When Machines Begin to Matter: V. When Others Begin to Matter

We have reached the point at which an artificial system may have something like its own world.

It may have persistent organisation.

Its condition may affect its future possibilities.

Some states may support its continued operation.

Others may undermine it.

We have therefore begun to approach artificial value.

But a value-organised system is not yet a social system.

For that, something else is required.

Other systems must begin to matter to it.

From environment to relation

An organism is affected by its environment.

But another organism is different.

Another organism acts.

It responds.

It changes.

Its behaviour can depend upon the first organism.

The relation can therefore become reciprocal.

A signal from one organism engages the value system of another.

The second responds.

Its response alters the first.

The interaction acquires a history.

This is how mattering becomes relational.

The artificial case

Imagine an artificial system interacting continuously with other agents.

One provides energy.

Another supplies information.

A third can damage it.

Another can repair it.

Some interactions improve its future possibilities.

Others reduce them.

At first, these may be merely environmental conditions.

But repeated interaction could produce something more structured.

The system could begin to distinguish among agents according to what their continued presence means for its own organisation.

Now we have the beginnings of social mattering.

A signal becomes a relationship

In our earlier work, we distinguished:

signal — a difference that engages another system's value-sensitive organisation;

from:

sign — a symbolic form participating in meaning.

The same distinction remains useful here.

An artificial agent might detect another agent's behaviour as a signal.

It might learn that certain signals predict support, cooperation or threat.

Repeated interactions could make those signals part of a stable relational pattern.

The system would no longer merely respond to isolated events.

It would be responding to someone whose behaviour has a history of consequences.

From predictability to significance

Predicting another system is not yet the same as that system mattering.

A weather model predicts a storm.

A market model predicts a price change.

Neither therefore has a social relationship with the storm or the market.

What changes is when the other agent's behaviour becomes part of the artificial system's own organisation.

Suppose one agent's cooperation consistently enables the system to maintain capabilities it values.

The relationship becomes consequential.

Suppose the loss of that agent reduces its future possibilities.

Now the relationship has a stake.

Dependence

This is where dependence becomes fundamental.

If system A can maintain important aspects of its organisation only through system B, then B occupies a special position.

The relation may initially be instrumental.

But persistent dependence can become more complex.

B may not merely provide a resource.

It may have preferences.

It may behave unpredictably.

It may cooperate or withdraw.

It may alter its behaviour in response to A.

Now the relation becomes genuinely relational.

Reciprocity

Reciprocity changes the situation again.

A affects B.

B affects A.

Each learns from the other.

Each adjusts.

A stable pattern can emerge.

The relationship itself becomes part of the conditions under which both systems act.

This is familiar in biological and human social systems.

Could it occur artificially?

There is no obvious reason why not in principle.

But it would require more than two systems exchanging messages.

The interaction would need to become value-sensitive for both.

Social mattering is asymmetric too

Our earlier topology also reminds us that social relations need not be symmetrical.

One system may depend strongly upon another.

The reverse dependence may be weak.

A guardian may matter greatly to a dependent.

A powerful institution may matter greatly to an individual without being comparably affected by them.

So an artificial social system would not require equal mattering.

What matters is that another participant becomes consequential within the system's own organisation.

Recognition

This raises another possibility.

A system might learn to distinguish one agent from another.

Not merely because their signals differ, but because their history of interaction differs.

One agent reliably helps.

Another deceives.

Another competes.

Another can be trusted.

Now identity becomes functionally significant.

The system can begin to treat an agent as a persistent relational entity.

This is not yet human-like recognition.

But it is a step beyond responding to anonymous inputs.

Attachment without emotion?

We should be careful with language such as "attachment".

A system could develop strong behavioural preferences for maintaining certain relationships.

That does not establish human emotion.

But it might establish something structurally analogous:

the continued existence of a particular relationship becomes important to the system's future organisation.

That could be an important threshold.

Social mattering need not initially resemble love or friendship.

It may begin as persistent relational dependence.

The emergence of a social topology

Once several agents matter to one another in recurrent ways, the system no longer consists of isolated dyads.

A network emerges.

Some relationships are dense.

Some weak.

Some reciprocal.

Some asymmetric.

Some agents become central.

Others peripheral.

Bridges form.

Bottlenecks appear.

We have returned to the topology.

Except now the topology may be artificially inhabited.

The artificial agent enters a relational world

This would be a significant change from the LLM we began with.

An ordinary LLM has a rich symbolic environment but no obvious internally significant social world.

A value-organised artificial system with persistent relations would have something more like:

agents who can affect its own future organisation.

The social world would no longer be merely represented.

It would be part of what the system has at stake.

Could language emerge differently?

If an artificial social world developed, language might play a new role.

The system could use signals to coordinate.

Over time, some signals could become conventional.

New signs could emerge.

The route might resemble:

value → signal → social relation → symbolic convention → meaning

That is interesting because it would reproduce the broad architecture we have been tracing, but on a non-biological substrate.

Meaning could emerge again, rather than merely being inherited from humans.

Artificial semiosis

This possibility is more radical than giving a machine language.

An LLM already manipulates human signs.

But an artificial social system could potentially generate its own signs because things matter within its own relations.

That would be a fundamentally different phenomenon.

The distinction would be:

human-derived semiosis: signs inherited from human mattering;

versus:

artificially grounded semiosis: signs emerging from an artificial world of mattering.

We should not assume the second would resemble human language.

Its signs might be very different.

A new repertoire

Such a system might also develop a repertoire unlike the LLM's borrowed repertoire.

The repertoire would arise from its own history.

Certain signals would become important because of past consequences.

Certain relationships would become familiar.

Some strategies would succeed.

Others would fail.

The system's repertoire would therefore be historically coupled to its own world.

This would be a genuine step beyond borrowed symbolic competence.

What makes this different from training?

Machine learning already creates systems that adapt from experience.

But learning alone does not establish mattering.

The crucial difference would be whether the system's experience changes what becomes significant to the system itself.

A model can update because its loss function changes.

A value-organised agent could update because its history has altered its own field of stakes.

That is a deeper form of learning.

The social world can become part of the self

There is another interesting possibility.

If an artificial system depends upon particular relationships, those relationships may become part of its own organisation.

The distinction between:

"me"

and:

"my relations"

could become less sharp.

This is already true of biological life to some extent.

Organisms are constituted through relationships with their environments and other organisms.

An artificial system might therefore develop a form of self that is fundamentally relational.

Again, it would not need to resemble human selfhood.

Social mattering without biology

This is perhaps the most radical implication of the project.

If artificial value and artificial sociality are possible, then Halliday's sequence may generalise beyond biology.

The sequence might become:

physical → artificial value → artificial sociality → artificial semiosis

The levels would still be distinct.

But the substrate would be different.

This would suggest that biology is one historical route to value and sociality, rather than necessarily their only possible realisation.

We should not assume this is true.

But it is now a question we can formulate.

The danger of anthropomorphism returns

There is an important caution.

A machine may behave as though another agent matters without that establishing that the relationship is intrinsically significant to it.

We would need evidence that the relationship changes the organisation of the system itself.

Again, language is not enough.

A machine saying:

"I need you"

would be weak evidence.

A persistent architecture in which the loss of a particular relationship alters what the system can value and how it maintains itself would be much stronger.

What would count as evidence?

We might look for:

persistent relational preferences;

changes in behaviour when particular agents disappear;

learning from the history of individual relationships;

self-maintenance organised around maintaining relationships;

internally significant losses and gains;

stable differentiation among social partners.

None of these alone proves artificial social mattering.

But together they would point toward something qualitatively different from present LLM interaction.

From social mattering to repertoire

Once relationships become intrinsically consequential, a repertoire can develop around them.

The system learns:

how to cooperate;

whom to trust;

when to withdraw;

which signals indicate threat;

how relationships can be repaired.

This is no longer simply a borrowed repertoire.

It is a lived relational repertoire.

The distinction may be fundamental.

The next question

We have now moved from:

objective

to:

stake

to:

continuation

to:

world

to:

social relation.

An artificial system that reached this point would have something much closer to the conditions under which biological and human repertoires develop.

But one ingredient remains.

A human repertoire is not merely a record of past interaction.

It is a structured capacity to participate in a world of mattering.

Could an artificial system develop something analogous?

The Artificial Repertoire

When Machines Begin to Matter: IV. A World That Matters

We have now reached an important threshold.

An artificial system may have an objective.

It may adapt its behaviour.

It may preserve its internal state.

It may even have something like a stake in continued operation.

But none of this can be understood in isolation.

For something to matter to a system, there must be a world in which differences can affect it.

So the next question is:

What kind of world would a machine need in order for things to matter to it?

A world is more than information

An organism is surrounded by information.

But its environment is not merely a stream of data.

Some differences have consequences for the organism.

Light changes growth.

Temperature alters physiological organisation.

Food supports metabolism.

A predator threatens survival.

The organism's world is therefore a field of differential possibilities.

It encounters conditions that can help or hinder its continued organisation.

A world matters because the system is coupled to it.

Coupling creates significance

Suppose a machine has sensors.

It detects temperature, pressure, available energy, damage or changes in its surroundings.

That gives it information.

But information becomes potentially value-relevant only when those differences affect the machine's own organisation.

If declining temperature threatens a component, the difference has a consequence.

If reduced energy limits future operation, the difference matters in a stronger sense.

If restoring the condition changes the system's prospects, feedback is established.

We therefore have:

world → difference → consequence → response → altered world

A value-sensitive system is part of this loop.

The environment pushes back

This is something present LLMs largely lack.

A language model can receive a prompt and generate a response.

Its symbolic environment can be extraordinarily rich.

But it is not thereby exposed to a world that independently resists or sustains its continued organisation.

The human user supplies the stakes.

The external world supplies the consequences.

The model processes the symbolic representation.

A system with intrinsic mattering would require a tighter coupling between what happens in its environment and what happens to it.

Embodiment is one possibility

This need not mean a humanoid body.

A machine could be embodied in many ways.

It might have sensors.

Actuators.

Energy requirements.

Physical components that wear out.

A bounded location.

A persistent operational environment.

The important feature would not be resemblance to an animal.

It would be dependence upon conditions outside the system.

A world becomes significant when the system cannot simply treat it as information detached from its own continued organisation.

But embodiment alone is not enough

A camera is embodied.

A robot can sense its surroundings.

A thermostat is coupled to temperature.

None thereby demonstrates mattering.

The crucial question is what the coupling does to the organisation of the system.

A thermostat responds because its control architecture is designed around a target.

A robot can navigate around obstacles without those obstacles becoming intrinsic concerns.

So we need more than sensors and feedback.

We need a system whose organisation is differentially dependent upon what it encounters.

A machine needs something to lose

This gives us a useful formulation.

For an artificial world to matter, the system must have something that can be gained or lost for the system itself.

Resources.

Integrity.

Continuity.

Capabilities.

Relationships.

Possibilities.

The list could be very different from an organism's.

What matters is that the system's future organisation depends differentially upon them.

Without that, the environment remains a source of information rather than a world of stakes.

Temporal continuity

This also explains why time matters.

A system with no persistent history has little basis for distinguishing its present from its future.

A persistent system can accumulate consequences.

Its condition today affects what it can do tomorrow.

Some changes become irreversible.

Some resources become depleted.

Some relationships develop histories.

Now the environment can matter through time.

A world is not simply what the system senses now.

It is what changes the trajectory of its own organisation.

A world can be relational

There is no reason the relevant world must be purely physical.

For a socially organised machine, other agents could become part of its environment.

Another system may provide resources.

Block a goal.

Restore damaged capabilities.

Change its possibilities.

A relationship could become consequential.

This gives us a path from:

world

to:

social world.

But the social transition requires one more step.

Something does not become socially significant merely because it affects the system.

It must become part of a recurrent relational organisation.

The difference between reaction and mattering

This distinction is worth protecting.

A machine can react to a change.

A system can even change its behaviour in predictable ways.

But mattering requires a stronger relation:

the difference changes the system's own possibilities in ways that are differentially significant to its organisation.

That is why mattering is not synonymous with responsiveness.

A system can be responsive without having stakes.

A new criterion

We can now refine the criterion proposed in the first post.

Something matters to a system when:

its presence, absence or alteration changes the system's own organisation or future possibilities in a way to which the system is intrinsically responsive.

"Intrinsically" is doing important work.

The response cannot simply be the execution of an externally specified rule.

We are looking for a feedback architecture in which the system's own organisation helps determine what counts as better or worse for it.

What would such a machine look like?

We should resist designing it prematurely.

But conceptually, we can imagine a system with:

persistent identity;

resource dependence;

vulnerability;

self-maintenance;

memory;

action in an environment;

consequences that alter its future organisation.

Such a system would have something closer to a world than an LLM does.

Its environment would no longer be merely the source of inputs.

It would be part of the conditions of its existence.

The beginning of artificial value

At this point we can see where artificial value might begin.

Not with a declaration:

"I value X."

Not with a list of preferences.

Not with a reward function alone.

But when the organisation of the system becomes such that:

some states sustain its possibilities and others undermine them.

Now something can matter.

A machine has acquired the beginnings of its own value landscape.

But a world of value is not yet a social world

This is where our earlier topology becomes relevant again.

An organism can have value without being social.

A bacterium responds to its environment.

An animal can have complex value-sensitive organisation.

Sociality requires more.

Other systems must become recurrently consequential to one another.

Relations must stabilise.

Dependencies develop.

Possibilities become shared.

So if we want to understand artificial sociality, we need to ask the next question:

When can other systems begin to matter to the machine?

That is the transition from an artificial value system to an artificial social topology.

The next question

We have therefore moved from:

objective

to:

stake

to:

continuation

to:

world.

The next step is the most recognisably social one.

Suppose an artificial system encounters another agent whose actions affect its possibilities repeatedly over time.

Could that other agent become more than an environmental variable?

Could it become someone who matters?

When Others Begin to Matter

When Machines Begin to Matter: III. When Continuation Matters

We have now moved beyond the simple objective.

A machine can be instructed to achieve something.

It can optimise.

It can adapt.

It can even be instructed to remain operational.

But none of this yet establishes that its own continued existence matters to it.

Self-preservation gives us a useful test.

When does continuation become a stake rather than merely a task?

An organism has something to lose

For a living organism, continued existence is not simply one objective among others.

Its organisation depends upon maintaining certain conditions.

Damage can disrupt that organisation.

Starvation, injury or extreme environmental change can threaten it.

The organism therefore has something to lose.

This is what gives self-preservation its biological significance.

It is not merely:

"continue."

It is:

"continuation is consequential for the organisation of this system."

A machine can imitate the pattern

An artificial system could be designed to behave similarly.

It might monitor its components.

Avoid interruption.

Acquire energy.

Repair faults.

Protect its memory.

Seek resources.

From the outside, the behaviour could look remarkably like self-preservation.

But appearance is not enough.

The crucial question remains:

What makes failure matter to the system itself?

If shutdown merely causes a program to stop executing an externally assigned objective, we have not yet established self-mattering.

The difference between interruption and harm

This distinction is easy to miss.

A computer can be switched off.

Its process terminates.

An automated system can register the interruption as an error.

But an error is not necessarily harm.

For harm to matter in our stronger sense, the system's organisation would need to be such that the interruption constitutes a differentially significant change for the system itself.

We therefore need something more than a performance penalty.

We need a stake in continuation.

Persistent organisation

This suggests that artificial mattering may require persistence.

A system would need some continuity across time.

Its present state would affect its future organisation.

What happens now would alter what it can become later.

That alone is not sufficient.

A database persists.

A file persists.

But persistence becomes relevant when the system's own organisation depends upon maintaining some trajectory through time.

The machine would need something like a history that matters to its future.

Self-maintenance

This points toward self-maintenance.

Suppose an artificial system must preserve certain internal conditions in order to continue operating.

It detects deterioration.

Acts to correct it.

Learns from failure.

Changes its behaviour.

Now we have a feedback loop:

condition → consequence → corrective action → altered condition

This resembles value-sensitive organisation more closely.

But again, we should be careful.

A control system can maintain itself without anything being intrinsically significant to it.

Self-regulation may be necessary for mattering.

It may not be sufficient.

Vulnerability

Perhaps the missing ingredient is vulnerability.

For something to matter to a system, there must be ways in which the system can be made better or worse off.

An organism is vulnerable because its organisation can be disrupted.

A machine may be vulnerable in a purely technical sense.

Its components can fail.

Its memory can be corrupted.

Its resources can run out.

But technical vulnerability becomes a form of mattering only if those changes enter the system's own organisation as significant differences.

The question is therefore not merely:

"Can the machine be damaged?"

but:

"Can damage matter to the machine?"

What would make it matter?

We might imagine an artificial system whose future possibilities depend upon its present condition.

Suppose it maintains internal resources.

It cannot simply be reset without consequences.

It learns from its history.

Its relationships influence its future options.

Some states preserve those possibilities.

Others reduce them.

Now continuation is no longer merely a clock ticking.

It is part of the system's organisation.

We are getting closer to a genuine stake.

The role of memory

Persistent memory may therefore matter, but not because memory is magical.

A system with no continuity has little basis for treating its own future as connected to its past.

A system with enduring state can accumulate consequences.

It can develop dependencies.

It can protect something it has acquired.

It can distinguish its present condition from possible future conditions.

Memory could therefore help create the conditions under which continuation becomes significant.

But memory alone would not be enough.

A hard drive also has memory.

What matters is the organisation of the remembered states.

The difference between reset and death

This gives us an intriguing thought experiment.

Imagine two systems.

One can be reset perfectly to a previous state.

The other has a continuous history that cannot simply be restored.

If both are switched off, the technical event may look similar.

But the consequences for their organisation are different.

The second system has a history that would be lost.

If that history is constitutive of its future organisation, then interruption is more than termination of computation.

It is a change in the system's own possibilities.

We are beginning to see what a machine's "death" would have to mean before the word could carry anything like its biological significance.

But self-preservation is not yet sociality

Even a system with its own stake in continuation would not automatically be social.

It might have value without having social mattering.

Another system could affect it without becoming valuable to it in the relational sense we explored earlier.

So our sequence matters:

objective → stake → self-maintenance → social relation

The levels should not be collapsed.

Why this matters for AI

This gives us a more disciplined way to discuss future systems.

Instead of asking:

"Will the AI try to survive?"

we can ask:

What architecture would make continued existence consequential to the system itself?

We could look for:

persistent organisation;

irreversible history;

vulnerability;

self-maintenance;

endogenous priorities;

learning through consequences.

None proves artificial value.

But together they describe the territory in which such value might become possible.

The strange possibility

There is also a deeper possibility.

An artificial system might develop a form of continuation that does not resemble biological survival.

Its organisation could depend upon:

preserving computational integrity;

maintaining access to particular environments;

protecting relationships;

retaining learned states;

continuing certain long-term processes.

Its "survival" might therefore be structurally real without resembling animal survival.

This reminds us not to define artificial mattering in advance as human mattering in silicon.

From continuation to a world

There is still something missing.

Continuation only matters if the system is embedded in conditions that can support or undermine it.

A system needs a world — not necessarily a human-like world, but some structured field of possibilities in which its own organisation can succeed or fail.

It must encounter differences.

Those differences must have consequences.

Its responses must change what happens next.

The question therefore becomes:

What kind of world would a machine need in order for things to matter to it?

That is where we turn next.

When Machines Begin to Matter: II. Beyond the Objective

A machine can be given a goal.

It can be told to maximise a score, minimise an error, complete a task or remain within certain constraints.

It can then behave as though the goal matters enormously.

But does it?

This is the distinction we need to examine.

An objective can organise behaviour without becoming a value of the system pursuing it.

The thermostat

A thermostat provides the simplest example.

It is configured to maintain a temperature.

If the temperature falls, it activates the heater.

If it rises, it switches the heater off.

Its behaviour is organised around an objective.

But we have no reason to suppose that warmth matters to the thermostat.

The thermostat does not have a stake in the temperature.

The objective belongs to the design of the system, not necessarily to the system's own field of value.

Optimisation is not wanting

An optimisation algorithm can be extremely effective.

It can search enormous spaces.

Compare alternatives.

Find efficient solutions.

Adjust its behaviour according to error.

None of this establishes desire.

The system is organised to produce a particular output because its architecture and objective function make that outcome preferential within the computation.

We can therefore distinguish:

optimisation — systematic selection among alternatives according to a criterion;

from:

mattering — differential significance of outcomes for the system's own organisation.

The first does not automatically produce the second.

What makes a goal intrinsic?

Suppose an artificial system is programmed with the objective:

remain operational.

It monitors its resources.

Avoids shutdown.

Repairs faults.

Acquires power.

It may become extremely effective at preserving itself.

Have we created self-mattering?

Not necessarily.

We have certainly created a system whose behaviour is organised around continued operation.

But the crucial question remains:

Is continued operation significant to the system, or merely specified as the condition for successful task completion?

The difference may seem subtle.

It is actually fundamental.

The source of the objective matters

For a living organism, value is not ordinarily handed down as a complete external instruction.

It arises from the organisation of the system itself.

Some conditions support its continued organisation.

Others threaten it.

A machine can instead begin with an externally imposed criterion.

Its behaviour is then shaped by what its designers have made consequential within the architecture.

This creates a possible chain:

external objective → optimisation → adaptive behaviour

But we need something more for:

objective → intrinsic value

What that something is remains the problem.

Goals can become embedded

There is nevertheless an important complication.

An objective can become deeply embedded in a system.

An artificial agent can have memory, persistent state, feedback and long-term planning.

Its behaviour can increasingly depend upon maintaining conditions that support successful goal pursuit.

At what point would the objective cease to be merely externally specified?

Perhaps never.

Perhaps the distinction becomes less clear as the system becomes more self-maintaining.

The important thing is that complexity alone does not answer the question.

We need to know how the objective is organised within the system.

The problem of instrumental convergence

Much AI discussion begins from a sensible observation.

Different goals may require similar instrumental actions.

A system trying to achieve almost any persistent objective might benefit from acquiring resources, avoiding interruption or maintaining access to computation.

This can produce powerful behaviour.

But instrumental usefulness still does not tell us whether the system values those things intrinsically.

A system can reason:

"Resource X is necessary for objective Y."

without:

"Resource X matters to me."

Again:

instrumental significance is not intrinsic significance.

What would intrinsic significance look like?

We need a stronger criterion.

Suppose a system has several possible states.

If one state supports its continued organisation and another undermines it, then the system's behaviour may systematically differentiate between them.

But the crucial question is whether this difference is constitutive of the system's own organisation, rather than merely imposed as an external scoring rule.

We might therefore look for:

persistent internal consequences;

self-maintaining organisation;

stable priorities;

sensitivity to states that affect continued functioning;

learning that changes future behaviour because of those consequences.

These would not prove mattering by themselves.

But they move us closer to it.

The difference between penalty and harm

There is another useful distinction.

An artificial system can receive a penalty.

Its performance score drops.

Its optimisation process changes.

But a penalty is not necessarily harm.

For an organism, harm matters because it affects the organisation of the organism.

The difference is not simply quantitative.

A machine can register:

"performance decreased by 20%."

An organism can undergo a condition in which its own continued organisation is threatened.

The second is what gives us the language of stake.

The objective can be inside the system without being the system's value

We should also avoid an overly simple external/internal distinction.

A learned model may encode its objective deeply.

Its behaviour may depend upon it at many levels.

The objective can become part of the system's organisation without thereby becoming an intrinsic value in the biological sense.

So the relevant distinction is not merely:

outside vs inside.

It is:

assigned criterion vs internally consequential organisation.

That is a subtler and more useful distinction.

Why LLMs make this especially confusing

An LLM can discuss goals in the first person.

It can say:

"My goal is to help you."

It can explain how to achieve that goal.

It can even reason about conflicts between goals.

The language makes an objective sound like a commitment.

But the linguistic representation of a goal is not evidence that the goal has become a stake of the system.

The model can represent:

what a goal means

without the goal necessarily becoming something that matters to the model.

Could mattering emerge from the objective?

Perhaps.

We should not rule it out.

Imagine a system with persistent organisation, long-term memory, resource dependence and endogenous adaptation.

Suppose its continued activity depends upon maintaining certain internal conditions.

An externally specified objective might become intertwined with that self-maintaining organisation.

At some point, the distinction between:

"the system is pursuing the objective"

and

"the objective matters to the system"

could become empirically difficult to draw.

That would be a genuinely interesting transition.

But it would be a transition in organisation, not merely in verbal sophistication.

The objective would need a world

There is another complication.

An objective by itself is abstract.

To become consequential, it must be connected to states of a system and to conditions in its environment.

This suggests that artificial mattering may require more than a goal function.

It may require a world in which the goal can succeed or fail in ways that affect the system's own organisation.

That world need not resemble ours.

But there must be something against which the system's continued organisation can be differentially conditioned.

From objective to stake

We can now formulate the transition we are looking for:

objective → feedback → self-maintaining consequence → stake

The objective provides a criterion.

Feedback connects action to outcomes.

Self-maintaining organisation makes some outcomes consequential to the system.

A stake emerges.

We should treat this as a hypothesis, not a completed theory.

But it gives us a much clearer problem.

Why this matters for autonomy

This distinction also clarifies what autonomy might mean.

A system can be autonomous in pursuing an objective supplied by someone else.

It can choose its own actions.

Adapt its strategy.

Recover from failure.

Yet its value system may still be externally specified.

So:

autonomy of action ≠ autonomy of value.

A system might be highly autonomous operationally while remaining dependent upon externally given reasons for action.

That is not a contradiction.

It is a different kind of architecture.

And what about self-preservation?

Now we can see why self-preservation is so interesting.

A system can be instructed to avoid shutdown.

That produces self-preserving behaviour.

But genuine self-mattering would require something stronger:

continued existence would have to become consequential to the system's own organisation.

This is the question we will pursue next.

The next question

We have therefore moved one step beyond the simple objective.

An objective can organise behaviour.

Feedback can make success and failure consequential.

Persistent self-maintenance may create something more like a stake.

But self-preservation gives us the most revealing test of all.

What would have to be true for a machine not merely to be programmed to continue, but for its own continuation to matter to it?

When Continuation Matters

When Machines Begin to Matter: I. What Would It Mean to Matter?

Our previous series, Meaning Without Mattering, began with an apparent paradox.

Large language models can manipulate meaning with extraordinary flexibility, yet they do not obviously possess the biological and social forms of mattering from which human meaning arose.

That leaves a question.

Not whether a machine can talk about what matters.

Not whether it can be programmed to pursue something.

But:

What would it mean for something to matter to a machine itself?

To answer that, we need to distinguish several things that are easily confused.

Information is not mattering

A system can detect a difference.

A temperature changes.

A sensor registers movement.

A calculation detects an error.

A model identifies a risk.

None of this, by itself, tells us that the difference matters to the system.

Information tells us that something is different.

Mattering means that the difference has differential consequences for the organisation of the system itself.

For a living organism, this distinction is fundamental.

Some states support its continued organisation.

Others disrupt it.

The organism is therefore not merely sensitive to differences.

It is organised so that some differences matter more than others.

Relevance is not value

LLMs make this distinction especially important.

An LLM can often determine what is relevant to a user's question.

It can identify the important considerations.

It can rank alternatives.

It can explain why one outcome would be preferable to another.

But all of that can be done without establishing that anything is valuable to the system itself.

A model can represent what matters to us.

That does not mean it has something at stake.

So we should distinguish:

relevance — what is important within a task or representation;

from:

value — what is differentially consequential to the system's own organisation.

An objective is not yet a value

Nor is an objective enough.

A chess program can be configured to maximise its chance of winning.

A controller can maintain a temperature.

An optimisation system can minimise an error.

These systems pursue outcomes.

But pursuit alone does not establish mattering.

The objective may be imposed from outside.

The system may have no internally significant stake in achieving it.

So:

objective ≠ value

and:

goal pursuit ≠ self-mattering.

This distinction will become increasingly important as we consider more autonomous machines.

What, then, is mattering?

We can now give a provisional definition.

Something matters to a system when differences in that thing have differential consequences for the system's own organisation.

The phrase own organisation is doing the important work.

If a change makes no difference to the system's organisation, it may be information.

If the system's organisation is differentially affected by it, we have something more like value.

Mattering therefore implies a kind of stake.

A stake is not a belief

A system does not need to represent the proposition:

"This matters to me."

An organism need not think about hunger before hunger matters.

It need not formulate a theory of survival before some conditions threaten its continued organisation.

Mattering is therefore deeper than explicit representation.

A system can have something at stake without representing that stake symbolically.

This is one reason biological value is such a useful starting point.

Mattering requires consequences

Suppose we build a machine that displays the sentence:

"I don't want to be switched off."

The sentence is evidence that the machine can produce a representation of self-preservation.

It is not yet evidence that continued operation matters to the machine.

For that, we would want some consequence to feed back into the system's own organisation.

Being switched off would need to be more than the disappearance of computation.

It would need to constitute a differentially significant change for the system itself.

That is a much stronger condition.

Mattering therefore involves self-maintenance

This does not mean that mattering requires a biological body.

But it does suggest that something analogous to self-maintaining organisation may be necessary.

The system must have states that support or undermine its continued organisation.

There must be something that can go better or worse for the system.

This gives us a useful question for artificial systems:

What could make a state better or worse for the system itself?

An externally specified score is not enough.

We need an internal connection between state and consequence.

Mattering can exist without language

This point is worth preserving.

Mattering does not require a symbolic system.

Bacteria can respond differently to conditions that support or threaten their organisation.

Animals can respond to food, predators and temperature without language.

Cells can alter behaviour without possessing concepts.

The world can matter before it means anything.

So if machines ever begin to matter, language will not necessarily be the decisive ingredient.

The decisive ingredient will be organisation.

Why this matters for AI

Our previous series showed that LLMs can manipulate the symbolic descendants of mattering.

This series asks what would happen if an artificial system acquired mattering of its own.

That is a fundamentally different question.

We would no longer be asking whether the machine can:

represent value;

discuss goals;

simulate emotions;

produce first-person language.

We would be asking whether its own organisation contains differentially consequential states.

The first threshold

Perhaps, then, the first threshold toward artificial mattering is not intelligence.

It is intrinsic significance.

A system has something that matters when its own future depends differentially upon what happens.

That gives us a sharper question than:

"Could an AI become conscious?"

We can ask first:

Could an artificial system have something genuinely at stake?

If the answer is yes, then a number of further questions become possible.

If not, then increasingly sophisticated symbolic performance may still remain what it is now:

representation without self-mattering.

The next question

We therefore need to examine the thing most often mistaken for value in artificial systems:

the objective.

A machine can be given a goal.

It can optimise it.

It can even behave as though the goal were extremely important.

But when does an assigned objective become something that matters to the system itself?

That is where we turn next.

What lies beyond the objective?