Originality: Can Anything Truly New Be Created?

Nothing comes from nowhere.

Every poem uses an existing language.

Every invention uses existing matter.

Every thought emerges from a brain shaped by:

• genes,
• experience,
• culture.

So what does it mean to create something original?

Can anything truly new be created?

Novelty vs Originality

Novelty means:

not previously encountered.

Originality usually means something stronger:

novel in a meaningful way.

A random number can be novel.

It is not necessarily original in the creative sense.

Ex Nihilo Creation

The strongest notion would be:

creation from nothing.

But human creativity does not work this way.

Creators transform existing materials.

Recombination

Many innovations combine existing elements.

Examples:

smartphone = phone + computer + camera + network.

Recombination is not trivial.

The important part is discovering a useful combination.

Combinatorial Explosion

If there are many existing components, the number of possible combinations becomes enormous.

Even recombination can explore genuinely unexplored space.

Newness does not require new primitives.

Language Example

A finite vocabulary can generate sentences never previously spoken.

Grammar gives:

finite rules, potentially unbounded expression.

Originality can emerge from composition.

Music

Western tonal music uses a limited set of notes.

Yet composers produce enormous variety.

Constraint does not eliminate novelty.

It makes structured novelty possible.

Chess

Chess has fixed rules.

Players still discover new strategies.

Originality can exist inside a closed formal space.

Search Space

A creative domain can be modeled as a space of possibilities.

Most points are:

• useless,
• boring,
• incoherent.

Originality involves finding rare meaningful regions.

Exploration

To create something new, a system must explore beyond familiar regions.

This can happen through:

• mutation,
• analogy,
• random variation,
• deliberate search.

Different mechanisms can produce novelty.

Randomness

Randomness can introduce variation.

But random output alone is not creativity.

It needs:

• selection,
• structure,
• evaluation.

Randomness supplies possibilities.

Intelligence filters them.

Mutation and Selection

Evolution creates novelty without foresight.

Mutation generates variation.

Selection preserves some variants.

Complex innovations accumulate.

This proves that novelty does not require conscious intention.

Emergence

New higher-level properties can arise from old components.

Water has properties not obvious from one molecule.

Similarly, creative systems can produce patterns not explicitly present in any component.

This is emergent novelty.

Is Emergence Really New?

A reductionist may say:

the outcome was always implicit in lower-level laws.

But “implicit” is doing a lot of work.

If the result cannot be feasibly predicted, it is new at the level of knowledge and organization.

Ontological Novelty

A stronger claim is ontological novelty:

something genuinely new comes into existence.

Examples might include:

• life,
• consciousness,
• institutions.

Whether these are fundamentally new or reorganized physical states depends on metaphysics.

Epistemic Novelty

A weaker and clearer notion is epistemic novelty:

the pattern was previously unknown.

This is enough for most creativity.

Historical Novelty

An idea may be:

new to one person

or:

new to humanity.

These are different standards.

Boden’s P-creativity and H-creativity capture this distinction.

Independent Discovery

Calculus was developed independently by Newton and Leibniz.

If one had already discovered it, was the other’s work unoriginal?

Independent generation still demonstrates creative achievement.

Originality is not only priority.

Rediscovery

A student can independently rediscover a theorem.

It is not historically new.

But psychologically it may be deeply creative.

Influence

No human creator is isolated.

Writers inherit:

• vocabulary,
• genres,
• myths.

Scientists inherit:

• concepts,
• instruments,
• theories.

Originality grows inside tradition.

The Myth of the Lone Genius

Culture often imagines originality as sudden inspiration from one exceptional mind.

In reality, innovation usually depends on:

• communities,
• predecessors,
• accumulated tools.

Creativity is socially scaffolded.

Standing on Shoulders

Newton’s famous metaphor of standing on the shoulders of giants captures cumulative knowledge.

New insight often becomes possible because old structure already exists.

Dependency does not eliminate originality.

Constraint and Originality

Paradoxically, constraints can increase creativity.

A sonnet has:

• meter,
• rhyme,
• form.

The challenge of satisfying constraints creates structured search.

Problem Reframing

Some of the most original acts do not solve the given problem.

They change the problem.

This is transformational creativity.

New Representation

A new representation can make previously difficult patterns visible.

Examples include:

• Cartesian coordinates,
• periodic table.

The novelty lies in reorganizing the conceptual space.

Conceptual Innovation

Inventing a concept such as:

gene, algorithm, entropy

creates a new way of grouping reality.

Conceptual originality changes what can be thought.

Naming Creates Cognitive Objects

A useful name makes a pattern easier to:

• discuss,
• compare,
• manipulate.

Representation can create new cognitive structure.

The world may be old.

The conceptual object is new.

Mathematics

Do mathematicians invent or discover mathematical objects?

If mathematical Platonism is true, the objects already exist abstractly.

Then creativity is discovery.

If mathematics is constructed, invention plays a larger role.

Originality depends on ontology.

Scientific Discovery

A new law of nature is not created by the scientist.

The law, if real, already governed nature.

The originality lies in:

• representation,
• explanation,
• discovery.

Creation and discovery overlap.

Art

A painting did not exist before it was made.

Its physical pattern is genuinely new.

But its:

• pigments,
• styles,
• motifs

have histories.

Originality exists within continuity.

Information-Theoretic View

A sequence can be algorithmically incompressible and therefore highly unpredictable.

But incompressibility does not equal meaningful originality.

Random noise can have high information content.

Meaning matters.

Compression and Creativity

A powerful creative insight can compress many facts into one concept.

Einstein’s equations unify phenomena.

A good metaphor compresses relationships.

Originality can mean discovering a better representation.

Generative Systems

A generative system contains rules capable of producing outputs never explicitly stored.

This is common in:

• grammar,
• mathematics,
• neural networks.

New output does not require new rules each time.

Determinism Objection

If the universe is deterministic, was every creative work fixed from the beginning?

Even if yes, originality need not disappear.

Deterministic processes can generate:

• complexity,
• novelty,
• unpredictability.

Causal determination is not repetition.

Predictability vs Determination

A deterministic process may be impossible to predict without running it.

This is computational irreducibility.

The future can be determined and still epistemically new.

Computational Irreducibility

Stephen Wolfram popularized the idea that some systems cannot be shortcut.

To know what happens, you must effectively run the process.

If so, novelty unfolds through time even under fixed rules.

Chaos

Chaotic systems are deterministic but highly sensitive to initial conditions.

Their detailed future can be practically unpredictable.

Again:

determinism ≠ predictability.

Algorithmic Generation

A short program can generate an enormous complex structure.

The Mandelbrot set comes from a tiny rule.

Simple origins do not imply trivial outcomes.

Can a Program Surprise Its Programmer?

Yes.

Programmers routinely build systems whose exact outputs they cannot predict.

The program’s rules are known.

The consequences are not.

This is especially common in:

• machine learning,
• evolutionary computation.

Is That Real Originality?

If originality requires that the creator not foresee the exact result, then generative systems can qualify.

But if originality requires autonomous goals, the criterion becomes stricter.

Definitions matter.

Training Data Objection

A model trained on previous works may seem derivative.

But human creators are also trained on previous works.

The key issue is not whether influence exists.

It is whether the system can transform structure into genuinely new configurations.

Statistical Novelty

A machine can produce combinations not present in training.

This is statistical novelty.

Whether the combination is culturally significant is a separate question.

Semantic Novelty

Stronger originality may involve a new idea that changes interpretation.

This requires not just new form, but new meaning.

Machines are less clearly established here.

Goal Novelty

An even stronger form would be:

inventing a new goal.

Most artificial systems optimize goals supplied by humans.

Autonomous goal formation would deepen claims of originality.

Value Creation

Originality often matters because it creates new value.

A new theorem changes mathematics.

A new technology changes possible action.

A new artwork changes culture.

Originality is relational.

Audience

Something can be original to one audience and familiar to another.

Novelty depends on a reference class.

No judgment of originality is context-free.

Simultaneous Invention

Technological history contains many simultaneous discoveries.

When prerequisites mature, several people may reach similar ideas.

Originality can be distributed across a historical moment.

Adjacent Possible

Stuart Kauffman popularized the idea of the adjacent possible:

new possibilities become reachable when current structures exist.

Innovation expands the space of what can happen next.

Innovation Creates Possibility

The invention of:

writing

made possible:

libraries, contracts, software source code.

New artifacts create new future search spaces.

Creativity changes the adjacent possible.

Recursive Creativity

Creative systems build tools for future creativity.

Humans invented:

• language,
• notation,
• computers.

These tools then expand what humans can create.

Creativity bootstraps itself.

AI as Creativity Amplifier

AI may similarly expand the space available to human creators.

It can generate:

• variants,
• drafts,
• hypotheses.

The system changes the creative environment.

Authorship

If output emerges from:

• model,
• prompt,
• training data,
• human curation,

who is the author?

Traditional authorship concepts may not map cleanly onto distributed generation.

Originality and Ownership

Legal systems often tie originality to rights.

But philosophical originality and legal originality are not identical.

Law creates operational thresholds.

Metaphysics asks different questions.

Nothing Is Completely Independent

Every cause has prior causes.

Every idea has history.

If originality required absolute independence from all influence, nothing would ever be original.

That definition is useless.

A Better Definition

A more useful notion of originality is:

the production of a meaningful structure not already explicitly available to the producing system or community.

This allows novelty within causal continuity.

Originality Without Magic

Creation does not need:

• supernatural inspiration,
• ex nihilo generation.

It can emerge from:

• recombination,
• abstraction,
• search,
• transformation.

Mechanism does not destroy originality.

The Philosophical Lesson

Truly new does not have to mean:

uncaused.

It can mean:

a new organization, a new concept, a new possibility, a new cultural pattern.

Originality is compatible with continuity.

The future can contain structures that did not exist before, even if they arise from old rules.

The Next Question

Modern generative AI makes these questions impossible to avoid.

Large language models generate:

• text,
• code,
• arguments,
• images through related multimodal systems.

What exactly are these models?

Why did they become so capable?

The next essay turns to:

Generative AI and Large Language Models.