
B_j(x)
\approx
H_j(x)
\quad
\text{for all }x\in\mathcal{I}_F,
\tag{5}
]

where:

• (B_j(x)) is the behavior produced by the simulated agent;
• (H_j(x)) is behavior that would be expected from a corresponding conscious human;
• (\mathcal{I}_F) is the set of interactions that the focal observer can actually initiate or encounter.

Equation 5 is weaker than unrestricted functional equivalence:

[
B_j(x)

H_j(x)
\quad
\text{for all possible }x.
\tag{6}
]

This difference is crucial. Chalmers worries that a system such as Block’s hypothetical lookup-table “Blockhead” would face combinatorial explosion if it had to contain a predetermined response for every possible human interaction. An observer-centered system need not store every response in advance or reproduce a human under every counterfactual condition. It can generate responses dynamically, exploit a model of the focal observer, and maintain fidelity only across the interactions that enter the simulated history. (Consciousness Network)

5.3 Interaction-bounded agents

An interaction-bounded agent is an apparent person whose implementation is designed to remain coherent within a finite social testing envelope. Such an agent might contain:

• a personality model;
• autobiographical records;
• relationship-specific memories;
• current goals and emotional variables;
• a model of the focal observer;
• language and action-generation systems;
• mechanisms for maintaining long-term consistency;
• compressed descriptions of activities performed outside the focal observer’s presence.

This architecture may be far less demanding than reproducing a complete human brain and its continuous private experience. It would still be more complex than a simple scripted character, particularly if the focal observer maintains close relationships, asks unexpected questions, and observes the agent over decades.

Research on generative agents provides an early technological illustration of observer-relative social coherence. Park and colleagues created language-model-based agents that stored experiences, formed higher-level reflections, made plans, communicated, developed relationships, and coordinated social activities in a shared environment. Their results do not establish whether such systems are conscious, nor do they approach unrestricted human equivalence. They demonstrate that coherent social behavior can be generated through architectures that are not conventional neuron-by-neuron human simulations. (arXiv)

The argument for consciousness-sparse simulations does not depend on current language models remaining nonconscious. It depends on the possibility that some systems can generate behavior adequate for the focal observer while omitting at least some of the organization required for humanlike consciousness. Whether this separation is possible is a substantive question in the science and philosophy of consciousness.

6. Strict Zombies Are Not Required

6.1 Philosophical zombies and social simulacra

A strict philosophical zombie is physically or functionally identical to a conscious person while lacking experience. The metaphysical possibility of such beings is highly controversial. A functionalist will generally deny that complete functional identity can coexist with a difference in consciousness, while a property dualist may treat the scenario as conceivable.

The present argument does not require strict zombies. The apparent people surrounding a focal observer need not be perfect physical or functional duplicates of humans. They need only produce sufficiently humanlike behavior across the focal observer’s interaction envelope.

This gives three increasingly demanding categories:

1. Scripted character: capable of a narrow and predetermined range of behavior.
2. Interaction-bounded social agent: dynamically generates coherent behavior across the focal observer’s accessible interactions.
3. Full human functional duplicate: reproduces human cognition and behavior across the relevant range of possible circumstances.

Only the third category raises the strongest zombie dispute. A consciousness-sparse simulation may rely primarily on the second.

6.2 Behavioral adequacy and hidden internal differences

Two apparent people can display similar bodies and behavior while having radically different implementations. One may contain a continuously running consciousness-generating architecture. The other may consist of a language-and-action policy, episodic memory store, and social consistency model.

This asymmetry would ordinarily be surprising in base reality because human bodies share evolutionary origins, homologous nervous systems, similar development, and comparable dependence on brain function. Under normal biological assumptions, the best explanation of another human’s behavior is that a brain similar to one’s own produces a mind similar in broad respects.

A designed simulation introduces the possibility of implementation asymmetry:

Apparent similarity among simulated agents need not imply similarity in the host-level processes implementing them.

One simulated skull could contain a detailed model of neural activity. Another could present the visible and instrumentally accessible consequences of neural activity while its social behavior is generated elsewhere by a more economical architecture. The focal observer could see brain scans, observe behavior following apparent injury, and conduct experiments without the apparent brain being the true causal source of the agent’s behavior.

This possibility weakens the inference from apparent biological similarity to implementation similarity. It does not eliminate the inference. A complete bottom-up simulation in which every brain evolves through the same laws would restore much of the ordinary abductive basis for belief in other minds.

6.3 Conditional escalation

A hybrid simulator could alter an agent’s implementation when the focal observer begins investigating it more deeply. A socially peripheral figure might initially be represented through a compressed behavioral model. If the focal observer develops a close relationship with that figure, the simulator could allocate a richer cognitive architecture or instantiate a conscious stream.

The reverse transition is also conceivable. A conscious agent might be suspended after leaving the focal observer’s life, while a compressed record preserves the information needed for a later encounter. Memories could be constructed to bridge the inactive interval.

These transitions raise serious metaphysical and ethical questions. A reactivated agent might regard the inserted memories as genuine memories of experiences that never occurred. A temporarily conscious agent might be created for an interaction and then permanently discontinued. Consciousness occupancy therefore has ethical significance beyond simple population counting.

7. The Simulation-Intensified Problem of Other Minds

7.1 The ordinary problem

Consciousness is directly available only from the first-person perspective. An observer experiences one mind directly and infers other minds from behavior, verbal reports, neural similarity, developmental history, and shared biological organization.

The classical problem of other minds asks whether these inferences provide knowledge. In ordinary life, the evidential case is nevertheless powerful. Other people respond flexibly, describe private experiences, show systematic relationships between brain states and cognition, and possess bodies generated by the same evolutionary and developmental processes as one’s own.

Solipsism remains logically difficult to disprove, but it is ordinarily explanatorily extravagant. The hypothesis that everyone else lacks consciousness adds an unsupported difference to otherwise similar biological systems.

7.2 Engineered asymmetry

The simulation hypothesis changes the character of this problem. It introduces a possible designer who may deliberately implement superficially similar agents through different underlying methods.

The question is no longer limited to:

[
\text{Can a physically similar organism lack consciousness?}
]

It becomes:

[
\text{Does the apparent organism possess the same underlying implementation at all?}
]

A simulation can preserve visible causal structure while altering hidden computational structure. The creator may have reasons to reserve expensive or ethically consequential consciousness-generating processes for selected agents.

Simulation theory therefore transforms an abstract skeptical gap into a possible engineering choice. The uncertainty about other minds becomes uncertainty about consciousness allocation.

7.3 The residual abductive case

The introduction of engineered asymmetry does not make it rational to distrust other people in ordinary life. Several considerations can still favor a population-complete interpretation.

A world with uniform laws and similar brains is simpler to generate and scientifically useful as a model of an entire civilization. Other people exhibit enormous behavioral depth, surprise, creativity, resistance, and private activity. Our apparent world contains institutions, records, conflicts, and discoveries that are not organized around one person’s convenience.

These observations fit naturally with a shared reality containing many autonomous minds. A sufficiently advanced observer-centered simulator could reproduce them, but possibility is not probability.

The present argument is conditional. It identifies an uncertainty internal to generic simulation belief. It supplies no empirical evidence that anyone is an artificial construct, and it does not weaken the practical or moral presumption that other people are conscious.

8. The Solipsistic Burden

8.1 A burden of conditional probability

Let (L) denote the singleton hypothesis:

[
L:
\text{The focal observer is the only conscious inhabitant of the apparent world.}
]

The law of total probability gives:

[
P(L\mid E)

P(L\mid S,E)P(S\mid E)
+
P(L\mid \neg S,E)P(\neg S\mid E).
\tag{7}
]

If engineered singleton worlds are more feasible under simulation than under nonsimulated biological reality, then:

[
P(L\mid S,E)

P(L\mid \neg S,E).
\tag{8}
]

Under that condition, increasing one’s credence in (S) increases one’s credence in (L), all else being equal. The increase may be extremely small, but it cannot be ignored while the generic simulation hypothesis remains unspecified.

This yields the Solipsistic Burden Principle:

A person who assigns substantial credence to the generic simulation hypothesis must assign some corresponding credence to singleton and consciousness-sparse simulations, unless independent considerations reduce their conditional probability to zero or near zero.

The principle does not claim:

[
P(L\mid S,E)\approx 1.
]

It claims that the value of (P(L\mid S,E)) requires justification. One cannot endorse a broad simulation hypothesis and silently set every socially sparse architecture aside.

8.2 Entailment, possibility, and rational openness

Three claims should be distinguished:

Entailment: If we are simulated, I am the only conscious inhabitant.

This is false. Population-complete simulations are possible.

Conceptual compatibility: If we are simulated, I could be the only conscious inhabitant.

This is true under at least some coherent architectures.

Credal obligation: If I assign substantial probability to unspecified simulation, I should allocate some of that probability to compatible singleton or sparse architectures unless I have grounds for excluding them.

This is the argument defended here.

The third claim is the philosophically important one. Simulation believers often acknowledge singleton worlds as logically possible while assigning them no practical place in their credence. That omission can be rational only if population-complete simulations are overwhelmingly favored by cost, motive, architecture, consciousness theory, or observer-selection effects.

8.3 Restricted simulation belief

A person can avoid the burden by endorsing a restricted hypothesis such as:

[
S_F:
\text{The world is a complete bottom-up simulation in which all human brains are implemented by the same consciousness-supporting laws.}
]

Under (S_F), the consciousness of other people follows from the same psychophysical principles that make the focal observer conscious. The problem of other minds remains in its ordinary form, but simulation adds little new asymmetry.

The cost of this solution is that (S_F) is more specific than (S). Evidence or arguments supporting generic simulation do not automatically support a complete bottom-up architecture. A person who endorses (S_F) must explain why simulators would choose that architecture rather than selective, compressed, observer-centered, or hybrid alternatives.

9. The Economics of Conscious Population

9.1 Chalmers’s solipsistic sim blocker

Chalmers observes that a large number of singleton worlds can contain fewer conscious observers than a small number of populated worlds. Suppose there are:

[
10^6
]

singleton simulations containing one conscious observer each, and one population-complete simulation containing:

[
10^9
]

conscious observers. The singleton worlds outnumber the complete worlds by one million to one, but conscious observers in the complete world outnumber singleton observers by one thousand to one. (Consciousness Network)

This is a decisive correction to reasoning that counts simulations rather than conscious observers. Under an observer-weighted framework, the relevant quantity is not the number of world instances. It is the number of observers or observer-moments contained in each class of world.

The blocker does not settle the matter because the number of simulation instances is not necessarily independent of their cost. If one populated world consumes the resources required to run many singleton worlds, the number of worlds generated under a fixed budget will vary accordingly.

9.2 A cost-adjusted model

Let:

• (m) be the number of apparent people in the world;
• (n) be the number of conscious people, where (1\leq n\leq m);
• (C(n,m,T)) be the cost of running that world for simulated duration (T);
• (B_n) be the resources allocated to simulations with conscious population (n).

The number of worlds that can be produced is approximately:

[
W_n

\frac{B_n}{C(n,m,T)}.
\tag{9}
]

The number of conscious observer-moments produced is:

[
O_n

\frac{B_n nT}{C(n,m,T)}.
\tag{10}
]

The relevant production efficiency is therefore:

[
\eta_n

\frac{nT}{C(n,m,T)}.
\tag{11}
]

A densely conscious architecture dominates observer counts when it produces more relevant conscious observer-moments per unit of allocated resource. A singleton architecture gains anthropic importance when its observer-centered compression makes one conscious stream sufficiently inexpensive.

For a full simulation with (N) conscious people and a singleton simulation with one conscious person, assume equal duration and equal budget allocation. Full simulations generate more conscious observer-moments when:

[
\frac{N}{C(N,N)}

\frac{1}{C(1,N)}.
\tag{12}
]

Equivalently:

[
C(N,N)
<
N C(1,N).
\tag{13}
]

Equation 13 means that one world containing (N) conscious observers is cheaper than (N) separate singleton worlds. This is an economy of conscious population scale.

Singleton worlds generate more observer-moments when:

[
C(N,N)

N C(1,N).
\tag{14}
]

Equation 14 describes diseconomies of conscious population scale. In this case, one world containing (N) independently conscious agents costs more than (N) observer-centered worlds containing one consciousness each.

9.3 Sources of population economies

Densely conscious simulations can share many resources:

• one environmental history;
• one physical world;
• common geographic and astronomical structure;
• shared institutions;
• shared cultural records;
• interactions computed once for all participants;
• infrastructure supporting many minds;
• common compression of laws and initial conditions.

If environmental simulation dominates total cost, adding another conscious mind to an existing world may be much cheaper than constructing a separate personal world. The cost per observer may decline sharply with population.

A complete bottom-up universe also gains simplicity from uniformity. The simulator can apply the same laws to every brain rather than maintaining special implementation rules for focal and nonfocal agents. Uniform causal evolution may be easier to verify, scientifically interpret, and protect from inconsistencies.

These considerations support Chalmers’s blocker. A relatively small number of population-complete simulations may produce far more conscious observer-moments than a much larger number of singleton simulations.

9.4 Sources of population diseconomies

Observer-centered simulations may gain their own efficiencies. A singleton world need not generate the full private lives of billions of agents, preserve every inter-agent relationship, or maintain independent causal detail in regions inaccessible to the focal observer.

A fully conscious society requires each person to have:

• a continuous stream of experience;
• private perception and memory;
• independent plans;
• interactions with many other agents;
• reciprocal modeling of social partners;
• causally consequential activity outside the focal observer’s awareness;
• long-term developmental and emotional continuity.

As the number of autonomous conscious agents increases, the number of socially significant relationships and possible interactions may grow faster than the population. This does not mean every pairwise relationship must be modeled at equal depth, but multi-agent autonomy can produce novel information and difficult counterfactual demands.

A singleton world can compress most of this structure. Apparent people need detailed lives only where those lives affect the focal observer’s causal and epistemic envelope. Their unobserved activity may be represented through summaries and generated consequences.

The critical empirical and theoretical question is:

Does simulated consciousness exhibit economies or diseconomies of population scale after environmental sharing and observer-centered compression are both included?

Neither Bostrom’s original estimate nor Chalmers’s numerical example answers this question. Bostrom noted that one-person simulations would need to be enormously more numerous than ancestor simulations to dominate observer counts, but the relative cost and motives determining those numbers remain uncertain. (Simulation Argument)

9.5 Budget allocation and simulator motives

Even cost efficiency is insufficient by itself. Simulators may allocate unequal resources to different architectures:

[
B_1
\neq
B_N.
]

The posterior distribution of observers depends on both production efficiency and simulator preference:

[
O_n
\propto
B_n\eta_n.
\tag{15}
]

A civilization may devote most of its simulation resources to historical reconstruction, in which case population-complete worlds dominate even if singleton worlds are cheaper. Another civilization may conduct huge numbers of individualized training, therapeutic, entertainment, or decision simulations. The same cost structure could then produce a very different observer distribution.

Observer counting must therefore model at least three unknowns:

[
\text{cost per architecture},
\qquad
\text{resource allocation},
\qquad
\text{number of conscious observer-moments}.
]

Counting world instances or apparent inhabitants alone is inadequate.

10. Possible Simulator Motives

10.1 Historical and scientific reconstruction

An ancestor simulation designed to study social evolution, institutions, disease, war, migration, technological development, or cultural change may require many autonomous minds. Removing consciousness might also alter behavior if consciousness contributes causally to learning, judgment, motivation, and social cognition.

Population-complete simulation is therefore a natural architecture for some historical purposes. It preserves reciprocal social causation and allows macrohistorical events to emerge from many independent perspectives.

A cosmological or evolutionary simulation might also generate consciousness unintentionally. The simulator could specify laws and initial conditions, after which minds emerge wherever the simulated dynamics produce consciousness-supporting systems.

10.2 Focused experiments

Other research questions may concern one decision-maker, one cognitive architecture, or one sequence of experiences. A simulator studying memory, trauma, creativity, social influence, or moral choice could focus resources on one target while generating the surrounding world as experimental context.

A full population might be unnecessary and methodologically undesirable. Additional autonomous minds could introduce uncontrolled variation, ethical complications, and computational expense.

10.3 Games, training, and personal virtual worlds

A simulated world could be built for one user or a small group of participants. Many contemporary virtual environments already distinguish human players from non-player characters, although present-day characters are far from unrestricted human equivalents.

An advanced version could contain a few conscious participants and a much larger apparent society. The apparent population would provide narrative, social challenge, education, or companionship without requiring a conscious perspective behind every character.

10.4 Ethical minimization of suffering

Ethically cautious simulators may avoid instantiating unnecessary suffering. A historically accurate simulation of famine, disease, abuse, warfare, and death would create severe moral concerns if every represented person consciously experienced those events.

Helton and Chalmers both identify this as a possible reason to prefer consciousness-sparse worlds. A simulator might preserve one target consciousness because it serves the purpose of the simulation while replacing other sufferers with nonconscious agents. (Grace Helton)

The ethical conclusion is not straightforward. Deceiving a conscious focal observer about the existence and love of others may itself constitute a serious harm. Ethical simulators might therefore reject singleton worlds for reasons similar to those that lead them to reject mass suffering.

10.5 Consciousness research

A civilization attempting to understand consciousness may create isolated conscious systems precisely because the dependent variable is subjective experience. It could vary cognitive architecture, embodiment, environmental complexity, or social input while limiting the number of subjects.

Such simulations might produce many singleton or small-group observers. They could also involve short conscious intervals, repeated memory states, or experimental branching, giving consciousness occupancy as much importance as population density.

11. Apparent World Size and Conscious Population Are Orthogonal

A natural intuition holds that a simulation of an entire universe would contain many conscious beings, whereas a tiny simulation would contain few. This correlation may hold for bottom-up world-generating simulations, but it is not necessary.

Consider four possible worlds:

Apparent world

Conscious population

One room

Ten conscious participants

One city

One conscious participant

One planet

A thousand conscious participants

One observable universe

One focal consciousness

The apparent universe in the final case might be represented through physical laws, astronomical records, sensorily accessible outputs, and procedural descriptions. It need not contain an independently updated microphysical history for every distant region.

Likewise, apparent human bodies need not correspond to continuously conscious minds. The size of the rendered world and the richness of focal phenomenology can be maintained while conscious occupancy remains extremely low.

This produces a form of ontological sparsity beneath phenomenological abundance. From one perspective, the world is filled with people, activity, culture, and history. From the implementation perspective, subjective experience may occupy only a small portion of the represented social structure.

12. Formal Propositions

Proposition 1: Consciousness-Population Underdetermination

The generic simulation hypothesis does not determine the number or distribution of conscious inhabitants within the simulated world. A population-complete, sparse, singleton, or intermittent architecture can satisfy the basic claim that the focal observer and apparent world are computationally implemented.

This follows because computational implementation of the observer does not logically require identical implementation of every apparent agent. Additional architectural assumptions are needed to infer that apparent population equals conscious population.

Proposition 2: Scale-Density Independence

Apparent spatial or social scale does not, by itself, determine consciousness density. A large apparent universe can be implemented around one focal consciousness, while a small environment can contain many conscious observers.

The apparent extent of a simulated world is a representational property. Conscious population depends on which represented agents instantiate the organization sufficient for experience.

Proposition 3: Interaction-Bounded Adequacy

A singleton or consciousness-sparse simulation does not require nonconscious agents to duplicate human functioning under every possible condition. It requires them to remain behaviorally adequate across the focal observer’s accessible interaction envelope.

This proposition weakens the computational objection based on an exhaustive lookup table. Dynamic models, memory systems, world knowledge, and observer-specific prediction can generate coherent behavior without enumerating every possible encounter in advance.

Proposition 4: Social Attenuation Without Focal Attenuation

A simulation can reduce consciousness density and occupancy while preserving the phenomenological richness of its focal observer. Resources saved by reducing surrounding consciousness can be allocated to the focal cognitive process, environment, or duration.

Declining social consciousness therefore does not entail declining focal consciousness. The loss is located in reciprocal subjectivity rather than in the vividness of the focal observer’s experience.

Proposition 5: Solipsistic Credal Transfer

If singleton simulations are conditionally more probable under (S) than under (\neg S), increasing credence in the generic simulation hypothesis increases credence in singleton simulation, all else being equal.

This result follows from Equation 7. The magnitude of the transfer depends on (P(L\mid S,E)), which must be estimated from architecture, cost, motive, consciousness theory, and observer-selection assumptions.

Proposition 6: Cost-Adjusted Observer Weighting

Under a fixed simulation budget, the anthropic weight of an architecture is proportional to the number of relevant conscious observer-moments it produces per unit cost, multiplied by the resources allocated to that architecture.

Consequently, Chalmers’s solipsistic sim blocker is strong when densely conscious worlds exhibit population economies of scale. It weakens when observer-centered singleton worlds are sufficiently inexpensive per conscious observer or receive disproportionately large resource allocations.

Proposition 7: Social Implementation Opacity

A focal observer’s access to the behavior and apparent physical structure of another agent may fail to reveal the host-level architecture implementing that agent. Apparent biological or neural similarity therefore provides less decisive evidence of consciousness under architectures that permit hidden implementation asymmetry.

The proposition does not imply that behavioral and neural evidence becomes worthless. It states that its interpretation becomes conditional on whether the simulator preserves implementation symmetry.

13. The Socially Sanitized Simulation Hypothesis

13.1 Substrate change without social loss

Popular simulation discussions often present an emotionally sanitized version of the hypothesis. The world becomes computational, but its socially important contents remain untouched. Friends still possess private experiences, parents and partners still love from the inside, and every person continues to inhabit a first-person life.

This conception can make simulation theory feel less radical than traditional skeptical scenarios. The hypothesis appears to replace atoms with code while preserving truth, value, agency, and relationship.

Helton’s argument disrupts this reassurance. Knowledge that a simulated tree is structurally real does little to establish that an apparent friend possesses experience. Chalmers similarly acknowledges that a world in which others lack consciousness would be very bad for many people, even if simulated physical objects remain real. (Grace Helton)

The simulated-world question therefore divides into two:

[
\text{Is the environment real within the simulation?}
]

and

[
\text{Is the environment socially inhabited from multiple first-person perspectives?}
]

A positive answer to the first does not guarantee a positive answer to the second.

13.2 Affective revision and rational revision

Making singleton architectures salient may cause some simulation believers to feel less comfortable with the hypothesis or lower the probability they assign to it. This psychological response would not constitute evidence that simulation is false.

The response could reveal that their original belief concerned a narrower proposition than they realized. They may have been endorsing a population-complete shared simulation while describing their belief as generic simulation.

Affective discomfort can therefore have a diagnostic role. It helps disclose which features of the imagined scenario were tacitly held constant. It cannot determine whether those features are actually present.

13.3 Simulation belief as a bundle

An ordinary statement such as “I think we are probably living in a simulation” may contain several bundled assumptions:

[
S
+
\text{population completeness}
+
\text{continuous history}
+
\text{shared social reality}
+
\text{uniform physical laws}
+
\text{no targeted deception}.
]

The speaker may have arguments for (S) while possessing no independent arguments for the remaining components. Unbundling them permits a more accurate statement of belief.

A person might conclude:

I assign significant probability to a complete world-generating simulation, but very little probability to an observer-centered or singleton simulation.

That position is coherent. It is also substantially more informative than generic simulation belief.

14. Objections and Replies

14.1 Humanlike behavior may require consciousness

One objection holds that any system capable of sustaining lifelong humanlike relationships, flexible reasoning, emotional responsiveness, and self-reflection would itself be conscious. On this view, convincing nonconscious agents may be impossible.

If true, this would sharply reduce the feasibility of consciousness-sparse simulations. A simulator that produced sufficiently capable social agents would inadvertently create additional minds.

The objection does not eliminate all sparse architectures. Peripheral agents could be less behaviorally complete, behavior could be generated by centralized systems without stable individual perspectives, or social detail could be restricted to the focal observer’s actual trajectory. More importantly, whether consciousness necessarily accompanies the required architecture remains unresolved.

14.2 A centralized social model might itself be conscious

A simulation could generate thousands of apparent people through one large artificial system. Even if the individual characters lack separate consciousness, the centralized model might possess its own consciousness.

This possibility complicates population counting. The world could contain the focal observer plus one distributed artificial mind that controls the apparent society, rather than one focal observer alone.

Consciousness density should therefore ultimately track subjects of experience rather than software modules or visible characters. The relationship among computational agents, unified subjects, and fragmented or collective consciousness remains theory-dependent.

14.3 Full simulations may be much cheaper per observer

Shared physical and social environments may produce overwhelming economies of scale. Once a world is simulated, adding another conscious brain may cost comparatively little. One populated world could then generate more observer-moments per unit of computation than any collection of singleton worlds.

This objection is plausible and may be correct. It supports a low value for (P(L\mid S,E)), but it does not justify setting that value to zero without a cost model.

The paper’s claim is not that singleton worlds dominate. It is that their anthropic significance depends on relative cost and allocation, variables omitted by simple counts of worlds or people.

14.4 Ethical simulators would not create deceptive singleton worlds

A simulator concerned about suffering may avoid creating billions of conscious people but also refuse to deceive one focal observer into loving nonconscious constructs. A socially isolated truth may be preferable to a fabricated social life.

This ethical objection reduces the probability of singleton simulations under some moral theories. Other objectives, institutional failures, experiments, entertainment practices, or less benevolent simulators could yield different choices.

Bostrom’s original argument already depends on uncertain posthuman motives. Introducing consciousness distribution makes the motivational uncertainty more explicit rather than creating it for the first time.

14.5 Our social world does not appear observer-centered

Events occur that are inconvenient, incomprehensible, or unrelated to any particular observer. Other people resist our wishes, possess knowledge we lack, surprise us, and participate in vast social systems beyond our awareness.

These observations strongly favor ordinary autonomous agents over crude personalized constructions. They do not exclude a sophisticated observer-centered simulation capable of generating apparent independence.

The evidential question is comparative. A population-complete world may provide the simpler explanation of social complexity, especially when there is no independent evidence for simulation or targeted design.

14.6 Observer selection is controversial

Anthropic reasoning depends on disputed choices about reference classes, observer-moments, self-sampling, and self-indication. A different anthropic framework may assign different weights to singleton and populated worlds.

The minimal solipsistic burden does not require a complete solution to observer selection. It requires only that generic simulation belief be partitioned across compatible architectures rather than equated with population completeness.

The stronger claim that singleton worlds are probable would require a defended observer-selection theory. This paper does not make that claim.

14.7 The term solipsistic is misleading

A singleton simulation contains minds outside the apparent world, including the simulators. Calling it solipsistic may therefore suggest a stronger metaphysical thesis than intended.

The term is retained because it is established in the recent literature and captures the focal observer’s local epistemic condition. The more precise description is a locally solipsistic, singleton-conscious, or socially uninhabited simulation.

14.8 The argument could encourage unhealthy suspicion

Philosophical discussion of singleton simulations could be misread as practical reason to distrust friends, relatives, or strangers. Nothing in the argument supplies such evidence.

The ordinary evidential and moral presumption should remain that other people are conscious. The paper analyzes the internal structure of a speculative hypothesis. It does not recommend treating anyone as an automaton or interpreting unusual experiences as signs of simulation.

15. An Empirical Research Program

15.1 Measuring the hidden architecture prior

People who endorse simulation theory could be asked to distribute their conditional credence across several architectures:

• complete bottom-up universe simulation;
• full-population ancestor simulation;
• consciousness-sparse simulation;
• small-group simulation;
• singleton simulation;
• intermittent simulation;
• uncertainty about architecture.

This would reveal whether reported belief in simulation actually refers to the generic hypothesis. It would also provide the first direct estimate of the architecture assumptions carried by simulation believers.

15.2 The Social Completeness Framing Effect

An experiment could randomly assign participants to different descriptions.

Generic framing: An advanced intelligence is simulating our world.

Population-complete framing: Every person is implemented as a conscious simulated mind.

Population-unspecified framing: Some apparent people may be conscious and others may be behavioral constructs.

Singleton framing: Only the participant may be conscious within the apparent world.

Participants could then report:

• probability that the scenario is true;
• perceived scientific plausibility;
• emotional response;
• expected meaningfulness of life;
• confidence that relationships remain genuine;
• willingness to describe the scenario as “the simulation hypothesis.”

A likely prediction is that generic framing will initially be interpreted more like population-complete framing than population-unspecified framing. Making consciousness allocation explicit may reduce endorsement or increase uncertainty among some participants.

This Social Completeness Framing Effect would not show that simulation belief is irrational. It would show that informal descriptions recruit tacit assumptions about other minds.

15.3 Computational studies of observer-relative social fidelity

Artificial-agent research could estimate how computational cost changes as agents are required to pass increasingly demanding social tests. Relevant manipulations could include:

• length of relationship history;
• number of focal observers;
• frequency of unexpected questions;
• consistency across separate social contexts;
• depth of autobiographical memory;
• independent off-screen activity;
• resistance to adversarial testing;
• neural or physiological observability.

The resulting cost curves would not determine whether the agents are conscious. They would clarify how expensive observer-relative behavioral fidelity becomes compared with fuller cognitive simulation.

15.4 Population cost scaling

Future work could model:

[
C(n,m,T)
]

for worlds containing different conscious populations. The model should separate:

• shared environmental cost;
• cost per conscious cognitive architecture;
• cost per nonconscious social agent;
• inter-agent interaction cost;
• private-history cost;
• error correction and consistency;
• simulation speed;
• observer-centered compression.

The principal quantity would be conscious observer-moments per unit of host computation. This would turn the debate between singleton efficiency and Chalmers’s observer-number blocker into a tractable, if highly speculative, computational problem.

15.5 Consciousness theory as the decisive variable

The greatest uncertainty is the cost of consciousness itself. If any flexible, socially competent agent is conscious, sparse simulations may be difficult to construct. If consciousness requires a particular recurrent, integrated, or intrinsically causal architecture that can be omitted while preserving observer-relative behavior, sparse simulations become more feasible.

Artificial-consciousness research therefore bears directly on the solipsistic burden. It may reveal whether behavior and consciousness can be separated, how much additional organization consciousness requires, and whether one system can generate many apparent personalities without creating many subjects.

16. Discussion

The possibility of a simulated world is commonly discussed as though simulation determines ontology while leaving population metaphysics unchanged. Objects become virtual objects, bodies become simulated bodies, and brains become computational structures, but every apparent human remains a subject of experience.

That assumption is natural in a complete ancestor simulation. It is not guaranteed across the wider class of simulations that advanced intelligences might create. A world can be generated for one observer, for a selected group, for an experiment, or for a purpose that requires behavior without requiring billions of conscious lives.

The central insight is that apparent social population and conscious population can diverge. The divergence can occur in number, duration, degree, and implementation. A city may contain thousands of apparent biographies but only a few continuous perspectives. A person may be conscious during one interval and represented through compressed records during another. A socially sophisticated agent may possess enough memory and generative capacity to sustain a relationship without duplicating every internal process of a human mind.

Consciousness density and occupancy make this design space explicit. They replace the binary contrast between a normal shared world and absolute solipsism with a continuum of social implementation:

[
\delta=1
\rightarrow
\delta=0.5
\rightarrow
\delta=0.01
\rightarrow
\delta=\frac{1}{N}.
]

This continuum matters anthropically. Chalmers is correct that a world containing one billion conscious observers carries more observer weight than a singleton world when the numbers of worlds are fixed. The numbers of worlds, however, depend partly on cost and motive. A complete observer model must compare the number of conscious observer-moments produced per unit of computational investment.

The continuum also matters existentially. A simulated life may contain real objects and accurate nonsocial knowledge while lacking the reciprocal consciousness that gives relationships much of their value. The important distinction is not simply between physical and virtual reality. It is between a world that is socially populated from within and one that is merely populated in appearance.

Simulation believers need not conclude that they are alone. They should recognize that their belief contains an architecture question. A high probability assigned to generic simulation cannot automatically be transferred to a reassuring population-complete version.

The appropriate response is neither paranoia nor dismissal. It is conceptual precision. One should distinguish the probability of being simulated from the conditional probability that a simulated world contains a complete community of minds.

17. Conclusion

The generic simulation hypothesis does not specify how consciousness is distributed across the apparent inhabitants of a simulated world. Ancestor simulations may contain billions of conscious people, but selective simulations may contain a small focal group, one observer, intermittent minds, or a mixture of conscious and nonconscious agents.

Bostrom briefly recognized the possibility of single-person simulations and shadow-people. Helton developed the threat that ethical or resource-conscious simulators might instantiate only some sentient creatures. Chalmers accepted the coherence of the scenario while showing that populated simulations can dominate observer counts because they contain more conscious beings.

The framework developed here extends these arguments by treating consciousness distribution as a variable. Consciousness density measures the proportion of apparent agents who are conscious, while consciousness occupancy measures the proportion of apparent agent-time accompanied by experience. Observer-centered simulations can preserve a focal consciousness and its apparent social environment while reducing the number, duration, or implementation depth of other minds.

Such simulations need not rely on strict philosophical zombies. Their surrounding agents need only remain behaviorally adequate within the focal observer’s interaction envelope. This weakens objections based on the difficulty of reproducing unrestricted human behavior without consciousness, although it does not establish that advanced nonconscious social agents are possible.

The anthropic significance of singleton worlds depends on cost-adjusted observer production. Densely conscious simulations dominate when shared environments and uniform laws create economies of population scale. Singleton worlds gain importance when observer-centered compression makes them cheaper per conscious observer or when simulator motives direct more resources toward individualized worlds.

These uncertainties yield the solipsistic burden of simulation belief:

Anyone assigning substantial credence to an unspecified simulation hypothesis must assign some conditional credence to consciousness-sparse and singleton worlds, unless additional arguments exclude those architectures.

This is not evidence that we are simulated, and it is not evidence that other people lack consciousness. It is a consequence of taking the generic hypothesis seriously enough to separate it from its most socially reassuring interpretation.

A simulated world need not be a shared conscious world. The substrate of experience and the population of experiencers are separate questions.

References

Bostrom, N. (2003). Are you living in a computer simulation? The Philosophical Quarterly, 53(211), 243-255. doi:10.1111/1467-9213.00309. (Simulation Argument)

Chalmers, D. J. (2024). The simulation hypothesis: Metaphysics, epistemology, value. In U. Kriegel (Ed.), Oxford Studies in Philosophy of Mind (Vol. 4, pp. 499-513). Oxford University Press. doi:10.1093/9780198924159.003.0017. (Consciousness Network)

Helton, G. (2023). On being a lonely brain-in-a-vat: Structuralism, solipsism, and the threat from external world skepticism. Analytic Philosophy. doi:10.1111/phib.12291. (Wiley Online Library)

Helton, G. (2024). The simulation hypothesis, social knowledge, and a meaningful life. In U. Kriegel (Ed.), Oxford Studies in Philosophy of Mind (Vol. 4, pp. 447-460). Oxford University Press. (PhilArchive)

Park, J. S., O’Brien, J. C., Cai, C. J., Morris, M. R., Liang, P., and Bernstein, M. S. (2023). Generative agents: Interactive simulacra of human behavior. In Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology. doi:10.1145/3586183.3606763. (arXiv)