Description, Explanation, and Function

A scientist observes:

when people report seeing red, neural pattern X appears.

This is a useful discovery.

But what exactly has been achieved?

Have we:

• described consciousness,
• explained consciousness,
• identified its function?

These are different intellectual tasks.

Confusing them creates many false solutions.

Description

A description states:

what happens.

Example:

water boils near 100°C at standard atmospheric pressure.

This records a regularity.

It does not necessarily explain it.

Explanation

An explanation tells us:

why or how the phenomenon occurs.

For boiling, molecular and thermodynamic theory explains the transition.

Explanation connects the observation to a mechanism or law.

Function

A function tells us:

what a process does

or:

what role it plays.

For example:

the heart pumps blood.

This functional statement is different from:

the heart is made of cardiac muscle.

Three Questions

For any phenomenon, ask separately:

1. What is it like or what happens?
2. How is it produced?
3. What does it do?

These may require different answers.

Consciousness Makes the Difference Vivid

Suppose conscious access allows information to:

• guide speech,
• influence planning.

That tells us something about function.

It may not explain why the process has subjective experience.

Correlation as Description

If neural activity N correlates with conscious experience C:

[ N \leftrightarrow C ]

we have found a relationship.

This is scientifically important.

But correlation is not yet causal explanation.

Mechanism

A mechanistic explanation identifies:

• parts,
• operations,
• organization.

Example:

photoreceptors transduce light, signals move through retinal circuits, visual cortex processes patterns.

This explains aspects of vision.

Mechanism and Experience

Even a complete mechanism for:

• discrimination,
• report,
• memory

may leave the question:

Why does this processing feel like anything?

That is where the hard problem emerges.

Functional Explanation

A functional explanation says a trait exists or is organized because of the role it performs.

In biology, function often connects to:

• survival,
• reproduction,
• regulation.

Example: Pain

Pain has functions.

It can:

• protect injured tissue,
• motivate avoidance,
• support learning.

These functions explain why pain-like control systems may be adaptive.

But Does Function Explain Hurt?

A critic asks:

Why could those protective functions not occur without the unpleasant feeling?

This separates:

pain behavior

from:

pain phenomenology.

Teleology

Function language sounds purposeful.

“The eye is for seeing.”

Biology explains this without future causation.

Natural selection shaped structures because ancestral variants had consequences.

Selected Effect Function

One theory says a biological trait’s function is the effect for which it was selected.

The heart’s function is pumping blood.

Its sound is an effect, but not its biological function.

Causal Role Function

Another notion defines function as the role a component plays in a system.

The heart’s causal role includes circulating blood.

This does not require evolutionary history.

Why Function Concepts Matter

Different meanings of “function” can create confusion.

A neural process may have:

• current causal role,
• evolutionary function.

These are not always identical.

Marr’s Levels

David Marr distinguished three levels of analysis.

Computational level

What problem is being solved?

Algorithmic level

What representations and procedures are used?

Implementational level

How is it physically realized?

These levels clarify cognitive explanation.

Example: Vision

Computational:

infer structure in the environment.

Algorithmic:

use feature extraction and inference.

Implementational:

neural circuits in visual pathways.

Each answers a different question.

Consciousness Across Marr’s Levels

For consciousness, we might ask:

Computational: What role does conscious access serve?

Algorithmic: What information-processing architecture produces it?

Implementational: What neural mechanisms realize it?

But phenomenal consciousness may raise an additional question.

Is Phenomenology a Fourth Question?

Some philosophers think the standard levels leave out:

Why is there experience?

This is not obviously a computational or implementational question.

It is the hard problem.

Explanation by Reduction

One strategy explains a higher-level phenomenon by reducing it to lower-level mechanisms.

Temperature becomes:

average kinetic behavior of molecules.

Could consciousness be reduced similarly?

That is a central hope of physicalism.

Successful Reduction

A reduction is satisfying when it shows why the higher-level properties follow from lower-level structure.

Example:

gas pressure follows from molecular collisions.

The bridge is intelligible.

Explanatory Gap

With consciousness, critics argue the bridge remains obscure.

Even knowing every physical fact, why should:

neural state N

equal:

experience E?

This is the explanatory gap.

Identity Can End Explanation

Sometimes science discovers identities.

Water = H₂O.

Once the identity is established, we do not ask:

Why does H₂O additionally produce water?

They are the same thing.

Consciousness Identity Theory

A physicalist might argue similarly:

experience E = brain state N.

Then no extra production relation is needed.

The challenge is explaining why the identity should be believed and why it feels conceptually surprising.

Lightning Analogy

Lightning was once mysterious.

Science identified it with electrical discharge.

Perhaps consciousness will undergo a similar conceptual reduction.

This is an optimistic physicalist analogy.

Why Critics Resist the Analogy

Lightning has no private qualitative character.

We can understand electrical discharge entirely from third-person concepts.

Consciousness seems to include an irreducible first-person fact.

That makes the analogy controversial.

Explanation by Deduction

In the deductive-nomological model, an explanation derives an event from:

• laws,
• initial conditions.

But many scientific explanations are not simple deductions.

Mechanisms and causal models matter.

Causal Explanation

A causal explanation identifies:

what produces the phenomenon.

If stimulating region R causes experience E, that is stronger than correlation.

Still, causal production may not explain phenomenal character.

Intervention

Modern causal science asks:

What happens if we intervene?

If changing N systematically changes C, N is causally relevant.

This strengthens the mind–brain link.

Necessary vs Sufficient

A neural mechanism may be:

• necessary,
• sufficient,
• neither alone.

Careful experiments distinguish these.

Consciousness likely depends on distributed systems.

Constitutive vs Causal Explanation

Suppose molecules constitute water.

They do not cause water as an external effect.

Likewise neural processes may constitute consciousness rather than cause it.

This distinction matters.

Constitution

A constitutive explanation tells us what a phenomenon is made of or realized by.

Example:

a traffic jam consists of interacting cars.

No separate “jam substance” is added.

Consciousness as Constitution

Physicalists may say:

consciousness is an organized neural process.

The neurons do not produce a second thing.

Their organized activity is the conscious state.

Why the Puzzle Remains

The critic replies:

Even if we call it constitution, why does this organization have experiential character?

Changing vocabulary does not automatically close the gap.

Functionalism

Functionalism identifies mental states by causal role.

For many cognitive states, this is powerful.

Belief can be characterized through relations to:

• evidence,
• inference,
• action.

Functionalism and Consciousness

If consciousness is defined functionally, then explaining:

• access,
• report,
• integration

may explain consciousness completely.

This is one major position.

Phenomenal Objection

Critics argue:

those functions could in principle occur without qualia.

If so, functional explanation leaves something out.

This is the zombie intuition again.

Multiple Realizability

Function can be realized by different substrates.

If consciousness is functional, then:

• biological,
• silicon,
• other systems

could all be conscious if they realize the same organization.

This has major implications for machine consciousness.

Biological Function

If consciousness evolved, it may have adaptive consequences.

Candidates include:

• flexible planning,
• integration,
• social modeling.

But evolutionary usefulness does not by itself explain phenomenology.

Epiphenomenal Challenge

If conscious feeling has no causal effect, how could natural selection favor it?

Selection acts on causal differences.

This creates difficulty for epiphenomenalism.

Reportability

A common functional property of consciousness is reportability.

If a person can say:

“I see red,”

the information influenced:

• language,
• memory,
• decision.

But report is not identical to experience.

Global Availability

Another function is broad availability.

Conscious information may be accessible to many systems.

This idea becomes central in Global Workspace Theory.

Integration

Consciousness also appears integrated.

Information from:

• vision,
• memory,
• goals

forms one coherent scene.

This motivates Integrated Information Theory.

Function Can Guide Theory

Even if function does not solve the hard problem, identifying function narrows candidate mechanisms.

Science progresses by decomposing questions.

Not every partial explanation must answer everything.

Explanatory Modesty

A good theory should state what it explains.

For example:

“This model explains conscious access.”

That is stronger than claiming:

“This model explains consciousness”

without qualification.

Label Inflation

Researchers sometimes attach the word:

consciousness

to mechanisms of:

• attention,
• working memory.

This can make empirical progress sound more metaphysically complete than it is.

Precision matters.

Operational Definitions

Experiments need operational definitions.

Example:

conscious perception = stimulus correctly reported.

This is useful.

But operational definition is not necessarily metaphysical identity.

Measuring vs Defining

A thermometer measures temperature.

It does not define what temperature ultimately is.

Likewise report can measure conscious access without exhausting consciousness.

Functional Equivalence

Two systems may perform the same function.

Does that imply the same experience?

Functionalists lean yes.

Qualia-focused theories allow no.

The dispute cannot be solved by behavior alone.

Explanation and Prediction

A theory that predicts new results is scientifically valuable.

But prediction and explanation are not identical.

A black-box model can predict without making the mechanism intelligible.

Black-Box Prediction

Suppose an AI predicts when someone will report awareness with 99.9% accuracy.

This is impressive.

Yet we may still not understand why those neural conditions produce experience.

Prediction can outrun explanation.

Compression and Explanation

A good explanation often compresses many observations into one mechanism or principle.

Newtonian gravity compressed planetary motion.

Explanatory power is not mere description length, but compression helps.

Understanding

An explanation should improve our ability to answer:

• why,
• what if,
• what would change it?

Causal and counterfactual structure matters.

What Would Count as Explaining Consciousness?

Possible standards differ.

A theory might need to:

• identify neural mechanisms,
• predict conscious states,
• explain functional roles,
• derive phenomenal structure.

Different philosophers demand different levels.

No Single Test

There is no universally agreed criterion for a complete consciousness theory.

This is part of why debates persist.

Scientific and philosophical standards overlap imperfectly.

The Risk of Mystery Inflation

We should not assume consciousness is permanently inexplicable.

Past mysteries have become ordinary science.

Conceptual difficulty today does not prove impossibility tomorrow.

The Risk of Premature Dissolution

We also should not declare the problem solved merely because correlates and functions are known.

A complete account should explain the target that generated the puzzle.

The Philosophical Lesson

Description tells us:

what happens.

Function tells us:

what it does.

Explanation tells us:

how or why.

For consciousness, all three matter.

But they should not be confused.

The Next Question

Now the distinction is clear enough to confront the deepest challenge directly.

Even if we explain every function of consciousness, one question appears to remain:

Why is there anything it is like to perform those functions?

That is:

The Hard Problem of Consciousness.