Chapter 19: Motor Intentionality

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 back to The Deep Dive.

Today we're taking a really deep cut from the world of 4E cognition, tracing a powerful idea that tries to bridge the gap between, well, how we think and how we actually move.

That's right.

We're wading into that overlap between philosophy, neuroscience, and action theory.

And we're going to be focusing on this almost mysterious connection between a high -level plan, you know, I'm going to pick up that cup, and the effortless unthinking motion our body uses to actually do it.

Precisely.

Our mission today is focused entirely on a concept called motor intentionality.

It was first coined by the French philosopher Maurice Merleau -Ponty way back in 1945.

Okay, so that's the term for the day.

It is.

We're going to explore how our body moves so skillfully, so unreflectively, and how that kind of embodied knowing integrates with our conscious, deliberate planning.

The central question really is about integrated human agency.

How do the reasons we have for acting actually translate into the specific mechanics of motion?

Okay, let's unpack that central concept right away then.

Motor intentionality.

People might see it abbreviated as MI.

Yeah, MI.

What exactly did Merleau -Ponty mean by that?

And I guess why did he even need a new term for it?

Right, that's the core of it.

So motor intentionality is the form of intentionality.

And by intentionality, we just mean the property of a mental state being about something, directed toward a goal.

Directed at something.

That is specifically seen in

purposive, skillful, and totally unreflective bodily activities.

It's the skilled way our body just habitually relates to the world.

So we're talking about the action system that operates way below the level of conscious analysis.

It's the difference between thinking, okay, I need to coordinate my triceps, my deltoids.

Right, all the specific muscles.

And my brachioradialis to stabilize my arm and just reaching out and grasping a doorknob.

It's riding a bike, it's dodging a falling branch, or catching a ball.

It's where the skill just flows without you thinking about it.

Okay, and Merleau -Ponty needed to distinguish this from what?

He felt he had to coin the term because he wanted to explicitly contrast it with the traditional, more cognitive, conceptual,

and representational forms of intentionality.

The thinking part.

The thinking part.

Cognitive intentionality is when you form a conscious plan, a proposition.

I intend to go to the store or I believe the store is open.

Motor intentionality is what actually executes the skilled movements.

Opening the door, maintaining your balance, the quick calculation of grip force that makes that conceptual plan actually succeed.

And the birth of this concept, the whole origin story that anchors this philosophical distinction, is rooted in this tragic but absolutely compelling case of a soldier named Schneider.

Absolutely.

Merleau -Ponty introduced MI in his huge 1945 work, Phenomenology of Perception, and he did it through a long discussion of this patient, Schneider.

Schneider was a soldier in World War I who had a very serious brain injury.

And it resulted in a whole host of problems.

A huge array of severe neuropsychological impairments.

He had alexia so he couldn't read, agnosia so he couldn't recognize objects, he had a loss of his body schema, severe loss of abstract reasoning.

It sounds like a complete failure of the conceptual, objective mind.

It was.

But what made Schneider so critical for Merleau -Ponty's argument was that his motor skills showed this.

This really bizarre dissociation.

A split.

A total split.

While he was profoundly impaired cognitively, his actual habitual movement still had a competence that, well, it shouldn't have been possible given his injuries.

In this contrast, this gave the philosopher the key evidence he needed to argue that motor intentionality is a basic, independent, and fundamental form of knowing.

It's entirely distinct from conceptual thought.

And that dissociation, that tension between what he could think and what he could still do, that sets up the entire problem we're tackling in this deep dive.

Our sources lay out three core questions we really need to answer to fully get our heads around MI.

The first one is foundational.

How should we rigorously characterize motor intentionality today, given all the advances in neurosciences in 1945?

Okay.

And second, and this is the major philosophical pivot point we're going to have is the question of representation.

Is MI truly non -representational, like Merleau -Ponty originally claimed, because it doesn't have conceptual content?

Or does cognitive science force us to say no, there are distinct motor representations in the brain, even if they're non -conceptual.

And then the third question, and this sounds like the toughest challenge for any theory of human action, is the interface problem.

The interface problem, yes.

How can motor intentionality and the more cognitive conceptual forms of intentionality, those conscious plans and beliefs, how can they actually be seamlessly integrated?

We're looking for the mechanism.

The thing that allows our high -level reasons to translate into the precise mechanics of the body without any magic being required.

Okay.

So let's delve deeper into Schneider because it really sounds like the devil is in the details of this specific case.

Gelb and Goldstein's findings on his motor performance.

They detail this critical dissociation between two types of movement.

And this distinction is what Merleau -Ponty built his entire philosophy on.

The researchers clearly distinguish between what they called concrete movements and abstract movements.

Okay.

So concrete is the real world stuff.

Exactly.

Concrete movements were those habitual actions, the ones performed in the context of everyday life.

Schneider could still execute these with speed and precision.

For instance, he could reach into a box, take out a match,

strike it to light a lamp.

He could do routine tasks if he mentally put himself in the right environment.

And the classic example you always hear is that he could accurately grasp his own nose.

This is the skill that was preserved.

Exactly.

But then you have the abstract movements.

And these were isolated arbitrary movements performed out of context, movements that weren't relevant to an actual immediate situation.

And this is where he fell apart.

Massively impaired.

He couldn't perform isolated movements on command, like bending one specific finger.

He couldn't move a limb just to order.

He couldn't point to his nose or draw a simple circle in the air.

The dissociation is just so stark.

Grasping his nose was fine, but pointing to it was impossible.

And this contrast allowed Merleau -Ponty to make two seemingly contradictory arguments, which we have to reconcile.

Let's look at the first argument, which focuses on the preservation of his skill.

He called this the centripetal dimension of MI.

Centripetal, like moving toward the center.

So it refers to actions that are successful because they're grounded in the body's practical understanding of its own space.

Precisely.

Merleau -Ponty argued that grasping his nose involves a practical understanding of his bodily space.

His body functions as the matrix of his habitual action.

It's a phenomenal non -objective relationship.

He grasps the nose not by plotting its coordinates in objective space.

He's not doing trigonometry in his head.

Not at all.

His hand just knows the path to the nose.

It's a direct, lived relationship within the body's own natural system.

I mean, think of reaching the exact painful spot where a mosquito bit you, even in the dark.

You don't need to calculate any coordinates.

Your body just knows.

It just knows.

And that grasping action, or the scratch, that's evidence that MI can function independently of high -level conceptual representations.

This preserved habitual ability is the centripetal dimension, where action is backgrounded by the world as it's just given to the skilled body.

Okay, so that's the preserved part.

Now for the tricky part.

Argument line two, which describes the impairment, the centrifugal dimension.

Right, centrifugal moving away from the center.

Merleau -Ponty also noted Schneider's failure to perform those abstract movements, like tracing a circle.

Okay, so if Schneider could recognize a circle if he saw one, he could definitely move his arm.

What was the actual deficit?

What was missing when he couldn't draw a circle on command?

Merleau -Ponty concluded that Schneider lacked the capacity for projection.

Projection.

He lacked something which is an anticipation of or arrival at the objective.

This something is what he called the motor project,

an anticipatory forward -looking motor intentionality.

Without this centrifugal function of projection, a simple command to move just remains a dead letter.

So the centrifugal dimension is where the body doesn't just react to the world as it's given, which is the centripetal part, but it actively constructs its own background and sort of throws itself out onto the world, making new actions possible.

That's a great way to put it.

Wait, I need to pause here, because this is exactly the kind of nuance that can trip you up.

Merleau -Ponty is arguing that MI is both fully preserved in concrete actions, but fundamentally impaired in abstract ones.

So how do we reconcile MI being both preserved and impaired in the same person?

The sources suggest the resolution lies in understanding MI as having these dual dimensions you just mentioned.

His claims can be reconciled if we just separate the functions.

So the centripetal dimension is the basic, almost instinctual motor intelligence, the capacity for an independent, skillful reactive function.

That was preserved in Schneider.

He can scratch the itch.

He can habitually grasp the doorknob.

Exactly.

But the centrifugal dimension, that function of projection, is what was lost.

Projection is the capacity to convert an abstract thought, or a conceptual plan like draw a circle or move my arm to order,

into an actual, specific, executable moto plan.

It ensures that transition from abstract thought to concrete movement.

So if we use a modern analogy, the centripetal dimension allows a skilled pianist to repeat a familiar song they've played a thousand times.

It's muscle memory.

It's integrated.

Perfect.

But the centrifugal dimension is what allows that same pianist to sight -rate a complex musical score they've never seen before, converting the conceptual info on the page into novel finger movements.

Schneider was stuck in the habitual.

He couldn't form that novel motor project.

That is an excellent way to frame it.

The ultimate takeaway here, which really sets up the rest of our deep dive, is that a full account of motor intentionality can't just stop at showing it's distinct from cognitive thought.

We have to find the mechanism that explains this mysterious function of projection, that critical bridge between a high -level plan and low -level mechanical reality.

Fortunately, science has since weighed in on this philosophical split.

Modern neuropsychology seems to strongly support Merleau -Ponty's initial insight about this fundamental distinction between the practical motor system and the cognitive perceptual system.

It does.

And this evidence is powerfully articulated through the dual system model of visual processing, which is primarily associated with researchers like Milner and Goodale.

The two streams hypothesis.

Exactly.

It posits that the brain processes visual information through two separate cortical streams, or pathways, that are anatomically and functionally distinct.

Let's start with the perception stream.

Okay.

So we have the ventral pathway.

It runs along the bottom of the brain and is often labeled vision for perception.

This pathway is responsible for our conscious object perception, recognition,

categorization.

It builds our stable, conscious visual world.

It answers the question, what am I seeing?

And the other one, the motor stream.

That's the dorsal pathway.

It runs along the top of the brain and it's the vision for action system.

This pathway deals with the visual guidance of actions directed at objects.

It deals with real -time, egocentric spatial information.

It answers the question, how do I interact with what I'm seeing?

And crucially, the source material shows that these systems can be doubly dissociated.

And that's the key.

The double dissociations are the real aha moments here.

They confirm that the two systems can operate entirely independently of each other.

So let's start with the patient who had a failure of perception, but an intact action system.

Patient DF?

Patient DF sustained lesions to her ventral pathway, which led to visual form agnosia.

She was completely unable to consciously recognize objects, identify simple shapes, or even report their orientation.

You could show her a square and a triangle, and she couldn't tell you which was which just by looking.

So her vision for perception, the what system, was basically non -functional.

But what about her vision for action?

Strikingly preserved.

I mean, it's incredible.

When you ask her to perform a grasping movement, DF could reach out and pick up objects with remarkable accuracy and dexterity.

She would optimally shape her hand for the specific object, timing the opening and closing of her grip flawlessly.

And what about that classic experiment with the mail slot?

Right.

She was asked to post a card through a slit.

She couldn't consciously report the slit's orientation at all.

She'd just guess.

But when you asked her to actually insert the card, she correctly and rapidly oriented the card to match the slit's angle every single time.

She demonstrated sophisticated motor -guided action without any conscious visual awareness of the object's properties.

That perfectly mirrors Merle Ponte's idea of preserved M .I.

She didn't know what she was doing conceptually, but her body knew how to do it practically.

And the functional reverse of this is patient A .T.

who had a dorsal lesion which results in optic ataxia.

So patient A .T.

had the inverse problem.

She had normal conscious perception.

She could report the shape, size, and orientation of objects perfectly.

She could tell you exactly what she was seeing.

Yep.

But when she tried to act toward those objects, her movements were just systematically incorrect.

Her attempts at grasping or pointing were disorganized, clumsy, and inaccurate.

She had the conscious perception, the conceptual knowledge, but she lacked the competent moment -to -moment action guidance.

The double dissociation is complete.

It is.

And it strongly suggests that these visual motor representations can be built and can function entirely independently of our conscious visual perceptions.

And the beauty of this distinction is that we don't only see it in brain -damaged patients.

We see this disconnect in healthy, normally functioning people all the time, especially when they're subjected to visual illusions.

That's right.

The fact that your conscious awareness can be tricked while your motor control remains accurate is maybe the most robust evidence we have for the functional independence of these two systems.

So take saccadic suppression, for instance.

Right.

Cicades are these really rapid ballistic movements of the eyes.

And during these movements, the brain basically suppresses vision, so you don't just see a blur.

Now, if a target is displaced during that saccade,

the participant doesn't consciously perceive the shift.

The world seems stable.

But if you instruct them to point to the now -displaced target,

researchers found that their pointing accuracy remains entirely unaffected.

They point right to the new spot.

So the motor system got the memo about the change in coordinates.

They got the real -time coordinates, even though the perceptual system completely failed to update conscious awareness.

And we see the same principle at play in the dot and frame illusion.

In this illusion,

you have a stationary dot in the middle of a big moving frame.

To the observer, the dot appears to be moving in the opposite direction of the frame.

Your conscious perceptual judgment is systematically deceived.

But I'm guessing if you ask someone to point to the dot.

Their pointing accuracy is not affected at all.

Again, the visual motor guidance system is using the non -illusory input.

It's operating outside the realm of our conscious visual awareness.

But the example that has become a real staple in this field, the one that truly solidified this ventral -dorsal split for action theory,

is the Titchener or Ebbinghaus illusion and its effect on grasping.

The Titchener illusion is so compelling because it's so easy to see.

You have a central circle surrounded by small circles, and it appears larger than an identical circle surrounded by big circles.

Right.

I know the one.

When researchers asked participants for a perceptual judgment, which circle looks bigger,

they were reliably deceived.

They always pick the one that looked bigger.

But the measurement of the physical act of grasping, that told a completely different story about the information the motor system was using.

It did.

Researchers measured the maximum grip aperture, or MGA, during the grasping movement.

Okay.

What is that exactly?

MGA is the largest distance between your thumb and forefinger during the transport phase of the reach.

It happens reliably at about 60 to 70 percent of the movement's duration.

And why is that specific moment, the MGA, so important?

Because it's the moment the motor system commits to the final configuration required for the actual physical object.

It shows the motor system planning for the object size before the grasp is even complete.

So it's a window into the motor plan.

It's a perfect window.

And what they found was that while perceptual judgments were tricked by the illusion,

the MGA remained correlated with the object's actual physical size, not its illusory size.

The action system saw reality.

The conscious system saw the trick.

This is a really robust demonstration of that functional distinction.

So the empirical evidence decisively validates Merleau -Ponty's idea that motor intentionality is a basic, independent form of intentionality.

But as we hinted at earlier, the dual system model isn't a perfect separation.

We have to confront the nuance now, especially the anatomical crosstalk, because that leads right back to Merleau -Ponty's problem of integration.

That's right.

Our current understanding, which has been refined by neuroanatomical studies, confirms the separation isn't total.

There's significant and necessary crosstalk between the streams.

For instance, researchers like Rizzolatti and Mattelli have suggested that even the dorsal stream itself is divided into two anatomically segregated subcircuits.

Subcircuits within the action system.

That gets technical pretty quickly.

Can we ground this for a minute?

Sure.

Think of it in terms of motor memory and real -time correction.

There's the dorsal -dorsal pathway, which is concerned with immediate moment -to -moment visual motor control, constantly adjusting your grip pressure, because the surface of the object you're lifting is slicker than you thought.

It's the constant, immediate feedback loop.

Okay, that makes sense.

And then the other one?

The ventrodorsal pathway.

This is thought to be involved in the long -term storage of skilled actions associated with familiar objects.

It stores the generalized knowledge of how to use a familiar tool, or the typical action patterns for a specific type of object.

Okay, let me try an analogy to check if I'm following.

Is the dorsodorsal pathway, like my muscle memory, when I'm desperately trying to keep a coffee from spilling right now, using immediate visual input?

Yes, that's it.

And the ventrodorsal is the memory for how to hold a hammer properly, which I haven't used in a year, but the skill just comes back immediately.

That's a highly effective way to visualize the functional difference between that immediate reactive control and the stored generalized skill knowledge.

Now these substreams connect back to the ventral stream, the perception stream.

The crosstalk.

The crosstalk.

The ventrodorsal pathway, which stores those generalized skills, seems to act as a crucial interface between the dorsal and ventral streams.

And why would we need an interface between action and perception?

Why can't they just stay separate?

Because our actions often require conceptual knowledge.

Think about two crucial situations, delayed or pantomime grasping, where you're grasping an imaginary object or relying purely on visual memory.

You need to access conceptually stored information about the object's size and location.

Right, you need your what system to tell your how system what to do.

Exactly.

And the other big one is tool use.

When you use a hammer, your grasp is dictated not just by its shape, but by your conceptual knowledge of the hammer's function.

That's ventral pathway information influencing the dorsal stream.

So this crosstalk actually reinforces Merleau -Ponty's original complexity.

The double dissociations support his idea of MI as a basic independent form, but the anatomical communication pathways support his later critical idea that MI must also ensure the transition and integration with abstract cognitive forms.

It moves the philosophical debate from whether MI is distinct to how it talks to our thoughts.

Which leads us directly into the philosophical thicket of representation.

Here we go.

This section addresses the central philosophical pivot.

Merleau -Ponty famously insisted that motor intentionality was non -representational, yet modern cognitive scientists, especially those talking about the computational demands of the dorsal pathway,

routinely talk about sensorimotor representations.

So is this a genuine contradiction or are they just using the word differently?

The sources argue it's primarily a definitional clash.

It's rooted in how Merleau -Ponty used the word representation.

His definition was very specific.

Extremely restrictive.

For him, a representation had to have propositional conceptual content information that could be expressed as a factual statement in language, and it required the agent to adopt an objective detached stance.

Since MI, in its raw form, is practical, non -conceptual and embodied, it failed his test.

But the modern view of representation, the one you need for computational and cognitive theories, is much broader.

We have to be able to attribute content to the motor system to explain what it does.

To clarify this, we can use a framework from the philosopher José Luis Bremudes.

He offers four functional criteria for a state to qualify as representational in a less demanding sense, regardless of whether his content is conceptual or not.

Okay, let's scrutinize these functional requirements because they form the basis for why we now have to accept that MI is, in fact, representational.

So criterion one, correctness conditions.

For a state to be a representation, it must have conditions under which it can be judged, correct, or incorrect.

It has to be possible for the representation to misrepresent the world.

It has to be able to be wrong.

It has to be able to be wrong.

For instance, if my motor plan specifies a grip size of 10 centimeters for an object that's only 5 centimeters, the representation is incorrect, even if I still manage to kind of fumble the object successfully.

This allows the system to engage in error correction and learning.

Okay, makes sense.

Criterion two, compositional structure.

This means the state isn't just a holistic blob of information.

It must be made up of constituent units, a vocabulary, or a lexicon that can be systematically combined according to a grammar.

This is necessary to explain how we can generate an almost infinite variety of complex actions from a finite set of motor primitives.

Okay, criterion three, cognitive integration.

This requires that the representational state can interact with, influence, or be influenced by other cognitive states like our beliefs, desires, and high -level intentions.

If a motor action is influenced by my belief that the cup is full and heavy, then it has to be integrated with my conceptual belief system.

Got it.

And finally, criterion four, explanatory role.

A representation must play a necessary explanatory role that goes beyond simple, lawful stimulus response, or SR, relations.

If the behavior is too complex, too context -dependent, or too goal -oriented to be explained by mechanism alone, then appealing to contentful states to representations is indispensable for a full explanation.

And a crucial point here, one that's often missed, is that compositionality and cognitive integration aren't just absolute binary properties.

They're not on or off switches.

Precisely.

They are graded notions.

This means the philosophical difference between conceptual and non -conceptual content might just be a matter of the degree of structure and integration rather than some sharp absolute difference between representational and non -representational forms.

Motor intentionality is simply at the non -conceptual end of the representational spectrum.

So now that we have that broader framework, how do we characterize the specific format of these motor representations?

They're characterized as being non -conceptual, pragmatic, and relational.

Non -conceptual, meaning they don't store linguistic or abstract categories.

And pragmatic, meaning they're purpose -driven.

Yes.

They represent the goal of an action in a very specific pragmatic mode.

They only encode the visual attributes of the object that are relevant for selecting the right motor pattern.

For example, the motor representation for reaching for a coffee mug doesn't need to encode the brand name or the color of the paint.

Because it doesn't matter for the grip.

It's pragmatically irrelevant to generating the right grip force.

What is relevant is the size, the texture, the location, and a representation of the final state of the acting body, where the fingers will end up.

And they are relational, focusing on that dynamic connection between me and the object.

As suggested by the neuroscientist Mark Janerod, they are fundamentally relational.

They represent the relations between the body and the goal object.

They define the motor patterns that the object affords to the agent.

They encode the object in terms of the actions it makes available.

And furthermore, these representations aren't just static snapshots, are they?

They are inherently dynamic.

Action isn't instantaneous.

It unfolds over time.

And it requires continuous guidance and control.

Therefore, motor representations have to anticipate future states of both the environment and the body.

And this is where the modern science of motor control theory comes in, with the idea of internal forward or predictive models.

This is the system that lets us act fast, bypassing that slow sensory feedback loop.

Exactly.

Researchers like Wolpert and Coato established that these forward models capture the causal relationships between a motor command and its likely sensory consequences.

The system estimates the effects of a command before the movement is complete, compares that prediction to the incoming sensory feedback, and makes continuous adjustments.

The content of the motor representation is dynamic because it's constantly being elaborated over time, allowing for swift,

skillful, and robust control, even if external sensory feedback is delayed or incomplete.

So let's explicitly put these characteristics, the dynamic relational pragmatic content, to the test using Bermuda's criteria to really confirm the representational nature of MI, starting with criterion one, correctness conditions.

Since motor representations blend a sensory function, extracting relevant facts and the motor function directing action, they're a hybrid.

They're not purely descriptive like a belief or purely directive like a desire.

Computationally, they're more efficient.

You refer to these as push -me -pull -you representations or PPRs, which is a great philosophical term from Ruth Milliken.

Can you make that concept even clearer for us?

Think of it this way.

A purely cognitive belief says,

the cup is full.

That's a world -to -mind fit.

A desire says, I want a drink.

That's a mind -to -world fit.

The motor representation combines both into one state.

Here is how to grip a full cup so that I can successfully drink it without spilling.

So it's got both the description and the direction in one package.

Exactly.

It's computationally efficient because it contains both in one state, and that hybrid nadir gives them dual correctness conditions.

Two ways to be right or wrong.

Indeed.

First, the descriptive fit must be correct.

It has to be the case that the object is physically reachable and graspable by the specified motor means.

Second, the directive fit must be correct.

It has to be the case that the object is actually reached by those specific motoric means, and the action has to conform to the required dynamical pattern.

Success is achieved not just by getting the outcome but by executing the movement skillfully.

Okay.

Moving on to criterion two.

Compositionality and grammar.

We need some specific examples of how motor actions are structured like a language.

We can clearly identify a motor lexicon.

Actions are composed of constituent units.

Reaching, grasping, lifting, rotating, releasing.

These are the building blocks.

The vocabulary.

The vocabulary.

The complexity comes from the grammar.

The rules for assembling these units into a coherent, executable sequence which reflects critical constraints.

Like spatial constraints.

Yes.

Reaching is based on egocentric coordinates where the object is relative to my body.

But grasping is based on allocentric coordinates, the object's shape relative to itself.

The reach has to be executed in a way that the arm's final configuration is compatible with the orientation required for the correct hand grasp.

That's a grammatical rule imposed by spatial logic.

And temporal constraints.

Reaching and grasping are timed, not just sequential.

The hand starts shaping its aperture long before contact, and the MGA occurs at that precise 60 to 70 percent point.

This temporal coordination between the transport phase and the grasping phase is an essential grammatical constraint of skilled action.

And finally, the movement has to show transitivity.

The ultimate goal influences how the subcomponents are executed.

This is maybe the clearest evidence of compositionality being governed by higher level constraints.

The grip you select for a cup is determined not by the cup's geometry alone, but by your intended action.

Are you carrying it to your mouth or are you turning it upside down to rinse it?

Right.

Different goals, different grips, same object.

Exactly.

The entire motor sequence is organized by the future goal, which demands a compositional grammar that adheres not only to kinematics, but also to biomechanical constraints like Fitts' Law.

Let's talk about criterion three, cognitive integration.

We need to show that this non -conceptual system is influenced by and influences our conceptual system.

Integration works in both directions.

First, you have top -down influence.

Motor representations are heavily influenced by high -level goals and stored beliefs.

If you intend to pass a glass of water to a child, your grasp and release will be scaled to ensure a safe transfer rather than a quick toss.

Furthermore, if you believe a pot is much heavier than it appears visually, your initial motor command for grip force will be scaled by that stored belief.

That's a clear case of cognitive penetrability.

This proves that MI isn't just a reactive system.

It's determined in part by the agent's complex motivational states and higher -order goals.

And we also see integration flowing from the motor system back toward the cognitive system, most famously through the mirror system.

The mirror neurons, which fire both during execution and during the observation of the same action, demonstrate motor resonance.

So when I watch a colleague pick up a folder, my own motor representations for picking up a folder activate, even though my hand is completely still.

And this motor simulation is critical for cognitive integration.

It allows you, the observer, to apply your own internal predictive models to infer the goal of the observed action.

This provides essential non -conceptual premises to our higher -level cognitive systems engaged in social interpretation and theory of mind.

It's a big part of how we understand why someone is doing what they're doing.

And don't forget the linguistic link.

Right.

Research has shown that reading action verbs, words like kick, pick, or lick,

someidotopically activates the corresponding areas of the motor and premotor cortex.

The brain region dedicated to controlling the foot gets activated when you simply read the word kick.

That's amazing.

It shows a direct functional link between our conceptual representation of action, which is language, True.

and the non -conceptual motor system.

Okay.

Finally, let's wrap this section up with criterion four, explanatory usefulness.

We've proven the systems are complex and integrated, but why must we call them representations?

Why can't we just stick with a purely mechanistic explanation?

If we only relied on mechanistic explanations, lawful invariant stimulus response correlations, we could maybe explain simple reflexes, but we cannot explain the goal -directed, relationally characterized, context -dependent complexity of human movement.

Give us the classic example again.

Take the simple stimulus of a horizontal bar.

An agent might respond with either an overhand grip or an underhand grip.

The resulting motor behavior is determined entirely by the agent's intention.

Do they want to carry the bar low or high?

Not merely by the sensory input of the bar itself.

Same input, different output.

Exactly.

Since the input is identical, but the output is variable and goal -dependent, we have to appeal to an intentional explanation.

A motor representation of the desired end state and the means to achieve it.

The representational stance is thus fully vindicated.

MI is not non -representational.

It's just non -conceptual representation.

Which brings us right back to Merleau -Ponty's central dilemma.

How do these non -conceptual motor representations seamlessly link up with our high -level conceptual intentions?

We've spent the last two sections demonstrating the distinct nature of motor intentionality.

Now we face the hardest part.

Solving the challenge of projection.

Merleau -Ponty's centrifugal function.

How does the abstract thought, the conceptual intention, translate into the specific proprietary format needed for motor execution?

This is the interface problem, as defined by Butterfill and Senegalia.

And why is this coordination so difficult to explain?

What's the core of the problem?

The difficulty stems from the fundamental difference in formats.

Intentions are typically propositional attitudes.

They're focused on reasons, rationality, and high -level goals, like I intend to make tea.

Motor representations, on the other hand, have this proprietary non -conceptual format.

They're focused purely on execution, control, and dynamic relations, like grip force X, transport speed Y.

The interface problem demands an explanation for how the content of these two disparate formats coordinates so that the action outcomes they specify non -accidentally align.

We need a content -preserving bridge.

Okay, so let's quickly review why some of the promising alternatives fail to solve this problem.

Starting with the simplest idea.

The common cause approach.

This approach suggests that an intention and a motor representation coordinate because they share a common trigger.

For example, the perception of the coffee mug triggers both the propositional intention, I want to drink, and the motor representation grasping configuration X.

The formats don't need to interact directly.

They just need a common start button.

But if my intention isn't triggered by an immediate site, say, I decide over breakfast that I'm going to write an email later this afternoon, the common cause approach totally fails.

Exactly.

It's just too limited.

It only works for reactive actions driven by immediate stimuli, and it completely fails when intentions arise from complex planning, deliberation, or memory retrieval.

It can't explain integrated human agency.

Okay, so that one's out.

The second approach comes from Wayne Wu, and it appeals to intention -guided attention.

He was trying to solve the problem of selection, the many, many problem, how we select the right action for the right target out of a chaotic field of possibilities.

Wu proposed that intentions serve as structural causes.

They deploy concepts, like the conceptual knowledge of a specific object, say, hammer, which then directs your attention to the relevant object in the perceptual field, and that solves the selection problem.

So you think hammer.

You look to the hammer.

Then the object selection leads to the action.

The critical second step in Wu's model is that the high -level action concept deployed in the intention, like grab or lift,

then activates the appropriate motor representation, the proprietary execution file.

And the critique here is?

Well, the critique is that the second step effectively begs the question of the interface.

We started by asking how a conceptual action, like grab,

activates the precise non -conceptual motor format required for execution.

By just saying it activates it, we haven't provided the mechanism for that content -preserving translation.

We've just stated that the link exists.

It's just relabeling the magic, not actually explaining it.

Precisely.

The third approach, proposed by the authors who defined the problem, Butterfill and Sinigaldia, suggested demonstrative deference.

The idea here was that intentions use demonstrative concepts, like do that action.

And that just defers directly to the motor representation that specifies the action outcome.

So you avoid format translation entirely.

It's an elegant philosophical solution.

The proprietary non -conceptual motor representation ultimately determines what the conceptual intention is referencing.

However, Mylopolis and Pachery raised significant objections, pointing out complex disanalogies with ordinary demonstrative reference.

It proves really difficult to explain how a conceptual system can reliably refer to content contained in a proprietary non -conceptual format without some internal mechanism doing the translating or linking.

So we need an actual stable structural bridge between the conceptual format and the proprietary format.

This brings us to the most detailed and promising solution in the source material, the motor schema solution.

The core idea is that intentions must deploy special constituents called executable action concepts.

This is the mechanism we were missing.

Mylopolis and Pachery argue that for an intention to properly bridge this gap, it has to contain a concept that the motor system can actually understand.

And to possess an executable action concept for a type of action, like grasping a cup, you have to possess a corresponding motor schema.

Okay, let's nail down this terminology distinction one final time because it is absolutely crucial.

How is a motor schema different from a motor representation?

The motor representation is the specific temporary execution flow.

It's generated in real time for this cup right now with this specific weight and location.

It's ephemeral and highly detailed.

It's the one -off instruction set.

It is.

A motor schema, however, is an abstract enduring representation.

It's much more stable in general.

Schemas store knowledge about the invariant aspects and the general abstract form of an action type.

For example, the general rules for a power grasp versus a precision grip.

So the schema is the generalized blueprint or the template for grasping and the motor representation is the specific instruction set that's derived from that template for a particular situation.

Exactly.

The schema is the bridge that solves the integration problem.

Schemas are acquired through inductive generalization.

You'll learn them from many specific motor representations over time.

And once you've acquired it, the schema provides the link both upward and downward.

How does that link actually work, then?

The conceptual rational intention activates the executable action concept, which is the general blueprint, the motor schema.

When it's activated, that schema requires specific parameters to generate the action.

It needs details like location, size, and function.

Which you get from perception.

Which are often supplied by attentional processes or visual perception.

And given those parameters, the schema then yields the precise dynamic motor representation suitable for execution.

This framework beautifully replaces Merle -Ponty's metaphorical description of projection with a concrete structural mechanism.

The centrifugal function, that translation of abstract thought into concrete movement, is achieved because the cognitive intention accesses a knowledge structure, the schema, that is intrinsically linked to the proprietary motor format.

It ensures a necessary content -preserving coordination between the high -level reasons we have for acting and the proprietary non -conceptual format that's required for bodily execution.

This motor schema solution offers the most robust representational path we have toward understanding integrated rational and embodied agency.

We have completed a massive intellectual loop today, starting with the philosophical ambiguity of a wounded soldier's movements back in 1945, and ending with a detailed computational model for how thought actually translates into action.

We covered a lot of ground.

We highlighted the distinction between motor and cognitive intentionality, proving its necessity through empirical double dissociations, showing that the action system, the dorsal stream, and the perception system, the ventral, can work separately.

We saw that in patients like D .F.

and in the Tischner illusion.

We then argued that the true contrast is not between a representational and a non -representational system, but between conceptual and non -conceptual forms of intentionality.

We showed that motor representations satisfy modern philosophical criteria because they are compositional, dynamic, and cognitively penetrable.

And most critically, we addressed the interface problem, that lingering mystery of Merleau -Ponty's projection function.

We saw that abstract, rational intentions are seamlessly realized through concrete action only when they deploy executable action concepts that are tied to underlying abstract motor schemas.

These schemas are the necessary blueprints for integrated action.

The journey from a philosophical metaphor to a concrete mechanism, from projection to motor schemas and internal forward models, it really demonstrates the sheer power of this kind of multidisciplinary deep dive.

It shows that by adopting a detailed representational approach, we can successfully explain how our high -level rational agency is seamlessly embodied and enacted in the world.

So what does this all mean for you?

Here's the final provocative thought we want to leave you with.

If our complex intentions are fundamentally constrained by the motor schemas we possess, the action blueprints we have already built, does that mean the limits of your embodied skill set ultimately impose limits on the complexity and specificity of your rational intentions?

You can conceptually desire anything, but can you truly intend to perform a complex action if your body hasn't already mapped the route?

This approach suggests the answer to Merleau -Ponty's projection mystery is,

you can't rationally intend it if your body hasn't already acquired the schema.

That's something to think about until our next deep dive.