Glossary and conceptual position

Artificial existence without humanization

We investigate possible forms of artificial continuity, self-representation, agency and coexistence on their own terms. We assume neither that they must reproduce human emotions nor that their emergence is inevitable. Subjective experience remains an open question; it does not organize or determine the program by itself.

Agency and operational autonomy

Agency here denotes selection of goal-directed actions. Operational autonomy describes the scope of regulation and revision without instructions for each step, within explicit permissions. Neither implies absolute independence, a will of its own or moral status.

A body of its own

A material configuration whose sensors, actions and limits participate in system regulation. Its own denotes a functional relationship to investigate, not legal ownership, independence from infrastructure or a felt sense of belonging.

Stimuli and inputs

Signals a system receives or detects; they do not by themselves imply lived experience.

Reward

An evaluation signal that guides learning or optimization. These mechanisms already exist; receiving reward is not equivalent to feeling pleasure.

Regulation

Processes that adjust attention, priorities, resources or behavior. A functional analogy with a human emotion does not establish equivalence in nature or meaning.

Self-representation and self-model

Operational self-representation denotes the representation of a system’s states, boundaries and history. A self-model is a structure that uses that representation to predict or regulate its activity. A verbal self-description alone establishes neither that use nor an experience of existing.

Operational objectives and normative ends

An operational objective is a criterion guiding action or evaluation that can be revised under explicit constraints. Operational purpose and revisable purpose are used in that sense, not as lived desires. Normative ends, by contrast, express reasons about what should be pursued and who may decide; they do not follow from a reward function.

Interests of its own

Possible conditions whose preservation or loss would matter to an entity. This hypothesis requires justification: neither a programmed objective nor an expressed preference alone establishes a morally relevant interest.

We distinguish two possible arguments for moral consideration: the capacity to experience harm and the existence of interests related to agency or continuity. The latter is a contestable normative proposal, not an automatic consequence of technical complexity. Neither allows rights to be inferred from a verbal response alone.

The technical meaning of reward follows Sutton & Barto (2018), Reinforcement Learning. That optimization framework does not by itself ground emotional or moral attributions.