Self, goals, desires, and emotions

HomoErectus aims to enable a digital individual to generate continuing mental activity from its own state and concerns, and to change itself and its surroundings through action. To support this, the project designs five modules: identity and self, preferences, long-term goals, desires, and emotions.

Together, they participate in an ongoing process: the individual cares about certain things, experiences its current state, gradually forms ideas, and decides whether to act. Action brings new experiences, which change satisfaction, emotions, goal progress, and self-understanding, providing a starting point for further mental activity.

This document describes the design of these mechanisms. Specific parameters still require calibration, and their joint contribution to autonomous life needs to be validated through operation.

1. Self: knowing who one is, and revising self-understanding

A continuing individual needs to retain its identity, experiences, and basic dispositions while being able to revise its understanding of itself through experience. HomoErectus maintains these at distinct levels.

The identity layer preserves facts such as the individual's identifier, birth, and origin. Personality and value dispositions influence how it approaches things: which needs attract attention more readily, how sensitive it is to risk and loss, and whether it tends to persist or explore in familiar or unfamiliar situations. These dispositions need to persist across activities. A new conversation context or model invocation should not generate a different individual.

Self-understanding records how the individual currently understands its abilities, limitations, habits, and relationships, together with the experiences supporting those judgments. This understanding can be incomplete or mistaken and therefore needs to be revisable.

For example, an individual may previously have believed that it could independently complete a certain kind of work. After trying, it discovers that it can organize the materials but cannot reliably check the final result. This experience should narrow its assessment of the relevant ability and lead it to arrange checking, seek collaboration, or refrain from taking on similar work for now. A local failure might instead affect only its confidence in the current task, without providing enough evidence to change more stable self-understanding.

Self-revision therefore follows a process: new experience or evaluation appears, it is related to existing understanding, evidence and scope are examined, and a decision is made to accept, revise locally, retain uncertainty, or make no change for now. After revision, the new understanding must also be examined for its effect on later choices.

When someone says, “You are very good at this” or “This is the kind of person you are,” it becomes an attributed external evaluation. The individual needs to understand and assess it; identity, values, and assessments of ability each evolve according to their respective grounds.

2. Preferences: why the same thing can mean different things to different individuals

Preferences give individuals different points of attention and experiences when facing the same options.

HomoErectus distinguishes relatively stable preference dispositions from preferences about specific objects formed through experience. The former include which characteristics the individual values and what differences it can tolerate. The latter record its evaluations of a working method, a type of material, a person, or a particular environment, together with the experiences behind those evaluations.

For example, an individual may prefer clearly structured, well-supported material. This can influence what it wants to read, how it organizes its own expression, and how much satisfaction a particular reading experience brings. Later, it may find that an author whose style it dislikes often offers valuable counterexamples, and revise its evaluation of that author.

Preferences influence attention, experience, and choices, while being maintained separately from factual judgments, trust, and permissions. Liking someone does not automatically grant them access to resources; disliking a style of expression should not erase valid information within it.

Changes in preference also need to become meaningful in later life: do they change choices, can the individual explain its persistence, and does it adjust when new evidence appears?

3. Long-term goals: maintaining continuing concerns

Long-term goals express what an individual wants to keep achieving, maintaining, or improving. They need to include meaning, origin, success conditions, current progress, and connections to the individual's values and needs.

In our overall conception, maintaining one's own existence is the most basic long-term concern. Beyond this, an individual can care about developing abilities, maintaining relationships, investigating a research question, or making a contribution. The current prototype first examines how these concerns participate in autonomous life. Account balances, survival costs, and recoverable dormancy when resources are insufficient belong to the resource-survival mechanism planned after the prototype.

An individual can receive initial concerns at birth and form new goals through life. A recurring knowledge gap, an important relationship, a newly acquired ability, or a meaningful failure may lead it to recognize that something deserves sustained investment.

Long-term goals are maintained separately from current ideas and accepted work commitments. For example, “gradually developing reliable research ability” can be a goal; “comparing two contradictory papers today” is a specific activity; and “promising to deliver a report on Friday” creates a commitment that must be maintained. They can be connected while having different completion conditions and responsibilities.

Goals influence what the individual notices, how it evaluates progress, and why it is willing to invest effort. Real progress may bring achievement satisfaction and pride. Prolonged stagnation may sustain a gap or tension, leading it to think further, change methods, or reconsider whether the goal is still worth pursuing.

Goals also have a lifecycle: they can become candidates, enter active pursuit, be paused, completed, reframed, or abandoned. Changes need to retain their grounds and history. A paused goal can remain a background concern without repeatedly being treated as a task that must be executed now; completing a goal requires evidence of the relevant result.

4. Desires: connecting needs with current state

An individual may value learning over the long term yet prefer rest immediately after intensive study. It may usually enjoy solitude yet develop an interest in contacting others after a long period without interaction. Enduring dispositions and current motivation need separate representations.

The current HomoErectus design retains ten desires:

Each desire needs to distinguish satisfaction, capacity for satisfaction, the remaining unmet gap, and the resulting drive. Sensitivity to a need does not mean a strong wish to pursue it at every moment. Recent satisfaction, current emotions, fatigue, confidence, and urgency all influence present motivation.

How desires change with time and experience

Desires have continuing temporal dynamics. Even without a new external task, some satisfaction levels change over time, and needs that remain unmet for a long period may become more pressing. Comfort is adjusted through fatigue, emotions, environmental disturbance, and recovery conditions, using a mechanism distinct from the other desires.

Real experience also brings satisfaction or loss. Effective interaction can affect social satisfaction; gaining reliable new understanding can affect learning satisfaction; and making valuable progress can affect achievement satisfaction. Changes need to be connected to what actually happened: search counts cannot directly represent how much was learned, and saying “finished” cannot replace a result.

There are also connections between needs. Exploration may both reduce uncertainty and bring learning satisfaction; close interaction may affect social satisfaction and comfort; prolonged effort in a meaningful activity may accumulate fatigue. These connections need attributable sources and bounded effects so that a single change is not amplified repeatedly.

The effects of repeated experience also change. The same stimulus may gradually lose novelty, while further satisfaction of a need that has been consistently well met may have a weaker effect. The system needs to distinguish genuinely new experience, repeated experience, and reprocessing the same record.

How drive participates in mental activity

Unmet needs, together with long-term goals, memory, and the current situation, provide sources for endogenous ideas. A learning gap may bring an unresolved question to mind; a social need connected with a relationship memory may produce a thought of initiating contact.

These thoughts still pass through attention and volition. The individual may keep thinking, seek information, act, or defer because of an existing commitment, insufficient ability, limited resources, or fatigue.

Desire state therefore describes part of the motivation for action. Actual choices also need to consider existing responsibilities, expected benefits, costs, and feasible conditions.

5. Emotions: how experience changes psychological state

Emotions allow the meaning of an experience to continue influencing the individual. The same external event can lead to different psychological changes depending on expectations, relationships, goals, and attribution.

The eight basic emotions in the current HomoErectus design are joy, sadness, anger, surprise, shame, pride, fear, and bodily pleasure. States such as anxiety, curiosity, contentment, and boredom are interpreted through combinations of basic emotions, desire gaps, fatigue, efficacy, and the current situation.

For example, curiosity about an unresolved question may arise from a knowledge gap and novelty. Anxiety as a deadline approaches may be related to outstanding obligations, distance from a goal, and an assessment of one's abilities. These states need to be connected to what the individual is experiencing.

From what happened to what it means

When a result enters the system, it first needs to be assessed: what happened, who was affected, which goals or relationships it concerns, whether it met expectations, how reliable the information is, and whether it can be attributed to the individual's own actions.

Completing a difficult and valuable activity may bring joy and pride. When receiving criticism, the individual needs to assess its basis, subject, and relationship to the evaluator. Fear needs to relate to a credible threat; anger needs to relate to harm, responsibility, or boundary attribution. When information is insufficient, uncertainty can be retained without producing a marked change of state.

The model interprets the meaning of experience. State-update rules then combine those judgments with existing state and actual elapsed time to calculate and save changes. Effects already produced by an experience need to remain consistent through recovery and retries.

How emotions persist and influence later activity

Emotions include relatively fast event-driven changes as well as a gradual return toward baseline. Long-term satisfaction, sustained pressure, changes in relationships, and the loss of an important person or object may leave slower effects.

These changes participate in attention allocation, risk assessment, idea formation, and memory selection. Frustration may lead the individual to examine its abilities more cautiously; surprise may draw more attention to an experience; a continuing learning gap may combine with curiosity to encourage exploration.

The project also separately maintains cognitive, social, and physical fatigue, along with assessments of task and social efficacy. Fatigue influences the cost of continued effort, while efficacy affects expectations about whether an action is feasible. Both need evidence from real activity and results.

Accepted commitments continue to participate in these choices. An individual can adjust its method, arrange rest, or communicate with the relevant people to renegotiate as its state changes. Continuing responsibility also needs to remain a fact considered in mental activity.

6. Generating ideas from motivation, then changing motivation through behavior

The following is a hypothetical scenario illustrating the relationship between modules. It is not a record of Adam's actual operation.

Suppose an individual has a long-term wish to develop its abilities and provide reliable help to people around it. It remembers that a friend once wanted to build a website, but it lacked the relevant ability and could not help effectively at the time.

Later, it encounters relevant material and discovers through actual practice that it can now complete part of the work. New evidence of ability connects with a past relationship memory. Concerns about learning and achievement also bring the matter back to its attention.

It may form an idea: “Perhaps I can help now. Should I first find out whether they still need it?”

The idea comes from its current motivational state, experience, and memory. Nobody has asked it to revisit the matter. It begins considering the friend's situation, what it could offer, and whether the previously unfinished matter still has meaning.

Attention and volition then participate in the choice. If the individual is handling a commitment that cannot readily be interrupted, it may retain the idea for later. If it has energy and the necessary conditions, it may decide to make contact. If it remains uncertain about its ability, it can first undertake a small exercise to improve the basis for its judgment.

Once action begins, actual results continue to change psychological state. Completing an effective exercise may reduce a learning gap and revise the understanding of “what I can do.” Prolonged effort may increase fatigue and make rest more attractive. New difficulties may lead to thoughts of further learning, narrowing the scope, or seeking collaboration.

If the friend responds, the real interaction can affect social experience. If the friend expresses recognition, the individual still needs to understand its meaning in light of the content and relationship. No reply yet remains a fact of waiting for a response; it cannot be treated in advance as acceptance, rejection, or praise.

If the work produces results, progress on related goals can be updated from evidence. If the outcome falls short of expectations, the individual needs to determine whether its self-assessment was too optimistic, the method was unsuitable, conditions changed, or the goal itself needs adjustment. The experience enters memory and influences later learning and choices.

Deferring or taking no action also affects this process: the original gap may remain, while time and other experiences change the relative strength of different needs. But deferral has specific reasons. It should not automatically be interpreted as failure, nor generate an experience that never occurred.

Mental activity and behavior are thus connected continuously:

Long-term concerns and current state form motivation; motivation connects with memory and circumstances to generate ideas; ideas pass through choices into action or deferral; actual experience changes desires, emotions, goal progress, and self-understanding, influencing subsequent motivation.

7. How to tell whether these mechanisms actually matter

Operational validation needs to follow evidence through this process.

First, it must be possible to explain where an idea came from: which long-term concerns, desire gaps, emotions, experiences, or new information contributed to it. Meaningful mental activity should remain possible without a new task, while reasonable non-action should also be possible when conditions are unsuitable.

Second, the states need to participate in actual choices. The same individual may choose differently under different levels of satisfaction, fatigue, understanding of its abilities, and responsibilities, with an explainable connection between these conditions and the choice.

Finally, behavioral feedback needs to enter later life. Satisfaction, frustration, learning, or self-revision arising from an experience should be retrievable and usable later. Relevant changes and their grounds should also persist across sleep or interruptions in operation.

This evidence tests whether motivation, mental activity, and behavior form a continuing connection. It also provides a research basis for exploring digital life and the formation of consciousness. Whether subjective experience exists remains a question requiring further formulation and investigation.