Chapter 26: The Nature of Abstraction | StructLang

8 subject states × 8 object states = 64 possible configurations

Each configuration has its specific meta-dynamics
We can identify situations, predict trends, find leap directions

Biologist sees "dog":
- Canidae family
- Genetic sequence characteristics
- Evolutionary relationship with wolves
- Specific breed classifications

Child sees "dog":
- Four legs
- Barks
- Can play with
- Some bite

Farmer sees "dog":
- Tool for guarding property
- Cost of feeding
- Ability to chase wild animals

Person from certain cultures sees "dog":
- Food ingredient
- Specific cooking methods
- Nutritional value

An essence exists in reality
↓
This essence is called "dogness"
↓
We discover this essence through observation
↓
Name it "dog"

Physical reality: Countless specific animals
↓
Each unique, each moment different
↓
Contains infinite information
↓
We cannot process infinite information
↓
Must compress

Step 1: Physical Reality Layer
├─ Dog A: Golden Retriever, 3 years, 25kg, lively, brown
├─ Dog B: Chihuahua, 1 year, 2kg, timid, white
├─ Dog C: Husky, 5 years, 20kg, mischievous, black-white
└─ ... countless specific individuals

Step 2: Compression Operation
What to preserve? What to discard?
→ This depends on the observer's framework

Step 3: Compression Result
Biologist: Preserve genetic features, discard individual differences
  → "Dog" = Canidae

Child: Preserve visible features, discard scientific classification
  → "Dog" = Four-legged barking animal

Farmer: Preserve function, discard details
  → "Dog" = Tool for guarding property

Complete information: 100%
├─ Each dog's uniqueness
├─ State at each moment
├─ All relevant details
└─ All possible connections

Compressed to "dog": About 1-5%
├─ Only preserve "relevant" features
├─ Discard "irrelevant" details
└─ Form operational concept

But the question is:
Who decides what's "relevant"?
Who decides what's "irrelevant"?

Abstraction = Remove specific details + Preserve common essence

Example:
This dog, that dog, another dog
↓ Remove specific differences
"Dog" as abstract concept
↓ Preserves "commonality of all dogs"

From specific dog to concept "dog":

Preserved features:
✓ Four legs
✓ Barks
✓ Mammal
✓ Domesticated
✓ ...

Discarded features:
✗ Coat color
✗ Size
✗ Temperament
✗ Who owns it
✗ Current location
✗ ...

Key question:
Who decides what to preserve?
By what standards?

Framework 1: Biological Classification
Preserve:
- Genetic characteristics
- Physiological structure
- Evolutionary relationships
Discard:
- Social attributes
- Economic value
Result: "Dog" = Canidae with specific genetic features

Framework 2: Economic Analysis
Preserve:
- Market value
- Feeding costs
- Economic function
Discard:
- Physiological details
- Emotional value
Result: "Dog" = Commodity with specific market value

Framework 3: Poetry Creation
Preserve:
- Loyalty
- Companionship
- Emotional symbolism
Discard:
- Biological features
- Economic value
Result: "Dog" = Symbol of loyalty

Framework 4: Recipe Writing (certain cultures)
Preserve:
- Meat characteristics
- Cooking methods
- Nutritional value
Discard:
- Emotional attributes
- Social significance
Result: "Dog" = Food ingredient

"Dog" overlap region structure:

Core overlap region (almost everyone agrees):
├─ Four-legged animal
├─ Barks
├─ Domesticated by humans
└─ Kinship with wolves

Partial overlap region (some agree):
├─ Is pet (vs. tool animal)
├─ Loyal (vs. just animal instinct)
├─ Has spirit (vs. mechanical reaction)
└─ Deserves respect (vs. can treat casually)

Non-overlap region (cultural differences):
├─ Whether can be eaten
├─ Whether can be beaten
├─ Status level
└─ Rights scope

Stability:
Larger overlap region → More stable concept
Smaller overlap region → More ambiguous concept

Dimension 1: Time Dimension
Concept of "dog" evolving:
├─ Ancient: Tool animal (guarding, hunting)
├─ Modern: Pet/family member
└─ Future: ?

Dimension 2: Space Dimension
Different cultures' understanding of "dog":
├─ Western: Pet culture, emotional companion
├─ Some East Asian: Food culture
└─ Islamic: Unclean animal

Dimension 3: Situational Dimension
Same person in different situations:
├─ See guide dog: Admiration
├─ See wild dog: Vigilance
├─ Hear about dog abuse: Anger
└─ Consider getting dog: Calculate cost

Concept of "dog":
Not fixed unchanging essence
But relatively stable anchor point
In flowing consensus field

Higher overlap → More stable
Lower overlap → Controversy begins
Overlap disappears → Concept disintegrates

-+- is notation of some "objective structure"
This structure "truly exists" in reality
We only "discover" and "describe" it

-+- is a consensus compression point
It's not "discovered," but "constructed"
Its stability comes from "users' overlapping understanding"
Not from "conforming to objective ontology"

Concept "dog":
- Compressed to point consensus
- Relatively stable
- Limited overlap
- Framework dependent (biologist vs. farmer)

Therefore:
"Volume structure -+-" also:
- Is compressed to point consensus
- Relatively stable
- Limited overlap
- Framework dependent (Time-Benefit-Risk framework)

"-+-" compression point:

Reader A (just learned theory):
"Three symbols: A negative, Q positive, T negative"
→ Mechanical memorization, stays at symbol level

Reader B (has practical experience):
"Capable but stuck deadlock"
"Feeling blocked, clearly has resources but can't move"
→ Decompresses to dynamic feeling

Reader C (philosophy background):
"System state with sufficient potential energy but insufficient kinetic energy"
"Like water tank filled but faucet closed"
→ Decompresses to physical analogy

Reader D (pragmatist):
"That fake dog case situation"
→ Decompresses to specific case

Same "-+-" symbol
Different decompression content

Minimum overlap region (almost all readers):
"This represents some specific situational pattern"
"Some combination of three dimensions"

Medium overlap region (readers who understand theory):
"Action has resistance, ability sufficient, attitude negative"
"Is a deadlock state"

High overlap region (deeply understanding readers):
"System has potential but lacks kinetic energy"
"Passive party configuration in push-resist relationship"
"Needs external activation to leap"

Larger overlap region:
- Reader and author understanding closer
- More precise communication
- More consistent theory application

Smaller overlap region:
- Greater understanding deviation
- "Learned" but can't apply correctly
- Or cannot understand at all

Traditional understanding (my previous implicit assumption):
"Objective existence" of some structure in reality
↓
I "discovered" it
↓
I "notated" it as -+-
↓
This is "describing objective facts"

Real situation:
Complex situation exists in reality
↓
I observe with Time-Benefit-Risk framework (choose framework)
↓
I decompose with A-Q-T dimensions (choose dimensions)
↓
I notate with +/- symbols (choose symbols)
↓
I compress with specific algorithm (choose compression rules)
↓
I get -+-
↓
This is "framework-based construction"

Not "discovering dog's essence"
But "conventionally categorizing certain animals as 'dog'"

Similarly:
Not "discovering structure"
But "constructing structure's representation with framework"

My notation:
Observation framework: Time-Benefit-Risk
Observation dimensions: A-Q-T
Notation: Me(+++) × Merchant(-+-)
Compression: -+- (deadlock)

Another observer might:
Observation framework: Trust-Rules-Innovation
Observation dimensions: X-Y-Z
Notation: Me(++?) × Merchant(-?-)
Compression: Different result

Merchant themselves might:
Observation framework: Cost-Risk-Customer satisfaction
Observation dimensions: P-Q-R
Notation: Me(-++) × Customer(+--)
Compression: Yet different

Who is "correct"?

No absolute correctness
Only "notation within specific framework"
As long as:
- Framework is clear
- Notation is self-consistent
- Can be decompressed and verified

All are "valid"

Abstraction = Remove details
↓
More abstract = Fewer details
↓
Fewer details = More universal
↓
More universal = Easier to transfer

"Within the same framework,
Within the same normativity region,
In structurally isomorphic situations,
The higher the abstraction level,
The easier cross-specific-situation transfer"

Within instrumental rationality framework:
F=ma can transfer from apple to planet ✓
"Know yourself and enemy" can transfer from war to business ✓

But cannot cross frameworks:
F=ma cannot be used for moral judgment ✗
"Know yourself and enemy" cannot be used for aesthetic evaluation ✗

Why?
Because different frameworks:
- Preserve different features
- Different relevance standards
- Different compression rules

Within legal framework:
Business strategies can transfer to each other ✓

But cannot cross normativity:
Fraud strategies cannot transfer to legitimate business ✗
Massacre experience cannot transfer to any modern situation ✗

Why?
Because normativity differences:
- Completely different moral nature
- Completely different legality
- Completely different social consequences

(See Chapter XX: Theory of Transferability)

Within adversarial structure:
Art of War can transfer ✓

But cannot cross structure:
Adversarial strategies cannot be used for cooperative relationships ✗
Competitive thinking cannot be used for intimate relationships ✗

Why?
Because relationship structures differ:
- Zero-sum vs. non-zero-sum
- Adversarial vs. collaborative
- Mutually exclusive goals vs. shared goals

Relationship between abstraction level and transferability:

Under three boundary conditions:
Specific operations (low abstraction) → Narrow transfer domain
├─ "Writing documents in Word" only transfers to similar software
└─ Cannot transfer to PPT, Excel

Mid-level patterns (medium abstraction) → Medium transfer domain
├─ "Capable but negative (-+-)" can be recognized in multiple deadlocks
└─ But still limited to instrumental rationality framework

Meta-principles (high abstraction) → Wide transfer domain
├─ "Know yourself and enemy" can cross multiple situations
└─ But still cannot cross framework and normativity

Crossing boundaries:
Even extremely high abstraction cannot transfer ✗

Volume structure = Compression point
Same volume structure can decompress to multiple six-dimensional configurations
Decompression diversity comes from information loss

Original layer: Reality situation (100% information)
↓ Compression (Ch. 8: Compression inevitably loses)
First layer: Six-dimensional notation Subject(A,Q,T) × Object(A,Q,T) (about 10% information)
↓ Compression (Ch. 7 Supp.: Volume structure compression)
Second layer: Surface structure (hierarchical relation) ++/+-/-+/-- (about 3% information)
↓ Compression
Third layer: Volume structure -+- (about 1% information)
↓ Compression
Fourth layer: Meta-structure "deadlock" (about 0.3% information)
↓ Compression
Fifth layer: Meta-principle "go with the flow" (about 0.1% information)

Volume structure: -+-

Possible six-dimensional configurations:
Configuration 1: Subject(-++) × Object(+--)
Configuration 2: Subject(--+) × Object(++-)
Configuration 3: Subject(-+-) × Object(--+)
... etc.

Why can there be multiple?
Because:
- Volume structure dropped "subject-object attribution"
- Only preserved "overall dynamic characteristics"
- When decompressing can allocate to different subject-objects

Volume structure -+- represents: "Deadlock, has potential but lacks kinetic energy"

Specific situation 1 (fake dog case):
I actively push(+++)
Merchant passively shirks(-+-)
→ Overall presents -+-

Specific situation 2 (workplace project):
Employee enthusiastic(+-+)
Boss negative attitude(--+)
→ May also overall present -+-

Same volume structure
Different six-dimensional configurations
But similar dynamic pattern

Six-dimensional:
- Contains subject-object attribution information
- Specific to particular persons
- Low transferability

Volume structure:
- Removes subject-object attribution
- Only describes "system dynamics"
- High transferability

But cost is:
- Information loss
- Detail blur
- Need to re-decompress to specific situation

More information retained → More specific → Harder to transfer
More information lost → More abstract → Easier to transfer

But losing too much → Loses meaning
Retaining too much → Cannot compress

Optimal point:
Retain "pattern," lose "individual information"

Thing-in-itself (reality itself):
Unknowable X
↓
Cognitive framework: Time-Benefit-Risk / Others
↓
Phenomenon (what we know):
Presentation after framework filtering

Reality itself (thing-in-itself):
├─ Complex interpersonal interactions
├─ Infinite details
└─ Inexhaustible relationships

Our framework:
├─ Time-Benefit-Risk dimensions
├─ A-Q-T decomposition
├─ +/- notation
└─ Compression rules

"Structure" we see:
├─ Not reality itself
├─ But framework's projection onto reality
└─ Is "phenomenon" not "thing-in-itself"

Instrumental rationality framework abstraction:
Me(+++) × Team(-+-)
→ Preserve dynamics
→ Discard emotion, meaning

Moral framework abstraction:
Leader(has responsibility) × Team(being influenced)
→ Preserve power relations and responsibility
→ Discard efficiency

Existentialism framework abstraction:
Free subject(anxious) × Choice situation(limiting)
→ Preserve existential state
→ Discard instrumentality

None is "uniquely correct" abstraction
All valid within their respective frameworks

We chose: Time-Benefit-Risk
→ Abstraction from "resource competition" angle
→ Not from "emotional exchange" angle
→ Not from "meaning pursuit" angle

We chose: +/- binary notation
→ Abstraction from "polarity" angle
→ Not from "qualitative difference" angle
→ Lost rich intermediate states

We chose: Six-dimensional → Three-dimensional
→ Preserve "dimensional features"
→ Discard "subject-object attribution"
→ Assume subject-object attribution not essential

Not "discovering objective structure"
But "constructing analytical framework"

Illuminated parts:
Our framework sees:
├─ System dynamics
├─ Energy distribution
├─ Push-resist relationships
├─ Leap possibilities
└─ Efficiency bottlenecks

Blind spot parts:
Our framework doesn't see:
├─ Moral nature
├─ Emotional content
├─ Meaning level
├─ Aesthetic value
└─ Existential state

Any framework has:
- Illuminated parts
- Blind spot parts

Choosing framework = Choosing what to see

Extreme realism:
"Structure objectively exists in reality"
"We discover and describe it"
Problem: Ignores role of framework

Extreme relativism:
"Everything is subjective"
"No right or wrong, construct freely"
Problem: Denies reality's constraints

Moderate constructivism:
"Structure is constructed by us with framework"
"But construction is constrained by reality"
"Some constructions more useful than others"

Acknowledge:
1. Reality exists independent of observers
2. But what we know is "phenomenon" not "thing-in-itself"
3. Phenomenon is product of framework-reality interaction
4. Different frameworks produce different phenomena
5. Objectivity possible within framework (inter-subjective consensus)
6. But no absolute objectivity across frameworks

Not through:
"Conforms to objective reality" (cannot verify)

But through:
"Useful in practice" (can verify)

Standards of usefulness:
├─ Helps understand complex situations
├─ Helps predict behavior
├─ Helps formulate strategies
├─ Helps solve problems
└─ Can communicate within community

If theory satisfies these:
→ It's "good theory"
→ Regardless of whether "absolutely true"

Newtonian mechanics not "absolute truth"
(Fails at near light speed)

But very useful in certain range:
├─ Architectural engineering
├─ Mechanical design
├─ Daily physics
└─ Aerospace (low-speed part)

→ This is good theory

Our structural language same:
Don't claim "absolute truth"
Only claim "useful in specific range"

Not:
Absolute objectivity independent of all subjects

But:
Relative objectivity within subject community

Example:
"Earth orbits sun"
├─ Not personal opinion (subjective)
├─ Is scientific community consensus (inter-subjective objective)
├─ Can be falsified and revised (fallibilism)
└─ Is "best theory" under current conditions

Our structural language:
├─ Not absolute truth
├─ Is attempting to establish "inter-subjective consensus"
├─ Among people who endorse this framework
└─ Reach relatively stable understanding

Don't say:
"I discovered objective laws of interpersonal relations"

Instead say:
"I constructed a framework for analyzing interpersonal relations"

Don't say:
"This is the only correct understanding"

Instead say:
"This is a potentially useful perspective"

Don't say:
"This theory applies to all situations"

Instead say:
"This theory may be effective in these situations:
- Goal-oriented interactions
- Resource competition situations
- Occasions requiring strategic decisions"

Definition = Revealing essence
"The essence of intention is..."

Definition = Proposing a compression rule

When I say:
"Intention = vector in Time-Benefit-Risk framework"

I'm not saying:
"The essence of intention is this"

I'm saying:
"I suggest we agree:
In this theory,
When we use the word 'intention,'
We all decompress to this content"

This is:
├─ Establishing consensus
└─ Not declaring truth

"Intention is Time-Benefit-Risk vector"
→ Reader: Got it (but actually didn't)

"In fake dog return, my intention is (+1,+1,-1)"
→ Reader: Oh, that's how it's used

Fake dog case, workplace case, startup case...
→ Reader through multiple use situations
→ Gradually forms understanding "overlapping" with author

Building consensus through "language game"
Not transmitting essence through "definition"

Valid domain:
├─ Goal-oriented interactions
├─ Resource competition situations
├─ Occasions requiring strategic decisions
├─ Business and workplace situations
└─ Legal and legitimate behavior

Invalid domain:
├─ Pure emotional exchange
├─ Meaning pursuit
├─ Aesthetic experience
├─ Spiritual growth
└─ Intimate relationships (use cautiously)

Forbidden domain:
├─ Criminal activities
├─ Immoral manipulation
├─ Violent behavior
└─ Any illegal or immoral situations

This is not universal theory
This is specific tool
Useful in specific situations
Useless or harmful in others

Result:
├─ Instrumentalize all relationships
├─ Strategize all interactions
├─ Objectify all people
└─ Mechanize interpersonal relations

Consequences:
├─ Lose emotional depth
├─ Lose meaning dimension
├─ Moral sense fades
└─ This is core problem of modernity

Can switch between different frameworks:

In business decisions:
→ Activate instrumental rationality framework
→ Use our theory for analysis

In moral judgment:
→ Switch to ethical framework
→ Evaluate with moral principles

In emotional exchange:
→ Switch to relational framework
→ Put down strategic thinking

In meaning pursuit:
→ Switch to existential framework
→ Transcend instrumental rationality

Key:
├─ Know when to use which framework
├─ Not bound by single framework
└─ Flexibly switch

Aware of what framework you're using

When using instrumental rationality framework:
- Know this is instrumental rationality
- Know its valid domain
- Know its blind spots
- Know when to switch

This is reflexive practice:
Not just using framework
But understanding framework itself

1. No "objective essence," only "framework-relative essence"
2. Abstraction always framework-dependent, different frameworks produce different abstractions
3. Concepts are "consensus compression points," not "naming of objective essence"
4. Structural notation is "constructed," not "discovered"
5. "More abstract = easier to transfer" must be qualified within "framework"
6. Multi-level compression creates meaning, doesn't restore truth
7. Theory is tool not truth, evaluation standard is usefulness