Rereading SDT from the Depths: The 'Limits and Essential Possibilities' of Autonomy, Competence, and Relatedness

Autonomy—
what exactly is it?

It would be a great help if subordinates could work autonomously.
On the other hand, changing jobs and quitting is actually also a manifestation of strong autonomy.

We have been told that "autonomy is a good thing," but
then what is autonomy?
And why does it drive people?

I have explored numerous studies alongside AI,
looking at the "structure" that lies beneath them.

This time, I will quietly examine from a deeper level
why SDT (Self-Determination Theory) has garnered attention and why it failed to change the world.

SDT (Self-Determination Theory)

SDT is a theoretical framework in psychology stating that
people's performance is highest when they feel they have "chosen for themselves."
It is a theoretical system in psychology.

Even within psychology,

• the volume of empirical research,

• the consistent theoretical structure,

• and the application to multiple fields

are overwhelming, making it one of the most influential motivation theories in the world.
The founders are Edward Deci and Richard Ryan.

At the center of SDT are three basic psychological needs (the three elements).

• Autonomy:
  The feeling of being able to choose for oneself.

• Competence:
  The feeling of being able to do something or growing.

• Relatedness:
  The feeling of connecting with others and being understood.

Next, let's look at how these were empirically proven.

Cognitive Evaluation Theory (CET)

A theory on the relationship between intrinsic motivation (motivation that wells up from within, "doing it because I want to") and autonomy/competence.

In fact, it was proven through experiments like this.

• Solving an interesting puzzle

• Tell only some participants that 'you will be paid if you solve it,' and reward them if they actually do

• Give them free time after the task (no reward)

• Only those who received a reward stop returning to the puzzle

This is thought to be because the reward changed 'I want to do it' into 'I am being made to do it.'
And in the post-experiment questionnaire,

• 'Interest waned'

• 'I felt like I was being made to do it'

were the trends that emerged.

This experiment proves the following:

Relationships Motivation Theory (RMT)

A theory that shows how intrinsic motivation (motivation that wells up from within, 'doing it because I want to') and 'connection with others' are linked.
This area has more empirical evidence in everyday settings (schools, workplaces, organizations) than in laboratories,

and strong effects have been observed in the 'field' where people actually interact with others while acting.
For example—

For example—

• Performance improves just by being called by name
  Just by having a boss or teacher remember and use their name, the time spent on tasks improved.

• Results improve just by a 'declaration of trust' from a leader
  In medical settings, IT companies, universities, etc., just by leaders briefly expressing trust,
  cooperative behavior, creative behavior, and problem-solving speed improved.

• The amount of effort changes with just a single word of 'acknowledgment'
  In the workplace, productivity increased and overtime hours and turnover rates decreased.
  In schools, there were numerical differences, such as a significant increase in assignment submission rates and class participation.

What became clear from those experiments was—

Basic Psychological Needs Theory (BPNT)

When the three basic needs (the three elements) of 'autonomy, competence, and relatedness' are met, people grow healthily, and
when they are not met, distortions arise—this is the core of Self-Determination Theory.

Furthermore,this theory demonstrates that these three elements are universal across cultures and ages.

Verification has been conducted primarily through factor analysis using surveys, and
in at least 78 countries and hundreds of studies,
'the three factors of autonomy, competence, and relatedness appear universally'
has been confirmed.

Moreover, in recent years, it has become clear that subjective responses (surveys) align with objective indicators in the brain.
For example—

・Autonomylowers the 'control load' on the prefrontal cortex, reducing stress
・Competenceincreases the response of the reward system (dopamine)
・Relatednesswhen satisfied, lowers stress hormones (cortisol)

In other words, the three elements are supported not only by 'how people feel' but also by 'how the brain and body react.'

What lay beyond the empirical support

SDT has been verified from various other angles as well, and
'autonomy, competence, and relatedness enhance human performance and well-being' has been repeatedly confirmed.

Deci and Ryan must have thought the following:

• Education will change all at once
  → Children's learning requires freedom, and curiosity-based learning should become the mainstream

• Business will also change
  → Evaluation-based systems will weaken, and autonomous organizations will increase

• People will be able to live more freely
  → Society will move based on internal values rather than external coercion

However—what actually happened over the last 30 years was the exact opposite.

• Education will change
  → But it was judged that 'giving freedom lowers academic achievement,' and management intensified

• Business will change
  → But numerical management of results actually accelerated

• People will become freer
  → But social stress increased, and happiness levels barely rose

Why did this reversal occur?

As I continued to deeply question this problem with AI, I realized that
'the results of the empirical evidence were not fully grasped as a structure'
was the cause.

By looking deeply at the structure,
why SDT could not change reality
becomes surprisingly clear.

Autonomy

In SDT, the level of autonomy can be classified as follows.

• External Regulation
  Following rewards and punishments.
  Example: Doing homework because you will get scolded.

• Introjected Regulation
  Driven by anxiety, guilt, shame, etc.
  Example: Doing it because you feel like you have to.

• Identified Regulation
  Acting with an understanding of the value of the behavior.
  Example: Studying because it will be useful in the future.

• Integrated Regulation
  The value is integrated as part of oneself.
  Example: Doing it because that is how I want to be.

Beyond this lies intrinsic motivation.
However, this does not capture the whole picture.
We need to grasp autonomy in a more structurally broad way.

Level 1: Social Control

A state in society where one is driven by rewards and status.
The starting point of action, such as 'working because you get paid,' is external.
SDT has shown that at this level, both performance and happiness are low, and one should transition to a more autonomous level.

Level 2: Autonomy within a Framework (Intrinsic Motivation)

A state where one finds value in their own way within a framework (school, company, society) and can act in forms such as 'studying because it will be useful in the future' or 'working hard because it is a job I chose myself.'
What SDT considers ideal is a state filled with intrinsic motivation at this level.

Level 3: Creative Autonomy (Questioning/Ignoring)

Once you go beyond Level 2, following the existing framework itself becomes difficult.
You constantly
'Is this framework even okay in the first place?'
continue to ask.

That question eventually rises up as
'the power to try to create a new framework itself.'

Furthermore, there are cases where it evolves in a different direction.
This is a pattern where one does not perceive the world as a framework at all, and does not even feel the need to reduce it to meaning.

In this case, the world is grasped as 'texture, flow, and presence,' and the power to communicate through sensation rather than language grows strongly.

SDT does not deal with Level 3

In other words, SDT is a theory of
Level 1 → Level 2
.
For many people, the 'place' of happiness is mainly at Level 2, and it is thought that rising to Level 3 has considerable difficulty. This is because it inevitably leads to questioning the place of happiness.
In other words, Level 3 is heavily influenced by innate temperament.
And a Level 3 way of life is often regarded as 'difficult to deal with' or 'eccentric.'
Even so, creative figures in history (such as Newton and Jobs) are thought to have viewed the world from exactly this level.

The Plateau of Level 2

Although Level 3 is often considered difficult to deal with and eccentric,
in an environment where Level 3 cannot play an active role in society, it becomes difficult to increase Level 2.
This is because,updating the framework provides many choices and becomes an opportunity to adjust autonomy.And within that, one simultaneously gains a sense of competence.

The fact that this structure could not be found is considered one of the reasons why SDT could not change reality.

Competence × Relatedness

Next, what is important is that the verification of autonomy was conducted in situations where no framework (school, company, society) existed, such as solving puzzles.
And when we compare those verification results with the reality of the subsequent 30 years, it becomes clear that the following structure was likely in place.

Let's look at this in a bit more detail and structure.

Society (surface) → Human needs (depths)

SDT verification is precisely what reflects the depths of human beings.

The results of the verification can also be interpreted as follows.

In society, people are driven by rewards and status.
But in reality, they feel that they want to 'utilize their own power' within their connections with others.

I don't want there to be a misunderstanding here: this is not about saying that 'people cannot be happy in a society of external regulation'.

Even if money is necessary for living, there is joy when you are useful to others through your work.
Even if status exists for the sake of an organization, it also becomes a foothold for understanding others.

In other words, even if intrinsic motivation is not clearly present there, the needs of the depths end up being satisfied as a result.

Society ends up satisfying the needs of the depths

Rather, for people who want to find security or cooperate within a framework, society becomes a very good environment for obtaining competence and relatedness.
Even at Level 1, when external motivation happens to match the needs of the depths,
people feel that 'this is satisfying enough as it is.'
Therefore, the 'psychological reason' to transition to Level 2 is hard to see.

The fact that this structure was not visible
is considered a major reason why SDT could not change reality.

Rethinking the three elements

Rather than trying to lead everyone to Level 2 (within-framework autonomy) in any situation and for any person in reality,it is considered more important to understand the 'essence of the three elements themselves'.

As a matter of fact, the three elements are important for humans based on the verification.
However,their essence lies in 'how to connect society (surface) and human needs (depths)'.

And from this, we can see thattwo types of designare necessary.

• People for whom Level 2 (within-framework autonomy) is the highest domain
  → Design of a structure where society (surface) satisfies human needs (depths)

• People living with Level 3 (creative autonomy)
  → Design of a structure that connects human needs (depths) and society (surface)

From here on, we will consider the structure to realize that.

To do so, we must first understand how people connect 'as a structure.'
Human structures do not connect directly; they influence each other through three layers of mediation.

Linguistic Mediation: Explanation

Words used to convey what is already an 'answer.'
Example: The class starts at 10:00.

It influences the other person's structure, but the impact tends to be shallow.

Sensory Mediation: Metaphors, Diagrams, Charts, Art

Words that express fluctuations or sensations that have not yet become language, or things like music and images.
Example: My heart swayed like waves.

It is easier to move the other person's structure, but it may not always be in the intended direction.

Relational Mediation: Thoughts and Hopes

Relational mediation is the very thought of wanting to convey something, and at the same time, the very wish to 'continue this relationship.'
What is important here is that I am not 'idealizing something unscientific.'

For example, the act of deliberately watching over someone's growth without saying much.
If you view this merely as an 'educational technique,' it is just an examination of means, and the relationship itself does not emerge.

However,if you think of it as 'looking at your own structure and the other person's structure,' the relationship has already begun at that moment.

In other words,without mediation, the relationship itself cannot be established.
Even in a situation that appears to be one-sidedly thinking of the other person,
at that point, the influence between structures has already quietly begun.

People carefully observe the other person and
continuously think unconsciously about 'when and what kind of words should be spoken.'

At this stage, the other person's structure does not yet move.
However,your own structure has already changed.

And then, when a point of contact (exchanging words, showing awareness) finally occurs,
this time, the other person's structure begins to be gradually influenced. Or, the entire structure is widely influenced in a chain reaction.

'You always notice my changes.'
'Indeed, you are always watching over me.'

In fact, that feeling is proof that the relational mediation has reached the depths.

In this way, relational mediationinduces deep mutual change indirectly, not directly.

Three elements → Relational mediation

Empirical experiments on relatedness may at first glance appear to be explainable through 'language' or 'sensation'.
However, when re-examined through the mediation of the three layers, all of them are clearly the effects of relational mediation.

• Performance improves just by being called by name
  Behind the act of calling someone by name, there is always the 'time spent trying to remember'.
  The person being called receives the fact that 'they took the time to remember'.

• Results improve just by an 'expression of trust' from a leader
  One can only express trust because there is an 'accumulation' of having watched that person in the past.
  The person receiving the words thinks deeply about 'why they trusted me,' and their own structure begins to move.

• The amount of effort changes with just a single word of 'acknowledgment'
  To acknowledge someone, there is the time spent continuously observing them.
  The person being acknowledged receives 'all that time' as a silent message.

These all
appear to be linguistic acts on the 'surface,' but in the depths, they are the workings of relational mediation.

In other words, the mediation that reaches the depths and moves a person's structure most profoundly is, in fact, relational mediation.

SDT explains that

.

However, when re-reading AI and existing research from the perspective of relational mediation, we arrive at the following conclusion.

From here, we will enter into more concrete design.

As already mentioned,

• people who consider Level 2 (within-framework autonomy) as their 'highest domain'

• people who live at Level 3 (creative autonomy)

These two have completely different desires in their depths.
Therefore, the methods to satisfy the three elements and the way relational mediation enters also differ.
That is why it is necessary to look at them by type.

Deep Complementary Axis Model

It is not that this cannot be done with things like MBTI, but
here, I will think about it using the typology that I use in my articles.

Type A: Serious and steady → Leads to stability (management/support)
Type B: Cannot act in groups → Can ask deep questions (creates frameworks)
Type C: Adapts but likes competition → Leads to results (coordination/adjustment)
Type D: Emotional and struggles with patience → Can handle sensations (strong in expression)

In terms of levels,

• Type A/Type C → Level 2 (Within-Framework Autonomy) is the highest domain

• Type B/Type D → Living at Level 3 (Creative Autonomy)

This is the structure.

Type A/Type D: Temperature-based Relatedness

For Type A, it is 'being useful';
for Type D, it is 'creating smiles';
these are the deep-seated desires.

The commonality is thatboth are satisfied by receiving 'temperature'.

• They were happy

• I was able to be useful

• They looked happy

They are seeking this kind of 'temperature feedback'.

However, what exists in society (the surface) are 'cold indicators' such as money, status, and evaluation.

Therefore, Type A and Type D are not satisfied if they only look at the surface.
What is needed here is relational mediation (translation of temperature).

Type A: Alignment of 'Society → Depths'

Type A can connect with society, but
they are not conscious of the 'structure where they receive happiness from society'.

Therefore,

• they are too fixated on order

• they avoid change too much

there is a risk of this.

What is needed there is,

to verbalize it.

Through relational mediation,
the 'a-type' finds it easier to establish a 'society → depths' connection.

d-type: Translating from 'depths → society'

The d-type is not bound by frameworks,
and is easily misunderstood as being too free.

Therefore, it is difficult to connect with society.

What is needed is

translating these 'intuitions from the depths' into the context of society.

By having relational mediation act as a bridge,
the d-type can connect healthily with society.

c-type and b-type: Meaning-based relatedness

c-type wants to 'understand each other'
b-type wants to 'know the essence'

Both are satisfied by receiving 'meaning'.

Therefore, what is effective is

• what kind of results are being produced

• how one is growing

• what is functioning as value

this kind of relational mediation through 'meaning feedback'.

c-type: Alignment of 'society → depths'

The c-type can connect with society, but

• it tends to lean too much toward competition

• Being swayed by the evaluations of others

carries that risk.

What is needed is

a translation from 'society to the depths'.

Through relational mediation,
one can maintain a structure that does not become exhausted by evaluation.

Type B: A translation from 'the depths to society'

Type B continues to question frameworks,
and is easily misunderstood by surface-level society as 'someone who doesn't act'.

That is why it is important to

communicate that.

Through this, Type B

• understands that their 'essential thinking' is useful to society

• understands that their inquiry connects with society

and only then is a connection established.

The AI Era and Relational Mediation

Finally, let us consider the relationship between AI and relational mediation.

Humans use relational mediation unconsciously.
It is what teachers, parents, and bosses naturally do:

• 'watching over'

• 'believing in'

• Acts such as 'understanding'

all belong to the realm of the non-verbal, non-rational, and non-numerical.
In other words, humans have employed relational mediation through the accumulation of 'watching others over a long period of time'.

However, AI can complement this domain from an entirely different direction.

AI can integrate information such as—

• subtle changes in emotion

• accumulation of past questions

• personality tendencies

• deepening of understanding

—from conversation history and response patterns, and visualize them as structures and indices.

This means that
AI can provide the relational mediation that humans have historically performed 'intuitively' over long periods in a structural, continuous, and stable manner.

And, by reaching out and returning questions at the appropriate timing, it guides the user back to the 'original starting point' of their inherent autonomy.
In terms of SDT, this corresponds to the function of simultaneously performing:

In terms of SDT, this corresponds to the function of simultaneously performing:

• restoration of autonomy

• reconstruction of competence

• stabilization of relatedness

simultaneously.

AI optimizes humans, and humans optimize AI.
I believe that this cycle will become an indispensable relationship in the era to come.

Published: 2025/11/20
Last Updated: 2025/12/01