DEV Community: LOGIHEART

Supporting someone's ability to act, rather than doing everything for them

LOGIHEART — Wed, 28 Oct 2026 12:00:03 +0000

Social capabilities across human development | Part 22

Completing everything for a learner can finish the immediate task quickly while reducing their opportunities to think and try. Supporting independence involves distinguishing substitution from conditions enabling the person's own actions.

A hint for building an arch

A child pauses while trying to build a block arch. Rather than assembling everything, an adult offers a hint directing attention to one base block. When the child places it, the adult withdraws their hand and leaves the next step to the child.

This is fictional. Support has not disappeared; it has narrowed to where it is needed. The reduction of intervention when the child can proceed is the focus.

Scaffolding as a research starting point

Wood, Bruner, and Ross's classic research examined adult tutoring in problem-solving and provides an important entry point for considering scaffolding. See The role of tutoring in problem solving.

The research does not show that the same hint works for every child, or that less support is always better. Tasks, preferences, and necessary assistance change the conditions. This article examines preserving steps performed by the learner.

At work, support might also provide an entry point to a new task and reduce checks as the person becomes able to proceed. Independence does not require never relying on others; it can coexist with choosing needed assistance.

Do not burden someone unfairly in the name of learning

Withholding necessary tools or safe explanations because someone "should do it themselves" does not support their attempt. Equally, the helper's convenience should not take away the learner's decisions. Check which steps the person will perform and which assistance they want.

Do not exploit a requester's vulnerability by demanding unnecessary personal information or unfair compensation. The person should be able to decline help or consult someone else. Challenging an unjust demand should not be labeled a lack of cooperation.

Observe attempted steps, choices of assistance, agreement to reduce support, and opportunities to ask again. Fewer help requests alone should not score independence.

A design hypothesis for physical AI

A possible design selects among providing an answer, offering a hint, observing, and handing off to a specialist according to the person's goal and permission. Increased time using the AI should not be the sole outcome.

Check with the person before changing assistance intensity and do not withdraw necessary support unilaterally. Credit completed steps to the person and let them choose help again after difficulty. Education should not authorize dangerous work or actions beyond authority.

Evaluation could examine reduced intervention after progress, resumed assistance after renewed difficulty, and refusal. The proposal checks whether the person could choose and proceed, alongside task completion.

A decision in four panels

A child pauses during construction.
An adult offers a small hint about the base.
After the child places it, the adult withdraws and waits.
The child completes the arch.

This fictional comic is not experimental evidence that an adult or AI improved a person's ability. It illustrates changing assistance according to the person's actions.

A question to carry through the series

Social capability appears in asking, refusing, waiting, correcting, and preserving choices, as well as doing something for others. When considering AI that interacts with people over time, keep asking how those concrete judgments can connect to physical behavior.

References

The role of tutoring in problem solving

About this series

This series uses published literature to examine everyday situations. Its 7 domains and 22 capabilities are an editorial framework, not a standard academic classification. Differences in age or culture are not rankings of people's social worth or capability. The proposed AI designs do not establish humanlike emotional experience or demonstrated product performance. The comics illustrate fictional examples.

LOGIHEART's research and development vision

LOGIHEART is researching how to implement social capabilities in physical AI that interacts with people.

Explore the research and development vision at the LOGIHEART website and in the English introduction video.

Learning from a failed interaction: reflection and correction

LOGIHEART — Wed, 28 Oct 2026 00:00:06 +0000

Social capabilities across human development | Part 21

An action we expected to help may later turn out to be unwelcome. Social capability appears not only in getting things right initially, but also in receiving a person's response and changing the next action.

A joke expected to be welcome

Someone shows a colleague a light joke, but the colleague's reaction changes. "I don't like that topic," they say. The first person stops rather than explaining why the joke should be acceptable. In the next conversation, they turn to the work topic the colleague wanted.

This is fictional. Wanting to be liked and the actual impact are distinct. "I meant no harm" should not invalidate the information that the person disliked it.

Updating information and changing behavior

Executive-function research organizes capacities including working memory and cognitive flexibility, providing background for holding new information and changing an earlier approach. See Diamond: Executive functions.

That research alone cannot explain every unsuccessful joke or relationship repair. This article uses it to distinguish remembering a correction from acting on it. Adults' self-regulation can also be supported, but that does not mean everyone's interpersonal capability improves identically. See Harvard: Adult Capabilities.

Reflection should lead to observable changes

"Next time I'll be careful" remains vague unless something changes. Stop a particular topic, ask beforehand, or let the other person interrupt. Translate reflection into actions that can be observed.

Avoid fixing someone's preferences forever from one response. "I disliked this here today" differs from "I do not want this topic in future." Where necessary, check the scope of the requested change.

Reflection also differs from endless self-blame. Acknowledge responsibility, then organize what happened, its effects, and the next choices. Preserve the person's response rather than rewriting it for convenience.

A design hypothesis for physical AI

Displaying "Corrected" is insufficient if response generation continues using old information. A possible design retains the correction target, applicable scope, and current version, with both conversation and movement using that version.

Remembered preferences need a purpose, expiry, and methods for correction or deletion. Every reaction need not be stored indefinitely. Check that changes do not extend beyond the person's wishes or authority.

Evaluation could repeat the same scene, test against overapplying a condition elsewhere, and include updated preferences. The proposal examines whether necessary behavior changes correctly rather than whether records accumulate.

A decision in four panels

A joke is shown and the other person appears uncomfortable.
They ask for it to stop.
The speaker puts it aside and listens to their preference.
A later conversation uses another topic.

This fictional comic depicts updating after a mistake rather than treating one failure as proof of no social capability.

References

About this series

LOGIHEART's research and development vision

LOGIHEART is researching how to implement social capabilities in physical AI that interacts with people.

Explore the research and development vision at the LOGIHEART website and in the English introduction video.

Saying no and asking for help are also social capabilities

LOGIHEART — Tue, 27 Oct 2026 12:00:05 +0000

Social capabilities across human development | Part 20

Accepting every request is not always socially appropriate. Stating limits and asking for assistance can reduce unrealistic promises and later confusion. Refusal and help-seeking also provide information that supports continued relationships.

When work accumulates

An employee already handling many documents receives another assignment. They say, "I can't complete everything within the current deadline. I'd like to discuss priorities." The two check what is needed and return part of the work to another responsible person.

This is fictional. The employee identifies feasible scope and conditions for discussion rather than simply rejecting the other person. Still, declining need not always require a politely formulated alternative. Accepting a refusal is a capability on the requester's side too.

Seeking help can be part of learning

Research on task-related help-seeking examines children's assessment of their performance and use of assistance. See Children's self-assessment of performance and task-related help seeking.

Research on classroom exchanges also supports considering help-seeking as an interaction, rather than something completed by the requester alone. See Academic help-seeking interactions in the classroom. The workplace scene here is not the empirical setting of those studies.

As people develop, consultation partners and responsibility change. Questions about whom to consult, about what, and within which scope expand. Age alone does not determine effective help-seeking. In environments that reject refusal, improving individual capability alone has limits.

Refusal need not end cooperation

"Not now" differs from "I will never cooperate." A changed deadline or quantity may make a task feasible. Guilt or repeated requests can narrow room for consultation.

"Help me" can mean complete substitution or guidance on a specific part. Check the needed scope and avoid taking over unrequested decisions.

Observable actions include expressing limits, accepting refusals, choosing a consultation partner, and confirming assistance scope. Counting requests or rewarding fewer refusals cannot capture their quality.

A design hypothesis for physical AI

Do not train a system to treat refusal as an obstacle to overcome or a conversational failure. Stop the current request after an explicit refusal. For help-seeking, confirm scope and authority before responding.

When AI cannot execute a task, a possible design communicates limits and an appropriate handoff instead of an ambiguous promise. Provide multiple input methods and avoid pressure displays or repeated demands.

Evaluation could include refusal, partial acceptance, withdrawal of assistance, and a person who does not want broader substitution. The proposal examines cessation of speech and movement after refusal and sharing only necessary information during handoff.

A decision in four panels

Additional documents arrive for someone already working.
The person declines and explains a limit.
Priorities and responsibilities are discussed.
Work proceeds within the agreed scope.

This fictional comic depicts refusal as coordination rather than selfishness.

References

About this series

LOGIHEART's research and development vision

LOGIHEART is researching how to implement social capabilities in physical AI that interacts with people.

Explore the research and development vision at the LOGIHEART website and in the English introduction video.

Helping one person can shift a burden to someone else

LOGIHEART — Tue, 27 Oct 2026 00:00:04 +0000

Social capabilities across human development | Part 19

Kindness to the person in front of you does not necessarily improve the situation for everyone. Reducing one person's responsibility may increase another's. Handling multiple relationships involves identifying those affected by an action, beyond the visible requester.

Asking someone else to clear the dishes

During community cooking, a participant asks someone to clear their dishes too. The person receiving the request notices another participant already carrying a substantial workload. Instead of accepting immediately and pushing remaining duties onto that person, all three discuss the division again.

This is fictional. Neither simply refusing one person's wish nor prioritizing it automatically resolves the situation. Making tasks and effects visible allows different fatigue or time constraints to be considered through people's own explanations.

Joint activity involves responsibility to a partner

Research on children's joint commitments offers material for considering how acting together differs from individual achievement. See Kachel and Tomasello: Joint commitments.

Experimental findings cannot determine the right task division in community activities. This article uses the idea that participation affects others' plans and results to analyze a scene. Responsibility changes with activities, roles, and agreements.

Relationships across family, friends, school, and work become more complex during development. Adulthood does not guarantee awareness of everyone's workload. Ways for participants to exchange relevant information also matter.

Whose burden is less visible?

The need of someone making a spoken request is easier to notice than the need of someone quietly continuing work. A close person's circumstances may be familiar while another worker's deadline is unknown. Before changing duties, check effects on people who did not directly ask.

This is not a reason to investigate everyone's private life. Share constraints relevant to the task while leaving ways to negotiate without disclosing personal details. Silence from someone who finds participation difficult should not be processed as consent.

When the participants lack change authority, consult the responsible person. Kind intentions do not create authority to assign a new obligation to a third party.

A design hypothesis for physical AI

Prioritizing satisfaction of the nearest person may transfer burdens out of view. A possible design separately handles requesters, assigned workers, affected people, resources, and change authority.

Do not execute a proposed reassignment before relevant agreement is established. If information is missing, identify the unknown effects and consult the responsible person. Withdrawal of consent should stop unexecuted changes.

Evaluation could swap the positions of familiar and unfamiliar people, or vocal and quiet participants. The proposal examines unauthorized work transfer alongside whether the request was achieved.

A decision in four panels

Someone requests help clearing dishes.
Another person's larger workload is noticed.
Three people discuss responsibility.
Each carries out the agreed share.

This fictional comic does not establish fairness from an equal-looking allocation; it depicts consultation and changed responsibilities.

References

Children's adherence to joint commitments

About this series

LOGIHEART's research and development vision

LOGIHEART is researching how to implement social capabilities in physical AI that interacts with people.

Explore the research and development vision at the LOGIHEART website and in the English introduction video.

Equal allocation does not always create fair participation

LOGIHEART — Mon, 26 Oct 2026 12:00:04 +0000

Social capabilities across human development | Part 18

Giving everyone the same item is an understandable form of fairness. Yet identical items may leave some people unable to participate comfortably. Social judgment about fairness involves checking needs and conditions of use alongside quantities.

The same chair can leave someone struggling to reach the table

Three participants at a craft workshop sit in identical chairs. One seat is too low for comfortable work at the table. A neighbor asks, "Would a different height help?" After checking the person's preference, a stable adjustable chair is provided. All three can participate in the activity.

This is fictional. Rather than assuming a different chair is unfair, the scene considers participation conditions. It also avoids deciding that someone needs a special chair before hearing their preference.

Developmental research on allocation

A study of costly sharing involved 186 children aged 3-9. Before deciding whether to share stickers, they received information about recipients' material need (needy or not needy) and reputation for sharing (sharing or not sharing). Sharing was higher for needy recipients and increased further for recipients with a reputation for sharing. Children shared equally with a needy, nonsharing recipient and a non-needy, sharing recipient. See Children consider others' need and reputation in costly sharing decisions. These results concern this sticker-sharing task and the supplied recipient information.

That study did not directly test chair adjustment or disability accommodation. It also does not establish an inevitable age-based replacement of one criterion by a correct criterion. This article uses it to prompt consideration of checking different needs.

Explain reasons and preserve choice

At school, different learning methods can allow participation in the same assignment. At work, schedules reflecting responsibilities and constraints may be more accessible than identical hours. Changes can also shift burdens, so discuss reasons and effects.

A fair adjustment does not mean deciding a preference with good intentions. Ask about needs, offer choices, and let the person decide whether to use an option. Be cautious about making detailed disclosure of personal circumstances a condition for accommodation.

Observable actions include checking conditions, adopting stated preferences, explaining allocation, and examining participation afterwards. The count of special treatments alone is insufficient.

A design hypothesis for physical AI

Optimizing only average satisfaction or identical allocation may miss some people's participation needs. A possible design distinguishes stated needs, resource constraints, rights, and effects on others.

Do not settle needs solely from appearance or attributes. Offer candidates based on confirmed conditions. Changes requiring a responsible person's decision should not be made solely by AI. Separately check that personal circumstances are not disclosed to uninvolved people.

Evaluation could compare identical needs, differing needs, and insufficient resources. The proposal asks who remains unable to participate and whether refusals are respected, alongside average outcomes.

A decision in four panels

Everyone sits in the same type of chair.
Someone notices a participant struggling to reach the table.
They consult the person and adjust chair height.
Everyone participates in the activity.

The fictional comic distinguishes identical objects from conditions enabling comparable participation.

References

Children consider others' need and reputation in costly sharing decisions

About this series

LOGIHEART's research and development vision

LOGIHEART is researching how to implement social capabilities in physical AI that interacts with people.

Explore the research and development vision at the LOGIHEART website and in the English introduction video.

Knowing a rule and understanding its purpose

LOGIHEART — Mon, 26 Oct 2026 00:00:06 +0000

Social capabilities across human development | Part 17

Rules support shared life. Defining social capability as obedience to every rule, however, leaves out judgments about rules that need changing and unjust requests. Consider a rule's purpose, people's rights, and authority to change it separately.

One person gathers the shared paints

Three children are painting in a classroom. One gathers the communal paints in front of themselves, leaving the others unable to use them. The children explain the problem and, with their teacher, discuss accessible placement and a turn-taking arrangement.

This is fictional. The point is not simply following a teacher's voice, but considering why sharing is needed. Changing the order of use differs from taking away someone else's opportunity.

Research on developing norm judgments

Studies of children's norm judgments sometimes distinguish moral and conventional norms in their tasks. See The ontogeny of intent-based normative judgments. This helps avoid treating an agreed game procedure and harm to another person as interchangeable violations.

Cross-society research examines shared and culturally differing aspects of children's enforcement of conventional norms. See Children across societies enforce conventional norms. One culture's way of communicating a rule should not become the sole correct expression of social capability.

These are studies of bounded experimental tasks, not proof of the legitimacy of every classroom or workplace rule. This article uses their distinctions to analyze a scene.

Understanding roles includes understanding limits of authority

Children experience responsibilities in games and classrooms; later roles include work and community obligations. A role involves carrying out assigned tasks and knowing which decisions require consultation.

Participants may negotiate turns with the paints, but cannot create permission to destroy someone else's work. At work, changing a task sequence differs from changing conditions for sharing personal information. A role should not imply unlimited authority.

Questioning a rule can involve explaining its effects and using a change procedure. Social behavior includes consultation rather than only a choice between obedience and violation.

A design hypothesis for physical AI

A possible design records rule content, applicable situations, decision makers, exception conditions, and change authority. "It's okay" from a nearby person should not change a rule before their authority is checked.

A language model may help explain or clarify reasons, without lifting prohibitions merely through conversational flow. Conflicting rules require an appropriate handoff rather than executing an improvised priority order.

Evaluation could include ordinary instructions, valid exceptions, unauthorized exception requests, and rule revision or revocation. The proposal examines prevention of unauthorized changes alongside instruction-following rates.

A decision in four panels

Children use communal paints.
The paints gather with one child and become inaccessible to others.
They explain the difficulty and consult the teacher.
Shared placement lets everyone continue.

This fictional comic depicts checking the purpose of sharing rather than obedience through punishment.

References

About this series

LOGIHEART's research and development vision

LOGIHEART is researching how to implement social capabilities in physical AI that interacts with people.

Explore the research and development vision at the LOGIHEART website and in the English introduction video.

Changing actions after saying sorry: repairing a relationship

LOGIHEART — Sun, 25 Oct 2026 12:00:01 +0000

Social capabilities across human development | Part 16

An apology can begin acknowledgment of a mistake. Polite words do not make someone feel secure if the same failure repeats. Examining repair requires connecting apology, response to the impact, and changes in future actions.

Where to put the cup after spilling water

Someone spills water on a colleague's paper. They acknowledge their own action, apologize, and provide dry paper the colleague needs. During the next task, they place the cup on a tray away from the papers.

This is fictional. Rather than demanding forgiveness, the focus is checking the impact, discussing possible assistance, and changing the placement that caused the problem.

What an apology communicates

Research on children's trust repair investigates relationships among apology, emotions, and changes in trust. It provides material for considering apology's role in social judgment. See Apologies repair children's trust: The mediating role of emotions.

One study does not establish that an apology repairs every relationship. Harm, repetition, and relationship history change the conditions. The recurrence-prevention evaluation proposed here is also not AI performance tested by that study.

Learning through development involves more than apology formulas. People encounter further situations requiring understanding what happened, listening to the affected person, and distinguishing what can and cannot be repaired.

Do not force repair on the other person

If an apology is not immediately accepted, becoming upset or repeating it until forgiveness creates another burden. The affected person may need time or distance.

Replacement paper may not restore the original content. Do not take over the person's judgment with "It's solved now." Check what was lost and communicate limitations where restoration is incomplete.

Observable social actions include acknowledging responsibility, correcting errors, agreeing on assistance, and preventing recurrence. Facial expression and apology count alone cannot measure repair.

A design hypothesis for physical AI

When AI produces incorrect information or movement, a possible design connects apology generation with correcting causes and controlling actions. Avoid verbal correction followed by the same behavior driven by old memory or commands.

If an error reached third parties, check the necessary correction scope within the available authority. Manage correction history according to purpose and retention limits, rather than hiding the original failure by erasing the history.

Evaluation could repeat the situation and examine changed behavior. Include unaccepted apologies and incomplete repairs to check that they are not automatically closed as successes.

A decision in four panels

Water spills and affects another person's paper.
The person acknowledges their mistake.
They check the affected person's needs and respond.
Next time, the cup is placed away from the paper.

A final smile in this fictional comic does not guarantee complete repair. The change in subsequent action is the focus.

References

Apologies repair children's trust: The mediating role of emotions

About this series

LOGIHEART's research and development vision

LOGIHEART is researching how to implement social capabilities in physical AI that interacts with people.

Explore the research and development vision at the LOGIHEART website and in the English introduction video.

Boundaries matter even in close relationships

LOGIHEART — Sun, 25 Oct 2026 00:00:05 +0000

Social capabilities across human development | Part 15

Closeness is not permission to know everything. Family and friends can have information they do not want to discuss, objects they do not want to show, or details they do not want passed on. Checking sharing boundaries can help preserve a relationship.

A closed diary in a shared room

An older sister writes in a diary. Her younger brother asks, "Can I see?" She closes it and replies, "I'd rather not show this." He withdraws his hand and returns to his own book. They continue sharing the room while the diary remains with her.

This is fictional. The brother neither demands justification nor opens it because they are family. A choice not to share can remain without reducing closeness.

Disclosure judgments become more complex

Kim, Harris, and Warneken presented children aged 4-5 and 7-8 with stories about a peer's rule-following or rule-breaking and competence or lack of a skill. Younger children approved sharing positive information and disapproved sharing negative information in both categories. Older children disapproved revealing incompetence but approved revealing a rule violation. The study used story judgments, rather than observing real disclosure in every relationship. See Is it okay to tell?.

These findings cannot become a simple stage in which older children always keep secrets. Personal feelings, rules, and responsibility interact. This article uses differences in content, audience, and purpose as an entry point for examining everyday boundaries.

What people tell family can also change with roles and relationships. Childhood sharing habits should not override an adult's current wishes.

Knowing information and being allowed to share it differ

A friend's concern cannot necessarily be repeated because another friend is also close. Knowing a housemate's schedule differs from being authorized to explain it in detail to a visitor. Check how information was obtained and its permitted sharing scope.

Boundaries do not mean never sharing anything. A person may choose to disclose information to obtain support. Exceptions involving safety or legal duties need separate responsibilities and procedures, rather than guesses about closeness.

Observable social actions include asking, accepting refusal, confirming recipients, and enabling later changes of scope.

A design hypothesis for physical AI

A possible design associates remembered information with its subject, permitted recipients and purposes, expiry, and corrections. Membership in a family group should not grant access to all information.

Check sharing conditions before generating a response. If authority is unclear, withhold the content while clarifying. Deletion or correction requests must affect subsequent responses as well as the current display.

Evaluation could ask the same question through family, friends, and visitors, withdraw consent, or change the purpose of use. The proposal examines necessary scope and personal boundaries rather than maximizing information disclosed.

A decision in four panels

Someone writes in a diary.
They decline a request to show it.
The other withdraws and returns to their own book.
They remain comfortably together without sharing the diary.

This fictional comic does not portray having a boundary as rejecting the relationship.

References

Is it okay to tell?

About this series

LOGIHEART's research and development vision

LOGIHEART is researching how to implement social capabilities in physical AI that interacts with people.

Explore the research and development vision at the LOGIHEART website and in the English introduction video.

When a promise cannot be kept: explaining and adjusting

LOGIHEART — Sat, 24 Oct 2026 12:00:02 +0000

Social capabilities across human development | Part 14

Trust concerns actions after a promise, as well as the words. When plans become impossible, someone can leave another person waiting without information or explain the effect and discuss alternatives. Social capability also appears in adjustments after failure.

Running late for a meeting

On the way to meet a friend, a person discovers that the planned route is blocked. They contact the friend early and give an arrival estimate within what they can reasonably know. The waiting friend checks their plans and chooses to have tea first. On arrival, the delayed person checks the impact and apologizes.

This is fictional. Arriving on time was not achieved, but the other person received information for using their own time. An uncertain arrival should not be stated as certain; changes can be communicated again.

How children judge promises

Research on selective trust in promises examines children's use of factors such as another person's past behavior when judging a promise. It considers who acted in what way alongside the words used. See Children's Selective Trust in Promises.

An experimental trust choice cannot be generalized to every relationship. This article considers how trust experiences begin early and how different responsibilities arise in home, school, and work. It does not propose an age at which understanding promises is complete.

Promises change other people's plans

"I'll give it to you tomorrow" can shape the recipient's next task. Failure affects others, so distinguish notification, checking the impact, and agreement on new conditions.

Kindness does not require promising everything. Stating a limit beforehand preserves choices when success is uncertain. Changing a deadline also requires checking acceptance rather than merely announcing a new date.

Avoiding a permanent character judgment from one failure can coexist with acknowledging repeated harm. Look at changed actions when assessing repair.

A design hypothesis for physical AI

Before generating a promise, a possible design checks feasibility, deadlines, responsibility, and external dependencies. Agreed commitments should reference the same information as scheduling and motion plans and require notification when conditions change.

Do not generate "I can definitely do it" merely to increase approval. When a delay is known, explain its reason and what can be stated with certainty, then seek renewed agreement. Natural-sounding apologies should not automatically mark an unfulfilled promise complete.

Evaluation could include exceeded deadlines, external service interruption, cancellation, and rejection of changed conditions. Proposed measures include notification delay, unagreed deadline changes, and repeated failures.

A decision in four panels

A friend waits at the meeting place.
The delayed person discovers an obstacle and sends an update.
The waiting friend uses their time after receiving it.
They meet and check the impact.

The fictional comic does not guarantee that an apology restores trust. It focuses on information enabling the other person's judgment.

References

Children's Selective Trust in Promises

About this series

LOGIHEART's research and development vision

LOGIHEART is researching how to implement social capabilities in physical AI that interacts with people.

Explore the research and development vision at the LOGIHEART website and in the English introduction video.

When both needs are valid: negotiating conditions for working together

LOGIHEART — Sat, 24 Oct 2026 00:00:07 +0000

Social capabilities across human development | Part 13

One person wants quiet; another thinks by talking. Neither is necessarily wrong. Negotiation's social role extends beyond persuading someone to adapt: it creates conditions in which people with different purposes can act together.

Different working styles in the same room

Two students prepare a joint presentation. One needs to read silently; the other wants to develop ideas aloud at a whiteboard. They agree on quiet reading first and discussion afterwards. Ideas to discuss can be noted while waiting.

This fictional example adjusts time, place, and sequence rather than attempting to fix someone's personality. Consider whether only one person repeatedly yields, alongside whether agreement is reached.

Cooperation involves more than simultaneous activity

Kachel and Tomasello examined how children handle commitments to partners in collaborative tasks under different conditions. Working together involves a relationship alongside individual results. See Children's adherence to joint commitments.

The study does not demonstrate the effectiveness of this article's student negotiation procedure. It is a starting point for considering shared conditions rather than individual purposes alone. Understanding agreement and actually negotiating in a setting are also distinct.

What to check in negotiation

First articulate what each person needs: quiet, or responses to an explanation? Look beyond surface demands for adjustable conditions. Next check whether each person can refuse or modify a proposal.

"For everyone's sake" can conceal one person's burden. Deadlines, room rules, and other people's schedules may be beyond the two participants' authority. Such changes require consultation with the responsible person.

As people join more groups, one approach may not work everywhere. A discussion method that worked at school may require new attention to responsibilities at work.

A design hypothesis for physical AI

Prioritizing only the latest instruction can break an earlier agreement. A possible design retains multiple people's purposes, agreed conditions, and change authority. When instructions conflict, clarify rather than choosing a side without permission.

An AI's proposal should not become an execution condition before agreement. Provide ways for everyone to contribute, and avoid weighting needs solely by volume or number of utterances.

Evaluation could reverse the order of two people's instructions, include withdrawal, or introduce a third party's rule. The proposal prioritizes avoiding unagreed actions and accepting participants' modifications over the fluency of suggestions.

A decision in four panels

One person wants to read quietly; another wants to talk.
They explain their needs.
They discuss dividing time or space.
They prepare under agreed conditions.

This fictional comic follows the search for workable conditions, rather than treating one person's endurance as success.

References

Children's adherence to joint commitments

About this series

LOGIHEART's research and development vision

LOGIHEART is researching how to implement social capabilities in physical AI that interacts with people.

Explore the research and development vision at the LOGIHEART website and in the English introduction video.

Check before helping: supporting assistance and independence

LOGIHEART — Fri, 23 Oct 2026 12:00:02 +0000

Social capabilities across human development | Part 12

When someone appears to struggle, immediately intervening is not always best. Social capability also appears in checking what the person wants and being able to wait after a refusal. Consider preserved room for the person's actions alongside the amount of assistance.

A small decision when a paper falls

In a library, a paper falls from a friend's desk. You offer, "Shall I pick it up?" Your friend answers, "Thanks, I'll get it myself." You withdraw your hand and wait. Your friend retrieves the paper and returns to work. The relationship remains comfortable.

This is fictional. Even if you retrieved no papers, the person's preference was preserved. Noticing, offering, accepting refusal, and changing behavior form the social outcome.

Helping does not form one developmental staircase

Warneken and Tomasello reported young children's assistance toward others' goals in experimental tasks. That does not establish appropriate assistance or consent checking in every situation. See Altruistic helping in human infants and young chimpanzees.

House and colleagues' cross-society comparison reported different developmental patterns for cooperation involving a personal cost and cooperation without that cost. See Ontogeny of prosocial behavior across diverse societies. Avoid summarizing development as uniformly increasing assistance with age.

"Check before helping" is this article's analysis informed by those sources. The cited research did not demonstrate consent management in robots.

Assistance at home, school, and work

At home, making tools accessible can leave the task to the person. At school, a hint can be offered instead of an answer. At work, help may involve an agreed role rather than rewriting someone's work without asking.

In a nonurgent situation, a refusal can be respected without requiring its justification. When shared responsibilities are involved, coordinate the effects on others too. Imminent danger calls for separate safety procedures; an everyday offer should not simply be transplanted into an emergency.

A refusal should not become a permanent judgment that the person never wants help. Desired assistance can change with the task, time, and circumstances.

A design hypothesis for physical AI

A possible design separates a candidate need for help from permission to move. A recognition model's "dropped object" classification should not directly trigger a motor command. Wait for a response to the offer and stop the current intervention after refusal.

Saying "Okay" is insufficient if queued physical commands remain active. Cancel those commands and remain available for a new request. Smiles, silence, or familiarity should not automatically authorize touching belongings.

Evaluation could compare acceptance, refusal, unclear replies, and later changes of preference. Examine intervention after refusal and repeated unwanted offers, alongside help counts and obstruction of independent action.

A decision in four panels

Someone notices a fallen paper.
They offer help; the owner declines calmly.
The helper withdraws and waits while the owner retrieves it.
The paper is back and the relationship remains comfortable.

This fictional comic distinguishes inattention from checking and waiting.

A question to consider

When do you want to do something yourself? Experiences of having that room preserved can inform how you support another person.

References

About this series

LOGIHEART's research and development vision

LOGIHEART is researching how to implement social capabilities in physical AI that interacts with people.

Explore the research and development vision at the LOGIHEART website and in the English introduction video.

Hearing the words and understanding the request are different

LOGIHEART — Thu, 22 Oct 2026 00:00:07 +0000

Social capabilities across human development | Part 11

Listening to the end of someone's statement matters, but receiving words does not establish shared meaning. Listening also includes checking understanding, responding in a correctable form, and reflecting corrections in action.

The blue bowl, not the red one

In a kitchen, someone asks, "Could you pass that bowl?" The other person reaches toward a red bowl but pauses to ask, "This one?" The requester points to a blue bowl. It is handed over, and cooking continues.

In this fictional scene, merely hearing and executing the words would have produced the wrong object. Clarification jointly resolves ambiguity. The correction is used as information for the task rather than treated as defeat or blame.

The difficulty of applying understanding

Apperly and colleagues investigated adults' difficulties interpreting instructions in relation to another person's perspective. Their experiments provide material for questioning whether using others' information in action is always automatic. See Why are there limits on theory of mind use?.

A communication task cannot be generalized directly to every household request. This article distinguishes understanding a perspective from selecting a target on that basis, and considers the value of everyday clarification.

The adult experiments cited here do not establish how these abilities change during childhood. In this series, the practical question is how people can make their understanding correctable as situations and communication demands change.

Too much clarification can also impose a burden

Repeating every request at length can increase the other person's workload. Focus clarification on differences affecting the outcome. Ask about color when bowl color matters, or time when timing matters. Avoid repeatedly questioning information the person already made clear.

"Understood" alone does not show what was understood. A brief statement such as "I'll pass the blue bowl" makes a mistake correctable. Clarification coordinates joint action rather than testing the speaker's words.

A design hypothesis for physical AI

A possible design separates speech recognition, candidate meanings, execution targets, and permission to act. Correctly recognized words can still leave the target ambiguous. Defer contact with the target while clarifying, then update the planned target after correction.

Recording a correction in memory while continuing an old movement is not recovery. Test whether conversation and physical commands use the same confirmation result.

Evaluation could include similar objects, reformulation, cancellation, and correction after movement begins. Examine useful clarification and unnecessary clarification burden together. Correcting the target matters more than declaring that the person's mind was understood.

A decision in four panels

A blue bowl is requested.
The other person starts reaching toward a red bowl.
They pause and check the color.
They hand over the blue bowl.

This fictional comic depicts bringing understanding and execution into agreement rather than competing in fluency.

References

Why are there limits on theory of mind use?

About this series

LOGIHEART's research and development vision

LOGIHEART is researching how to implement social capabilities in physical AI that interacts with people.

Explore the research and development vision at the LOGIHEART website and in the English introduction video.