Chapter 2: Extended Cognition

Welcome to Last Minute Lecture.

This free chapter overview is designed to help students review and understand key concepts.

These summaries supplement not replaced the original textbook and may not be redistributed or resold.

For complete coverage, always consult the official text.

Welcome to the Deep Dive.

Today we're tackling one of the biggest philosophical showdowns in cognitive science.

We're asking this fundamental question.

Where does the mind actually stop?

Do our thoughts, our memories, all our reasoning processes, do they live exclusively inside the skull?

Or do they leak out?

Do they incorporate the world around us?

It really does sound like a question from a science fiction movie, doesn't it?

But this debate is, I mean, it's happening right now in philosophy and psychology and the implications are profound.

They reach far beyond the lab, you know, they go deep into how we define ourselves as thinking agents and even what we mean by the word mind.

Our source for this deep dive is a really meticulous breakdown of this whole debate.

It's from a chapter titled Extended Cognition by Julian Keverstein featured in the Oxford Handbook of 4E Cognition.

And so you, the listener, you're stepping into this field 4E cognition, which covers embodied, embedded, inactive and extended cognition, a field that's often described as a broad church.

A broad church.

I like that.

So lots of different even conflicting views under one roof.

Exactly.

It's a theoretical home to some wildly conflicting perspectives on the nature of thought.

And the author's argument here is crucial.

They're not just saying they're disagreements.

They contend that these intense fights within 4E cognitive science, they aren't just, you know, minor squabbles about how we design experiments.

They are fundamentally debates about, about ontology.

Ontology.

We're not just arguing over causal mechanisms.

We are arguing over the very nature of existence.

Specifically,

what it is for any process to count as cognitive.

Exactly.

And that's our mission today.

We're going to unpack these competing ideas.

We're essentially joining the search for the mark of the cognitive.

The mark of the cognitive.

Yes.

The distinguishing properties, whether they're functional or representational or dynamic, the properties a system must have for us to say, yes, that's genuinely mental.

So this mark is the dividing line.

If we can define it, we can finally settle the boundaries of the mind.

And that search, as I understand it, is going to take us through three major philosophical positions.

First, the conscious, more traditional view of the embedded mind.

Then the radical functionalist challenge of the extended mind.

And finally, an even more radical solution offered by something called dynamic extended cognition.

And we'll get there by navigating through functionalism, some famous thought experiments involving notebooks.

We'll talk about highly skilled bartenders and even a really crucial moment in a game of Tetris.

And all of it is focused on establishing the true boundaries of the self.

OK, let's unpack this by first establishing the two main opponents in this philosophical arena.

That would be the cautious embedded view and the boundary smashing extended view.

So we have to start with the historical baseline, right?

For centuries, I mean, you go all the way back to thinkers like Descartes.

It just felt utterly intuitive to view the mind and cognition as fundamentally inner phenomena.

It's the common sense view.

Exactly.

We take in the world through our senses.

We process it.

We remember.

We reason all internally.

And then we send commands out to the body to act.

Mental processes happen entirely inside the individual's head.

That's the traditional Cartesian legacy that cognitive science inherited.

Now the embedded theory or EMT, that's the first critical step away from this really strict interview.

It's a recognition that, you know, pure Cartesian isolation is impossible.

EMT acknowledges the deep and often surprising causal dependence of mental processes on our bodily interactions with the environment.

OK, so under EMT, the body and the world definitely matter.

My environment causes my cognition to unfold in a certain way.

If I'm in a quiet library, I can focus better.

The quiet library is causally influencing my internal thinking process.

Precisely.

The environment is indispensable.

But, and here's the crucial distinction that defines EMT.

The EMT advocate claims that the cognitive process itself is nevertheless wholly realized by systems and mechanisms located inside the brain.

OK, so the brain is the seed of cognition.

The world is just a powerful tool, a scaffold.

That's the key term, scaffolding.

EMT thinkers, people like Adams and Aizawa or Robert Rupert, they interpret cognition along very traditional lines.

It's all about internal computational rule -based operations carried out on these internal representational structures that carry information about the world.

They just recognize that our thinking is constantly scaffolded by all these external resources.

But, the thinking itself, that remains an internal event.

The most accessible example has to be using a calculator to divide a restaurant bill.

Without the calculator, the arithmetic might be, well, for me, impossible in my head.

The calculator is undeniably essential.

It scaffolds the mathematical thinking.

It makes the problem trivially easy.

But the actual cognitive computation reading the result, applying the rules of arithmetic to the input, is still fully constituted and realized inside my brain.

The calculator is a necessary external aid, but it's not a thinking part of my mind itself.

And if EMT is that cautious step away from Descartes, just acknowledging external causality, then the extended theory, EXT, is the philosophical explosion.

EXT completely rejects the Cartesian legacy and makes a much, much stronger claim.

And this is the constitutive claim.

We should probably slow down and really nail this distinction.

What does constitutive mean here compared to causal?

It's the core of the whole debate.

When something is causal, it means it influences a process.

It makes it happen faster or differently.

When something is constitutive, it means it literally makes up the thing we are talking about.

It is an integral, necessary part of the process itself, like how hydrogen and oxygen atoms are constitutive of water.

You can't have water without them.

So EXT argues that bodily actions in the environmental resources an agent acts upon can, under certain conditions, count as constituent parts of a cognitive process.

They don't just influence the mind.

They are part of the process that is the mind.

And modern life offers these compelling, if maybe a little unsettling, examples of this thorough integration.

The chapter mentions philosophers who describe using their smartphones not just for,

you but for their internal mental life.

David Chalmers described using his iPhone to idly call up words and images when his concentration slipped.

And the question becomes inescapable.

If that information is accessible and used just like internal memory, is it really separate?

Right.

Mobile phones, smartwatches, even a well -organized filing system.

They're so interwoven into our daily routine that EXT suggests these resources might now be thought of as parts of our minds, at least temporarily.

But the problem is deciding which category they fall into.

Are they merely external causes influencing our internal cognition, which is the EMT view, or are they literally part of the cognitive system itself, which is EXT?

To make this distinction more concrete, let's explore a classic pre -digital scenario, the expert bartender from a study in 1988.

Right.

So imagine a bartender who is just slammed with a long, complicated order for five different drinks, a Manhattan, a Martini, a Cosmo, a beer and a whiskey.

Now, instead of trying to hold all five steps in his internal working memory, what does he do?

He lines up the appropriate glasses, a tall glass for the beer, a Martini glass for the Martini, and so on, in the specific order the drinks were requested.

It's a simple spatial arrangement, but it makes remembering the order dramatically easier.

Some of the work of remembering is literally offloaded onto the environment.

Yes.

The line of glasses functions as an external store of information.

It's like an analog cue,

and it guides and controls the bartender's action.

It significantly lightens the information load on his internal working memory.

And that initial action of arranging the glasses is crucial.

David Kirsch has a term for this, right?

An epistemic action.

Exactly.

An epistemic action is an action you perform not to achieve a physical outcome directly, like shaking a drink, but to structure information in the environment to better fit the goals of the system.

In doing so, you simplify the internal cognitive task.

So you're actively manipulating the environment before the main task begins,

specifically to make that main task easier.

OK, so how does EXT interpret this versus the EMT interpretation?

Well, EXT looks at the situation and argues that the external structure, that spatial array of glasses, it works together with interperceptual processes, working memory, attentional processes, to form a unified, goal -directed, softly assembled cognitive system that brings about the behavior.

I like that term, softly assembled.

It suggests flexibility, not some fixed machine.

What defines this assembly?

It's a contrast with traditional fixed systems, where every component has a pre -specified fixed function.

Softly assembled cognitive systems are characterized by an interaction -dominant dynamics.

Interaction -dominant dynamics.

OK, what does that mean?

It means the components, the bartender's brain, his hands, the line of glasses, they mutually and continuously influence each other in real time.

So when the bartender looks at the row of glasses,

that perception immediately influences his planning system,

which dictates his next action, which then changes the state of the external glasses.

He removes one, for example.

And that new state feeds back instantly into his perceptual system.

They are constantly regulating each other.

That sounds like a very powerful argument for integration.

But it brings us right back to the core philosophical disagreement.

EXT says these external structures are constitutive parts of the cognitive process.

But EMT counters by arguing that they're only causal contributors, even if they are very tightly integrated.

And they warn specifically against what they call the coupling constitution fallacy.

The coupling constitution fallacy.

Yes.

The objection states, just because two things are tightly linked, interdependent, and mutually influencing each other, that is, coupled, that doesn't automatically mean the external component is constitutive of the mind.

My pencil and paper are tightly coupled to my writing process, but no one thinks the pencil is part of my brain.

That's a powerful line of resistance.

I mean, if a surgeon is using a finely tuned surgical robot, the robot is coupled to the surgeon's hand movements, but the robot is still a tool, right?

It's not part of the surgeon's mind.

Exactly, and this is where the metaphysical deadlock begins.

But what's so fascinating is that despite this deep philosophical divide, both EMT and EXT actually agree on the methodology of cognitive science.

How so?

Both camps agree on the practical task.

To explain problem solving, you often have to reference bodily actions on external structures.

Internal cognitive processes alone are often not sufficient to explain our complex goal -directed behaviors, like driving a car or playing Tetris.

And this points directly to Spravak's thesis, which the chapter references.

Spravak suggests that this whole dispute might be sort of toothless in the lab.

Yes, from an empirical perspective, the theories aren't genuinely competing.

A cognitive scientist can frame their findings in embedded terms.

The external structure caused the internal computation to finish faster.

Or they could frame it in extended terms.

The external structure was a constituent part of the overall computation.

And the explanatory value for predicting the behavior, it remains largely the same.

They're observationally equivalent.

So the empirical data can't decide the winner.

This means the real stake here isn't about how we explain the behavior.

It's about metaphysics.

It's about ontology.

And it forces us to ask, why does it matter?

Why does it matter if we call the bartender's glasses cognitive or merely causal?

If the behavior is explained either way, who cares about the philosophical label?

Well, we care because the definition of the mark of the cognitive affects far more than just academic philosophy.

If Otto's notebook system is genuinely a cognitive process, then its damage isn't just the loss of a tool.

It's cognitive damage, maybe even a form of amnesia.

The boundary of the mind is the boundary of the self.

If the external is constitutive, we may have moral and even legal responsibilities toward our extended cognitive apparatus.

The whole debate shifts from just observing inputs and outputs to defining that boundary property.

If we can't define the mark of the cognitive clearly, we can't decide if the notebook is mind or just paper.

And since the traditional Cartesian definition is insufficient, we now have to examine the major philosophical defense of EXT, which tries to establish this mark using functionalist arguments.

So the first, and certainly the most famous, attempt to define this mark of the cognitive for the extended mind.

It comes from Extended Functionalism, FEX.

And this is championed by philosophers Andy Clark and David Chalmers.

Okay, so FEX, it maintains the orthodoxy of classical cognitive science, right?

Cognitive processes are essentially computational and representational.

The mind works by processing information using internal structures.

But FEX makes this radical move of arguing that some of those relevant computations can actually take place in the world via our bodily actions on external information -bearing structures.

And their foundational argument is the parity principle, which they introduced in 1998.

It's elegant precisely because it seems to appeal to a kind of philosophical fairness.

Let's look at the wording.

If as we confront some task, a part of the world functions as a process, which were to go on in the head, we would have no hesitation in accepting as part of the cognitive process, then that part of the world is, for that time, part of the cognitive process.

The central idea is just that cognition should be counted as cognitive regardless of its physical location.

If a process does the job of cognition, it is cognition.

It doesn't matter if it's implemented in neurons or silicon or line of glasses on a counter.

So applying this to our bartender.

If the bartender were forced to mentally visualize that line of glasses,

an internal process everyone would agree is cognitive.

The parity principle demands that we count the physical line of glasses in the world as equally cognitive because it's performing the same function.

The argument is powerful because it challenges what Clark called the skull -bound prejudice.

Why should the arbitrary boundary of our skin and skull determine what counts as mental if the functional role is identical?

But this is where the argument immediately runs into that definitional problem we talked about.

The source material notes that the parity principle, on its own, doesn't really settle anything.

No, it doesn't.

The parity principle only works if we already have a preexisting standard, the mark of the cognitive, for judging what counts as cognitive in the first place.

We're using the principle to try and settle the status of an external process, but its application relies on us having already settled that question for the internal case, it kind of risks begging the question.

So to establish this mark, Clark and Chalmers initially appealed to a blend of what, common sense functionalism, our folk psychological understanding of belief and memory and modern cognitive science.

And this is where we encounter the famous Otto and Inge thought experiment.

So Inge has a standard biological memory system.

She uses it to recall that Mama is on 53rd Street.

Otto, on the other hand, is suffering from advanced Alzheimer's.

He relies heavily on a notebook to store crucial information, including Mama's location.

And this notebook is seamlessly integrated into his life.

He consults it automatically without conscious deliberation, just as Inge automatically accesses her memory.

And the functional equivalence claim is the linchpin of the whole thing.

FEX argues that Otto's notebook system and Inge's declarative memory system are functionally equivalent dispositional belief states.

They play the same coarse -grained action guiding role in their lives.

I mean, they both know through their respective mechanisms that Mama is on 53rd Street.

Therefore, since Inge's memory is cognitive, Otto's sister, the notebook plus his brain, must also be cognitive.

But this is where the EMT proponents, especially Adams and Aizawa, launch their most effective counterattack.

They argue that FEX is defining the cognitive kind way too broadly by relying on these coarse -grained functional roles.

They say we need to look at the fine -grained functional differences.

Okay, what kind of fine -grained differences are we talking about here?

Well, they point out that Otto's external memory system would likely not exhibit the myriad typical memory effects that we see in Inge's biological memory.

For instance, in psychological tests, Inge might show the primacy and recency effects.

Right, remembering items better at the beginning and end of a list than the stuff in the middle.

That's a classic feature of human memory.

Exactly.

It's because of how our working memory and long -term storage interact.

But Otto, using a notebook, would show no such decay or enhancement effect.

If he wrote down a list, every single item would be equally accessible, assuming his notebook hadn't run out of ink or suffered some physical damage.

So these differences for EMT are critically important.

They're everything.

They suggest that the systems are not functionally equivalent enough to be considered the same cognitive kind.

If you define a memory kind by its specific causal roles, including these unique internal effects, how it degrades, how it retrieves information under pressure, its precise speed, then the external system fails the test.

It might accomplish the same general goal of storing data, but it does it with a fundamentally different functional architecture.

This disagreement forces the debate away from fuzzy folk psychology.

And towards a much more rigorous standard,

empirical functionalism.

So both sides start appealing to the findings of actual cognitive science to identify the causally explanatory properties that define a cognitive kind.

The question becomes, what verdict does our best science deliver by the essential nature of memory?

Mark Wheeler, who defends FEX, tries to accommodate these differences by arguing for the concept of generic kinds.

He concedes that there are functional differences between Otto and Inga, but he asks a really interesting question.

He asks, if psychologists found a person whose internal memory processes exhibited those very same differences, maybe an internal memory system that somehow bypassed normal biological decay effects, would they strip that person of the label declarative memory?

And Wheeler thinks not.

As long as the system exhibits the core generic function of context sensitive storage and retrieval of information, it still counts as declarative memory.

The implementation details, the physical substrate or the specific memory quirks don't matter as much.

For Wheeler, the differences are just implementation details.

The generic kind is still memory, regardless of whether it's realized wholly inside or partially outside.

And this led to a really crucial division within EXT itself, which the philosopher Stephen Sutton labeled first wave EXT versus second wave EXT.

Okay, so how do those two waves differ philosophically?

Well, first wave EXT, which aligns with Wheeler, emphasizes that extended processes must share common underlying computational mechanisms, or at least a strong family resemblance.

They're looking for sameness.

But the second wave EXT view is, I think, maybe more interesting.

It stresses complementarity and functional differences.

They argue that external structures, like Otto's notebook or the bartender's glasses, they don't necessarily mimic the brain.

They deliver new hybrid cognitive systems that are part biological, part cultural, and part technological.

Yes, the point of second wave EXT isn't to prove sameness.

The point is that different components play different complementary roles to create a new, powerful kind of cognitive system.

They embrace the fact that the underlying mechanistic principles may not be the same.

But the overall system still counts as cognitive because of its functional integration into the agent's life.

This complexity forced Wheeler to propose a more formal, scientific mark of the cognitive to trick and solidify FEX.

And that was Newell and Simon's physical symbol systems hypothesis, or PSSH.

Right, this is a classic hypothesis.

It suggests that cognition relies on the presence of sufficiently complex physical symbol systems.

These are systems that are capable of holding,

manipulating, and using symbols that refer to things in the world.

And Wheeler's argument is that Otto, plus his notebook, clearly forms such a system, and it shares a family resemblance with Inga's purely biological symbol -manipulating brain.

But Rupert counters this focus on general function by drilling down on mechanisms again.

This is sort of his final move in the functionalist debate.

It is.

Rupert argues that science only recognizes a generic, kind -like declarative memory if it's brought about by the same cluster of integrated and persisting mechanisms.

Mechanisms are the persistent, underlying structures that cause the behavior.

And since the mechanisms realizing memory in Otto, a notebook, plus a brain, are vastly different from Inga's, a brain alone, they're not the same kind.

So Rupert is essentially saying,

if the components that cause the function are structurally and operationally different, a biological neural network versus ink on paper,

then the resulting processes are fundamentally different cognitive kinds, even if their outputs sometimes look similar.

And this leads us directly back to the deadlock.

Even if EMT concedes that external storage is possible, acknowledging these hybrid processes, they can always insist that the environmental contribution is only causal, not cognitive.

So an EMT advocate can look at the bartender system, or the Otto system, and agree that it's an extended cognitive system, a system designed to accomplish a cognitive task.

But they'll insist that the external components are non -cognitive elements working on the cognitive element inside the skull.

They're essential tools, but they're not constitutive parts of the mind.

And the problem is that science on its own cannot resolve this distinction.

We need philosophy to interpret which causal properties are of cognition, but the philosophy we've used so far, functionalism, is relying on scientific evidence that remains ambiguous.

We're stuck in a loop.

We need a new philosophical starting point that doesn't rely solely on representation or functional resemblance.

We've seen that functionalism just traps us.

So to truly escape this deadlock and to find the mark of the cognitive, we have to question the one assumption that both sides currently share.

And that's the idea that cognition must be defined by internal representations.

It is time to take the radical option.

So to break this philosophical deadlock, we have to move to a new theoretical framework.

Radical Extended Cognition, or EX, sometimes called Radical Embodied Cognitive Science, or RECS.

And this position seeks to achieve what that first wave of EXT couldn't.

It rejects the traditional representational mark of the cognitive altogether.

And this approach is rooted deeply in the American pragmatist tradition.

You think of figures like William James and John Dewey.

And the core philosophical thesis here is the mutuality of animal and environment.

Dewey described this relationship not like marbles sitting in a box, separate distinct entities, but like events in history.

Events in history.

I mean, they're defined by their interdependencies, right?

And you can't understand the rise of one civilization without understanding the context of another.

Exactly.

The organism and the environment are interdependent.

They form what Dewey called a moving, growing, never finished process.

RECS translates this historical pragmatism into a modern scientific framework by aligning with dynamical systems theory, DST, and ecological psychology.

The goal is to drop the assumption that cognition must be defined by internal representations and computation, which was the Achilles heel of functionalism.

Okay.

Dynamical systems theory.

That sounds like pure math.

How does it help us define the mind?

Well, DST provides the mathematical language for modeling change over time in complex systems.

Any system that changes in a nonlinear integrated way.

Instead of focusing on discrete inputs and outputs, like in computation,

DST focuses on the continuous mutual influence between components.

And the key concept here is coupling.

How does DST define that?

So two systems, let's call them S1 and S2, are coupled when the equations describing the change in S1 contain variables whose value is a function of the variables in S2 and vice versa.

If you have a predator prey system, like the famous Lotka -Volterra equations, the change in the predator population depends on the prey population, and the change in the prey population depends on predation.

They are continuously locked in this dance of mutual influence.

This mutual influence is critical because if I only track the environment's effect on the agent, that's just mere causality.

EMT is fine with that.

But REX is claiming a deep, mutual, two -way link.

Yes.

In extended cognition, the agent and the environment are dynamically coupled.

Variables describing changes in the agent neural activity, bodily movement, are influenced by variables describing the external information -bearing structures.

And the agent's actions are simultaneously influencing and changing those structures.

And crucially, two systems that are coupled in this way resist decomposition into separately functioning systems.

This is the REX counterpunch to the EMT hybrid system objection.

The overall behavior isn't just the additive sum of the brain component plus the environmental component.

It arises from continuous, integrated, and coordinated mutual influence, what DST calls an interaction -dominant dynamics.

Could you give us a more human -scale analogy for interaction -dominant dynamics?

Sure.

Think about walking.

Or even better, pushing a heavy, squeaky shopping cart.

The way your muscles, S1, fire depends entirely on the feedback you get from the environment S2, the cart in the floor.

If the cartwheel gets stuck, your muscles immediately tense up and change the force you apply.

But your force also changes the cart's movement.

It's an inseparable, non -linear loop.

Trying to decompose that into first the brain computes, then the arm acts, that ignores the continuous, necessary regulatory feedback.

That makes a lot of sense.

The system's performance relies on the constant, real -time mutual influence, not some sequential processing.

And even Andy Clark, despite retaining his commitment to representation, he recognized this complexity.

He noted that understanding these extended ensembles is insufficient if you try to take them apart piece by piece and then add them back together.

Now, REX needs to address Rupert's major objection to this whole coupling argument.

Rupert claimed that in many cognitive models we don't find genuine coupling, but rather one -way causality.

He said the environment merely causes variation among a small number of internal organismic systems.

And he used the famous Tetris thought experiment to illustrate his point.

Right.

When a player rotates the falling geometric shape, the zoid in Tetris to determine if it fits.

Rupert claimed this is not true coupling.

Why?

Because the player's rotation doesn't change the object's fundamental dynamics.

Its intrinsic predetermined evolution in state space remains the same as it was before the rotation.

The player is simply intervening on an external process, not becoming coupled with it.

But REX provides a highly effective rebuttal to this.

It does.

Rotation is clearly a case of coupling because it fits mathematical definition.

The equations describing the game must include variables for the player's action, the rotation, and the equations describing the player's action must include variables for the zoid's orientation.

The player's ability to recognize the fit is fundamentally facilitated by that external act.

And Rupert was focused too narrowly on the single piece.

Way too narrowly.

If we look at the system over time, his argument fails entirely.

Coupling is often defined using what are called order parameters, the macro level variables that define the state of the entire system.

So if the order parameter is defined as the overall configuration of pieces on the board over the course of the game,

then rotation dramatically affects the outcome.

Absolutely.

Kirsch's research showed that rotation, when it's done early enough, has a sizable influence on the player's success.

Without rotating the pieces, the board layers fill up much, much sooner and the player loses quickly.

The speed at which the Tetris board fills up depends entirely on the player's action of rotation and the player's rotation decision depends entirely on the current state of the board.

The two systems, player and game state, are in a relation of continuous mutual causal influence.

It is a dynamically coupled system.

This dynamic approach allows REX to bypass the representation debate entirely.

Especially when REX integrates ideas from ecological psychology, which was developed by James Gibson.

Ecological psychology is the ultimate non -representational mark of the cognitive.

It fundamentally claims that the agent doesn't interact with the environment through the intermediary of internal representations or models.

So if the mind doesn't build models of the world, how does it know what to do?

Well, ecological psychology posits that the environment, or the ecological niche, is rich with structural invariance patterns and structures that specify affordances.

And affordances are the possibilities for action provided by the environment to a particular animal.

The central idea is direct detection.

That's it.

Perceptual systems guide action by immediately and directly detecting these higher order invariants.

Take the famous example of approaching a precipice, like a cliff edge.

The agent immediately detects the shearing off of the ground texture in their visual flow.

That's a higher order invariant that specifies the affordance, the potential to fall.

This information is carried directly in the light reflecting off the surface.

There's no need for the brain to construct an internal 3D model of the cliff's height and then compute the danger level.

The information that the surface is no longer walkable is simply there, in the energetic stimulation hitting the eye.

This dramatically changes the perspective on the mind's function.

It shifts from the mind as a processor of symbols to the mind as a mechanism for attuned participation with the world.

And REX uses this ecological framework to reinforce its core conclusion.

Interaction with the world isn't just about delivering a French stimuli for the brain to construct representations.

Instead, the agent dynamically couples with external information -bearing structures exerting continuous and mutual causal influence.

This requires the whole thing agent plus environment to be modeled as a single extended cognitive system.

And this dynamic coupling is the key to dismissing the most effective counterargument used by EMT, the hybrid system objection.

Exactly.

The claim that an extended system is merely a hybrid cognitive inside, non -cognitive outside.

That whole idea assumes the possibility of a clean decomposition.

It assumes we can separate the internal cognitive elements from the external non -cognitive elements.

Dynamic coupling resists this.

If the system's coherent behavior is generated by the continuous real -time mutual regulation of all the components, then decomposing it into separate cognitive and non -cognitive parts is arbitrary.

It misrepresents how the system actually works.

So the debate about hybrid systems, therefore, ultimately rests on whether representation is taken as the mark of the cognitive.

By rejecting the computer theory of mind,

REX breaks that essential link between empirical functionalism and representationalism.

And in doing so, REX offers its own distinct mark of the cognitive.

It's one drawn from the American naturalist tradition.

REX defines extended cognitive systems as perception -action systems based on the fundamental mutuality of animal and environment.

So the mark is not some computational property realized in symbols.

It is a dynamic relational property.

Cognition is defined pragmatically by the way these dynamics relate to the agent's adaptation to its environment, its ability to deal adequately with its affordances.

The mark of the cognitive is the continuous, inseparable, mutual causal influence between the agent and the world that generates adaptive behavior.

So what we've seen, really, is that this question of whether the mind extends into the world is fundamentally a debate about definition.

It forces us to ask what properties are constitutive of a cognitive process versus those that are merely causally relevant.

And the first major philosophical defense of extended cognition, extended functionalism, or FEX,

which is based on the parity principle and the requirement of representation.

It just hit a philosophical wall.

It really struggled to definitively distinguish between a hybrid causal system and a truly extended constitutive one without begging the question of what the mark of the cognitive actually is.

And here's where it gets really interesting.

The Radical Extended Approach, REX, offers a potential way out by discarding the requirement of representation altogether.

It suggests that the true mark of the cognitive resides in the dynamics of coupling and the mutuality of the agent environment system.

If the component cannot be removed without fundamentally changing the dynamics that generate the adaptive behavior, then it is constitutive.

It's part of the system.

So what does this all mean for you, the learner?

It means that when you're relying on your phone for memory, or when a surgeon relies on external monitors and robotic arms to perform a task, you are participating in a conversation that challenges the very boundaries of what it means to be a thinking -knowing agent.

The explanatory strategy is clear.

We have to look outside the head to explain complex behavior.

But the metaphysical answer is still evolving.

It depends on which philosophical foundation representational functionalism or non -representational dynamics you choose to stand on.

The choice of the mark of the cognitive changes the nature of the mind itself.

So a final provocative thought.

If cognition is defined by the inseparable dynamic coupling between you and your environment, as Ari X suggests,

then the state of your desk, your digital tools, the physical spaces you occupy, they are not just tools for your mind.

They are, for that time, integral to who you are thinking as.

Where do you see the most compelling, inseparable evidence of interaction -dominant dynamics in your own daily life?

Where does coupling become so seamless that the distinction between you and the tool just disappears?