Chapter 27: How Revisionary Are 4E Accounts?

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, the place where we turn complex research into, well, clear and captivating knowledge just for you.

Today we are wrestling with one of philosophy and cognitive sciences most enduring and frankly stubborn questions.

A really big one.

Yeah.

How exactly do human beings figure each other out?

Do we see into souls or do we, you know, just read bodies?

It really is the ultimate social puzzle.

And for decades, the answer felt,

well, it felt almost settled.

Social cognition, the way we understand others, was defined almost entirely by what we call the traditional mind -dreading account.

But then a seismic shift happened.

The 4E movement arrived.

That's the framework based on embodied, embedded, and active and extended cognition.

Our mission today is to dive deep into a critical note that appears in the Oxford Handbook of 4E Cognition, which evaluates the whole thing.

And the central question is, how revisionary are these 4E accounts of social cognition?

I mean, do they just scrap the old rules or do they just force a long overdue rewrite?

That's the core tension we're exploring, enrichment versus outright rejection.

This is a deep, highly philosophical dive into the very nature of human connection.

And we're going to focus strictly on the arguments and counter -arguments laid out in this specific critical analysis.

So to start, we really have to establish the intellectual landscape.

We have to know what the 4E theorists were actually challenging.

We absolutely have to.

We have to establish the baseline, the traditional mind -dreading account or MA for short.

This research program has just dominated since the late 1970s.

And MA characterizes human social cognition as fundamentally revolving around the capacity to mind -read.

Which means understanding other agents'

internal, unobservable mental states like their beliefs, their desires, their intentions in order to successfully explain or, you know, predict their behavior.

So if I were watching someone, I'm not just seeing a body moving through space.

I'm trying to solve for all the hidden variables inside their skull.

Exactly.

And this is critical.

The MA tradition studies this capacity from an individualist, internalist, and strictly cognitivist perspective.

OK, let's break that down.

It defines mind -reading as a set of representational cognitive capacities that are realized solely within the individual's brain.

The environment gives you input, sure, but all the heavy lifting, the interpretation, the calculation, that's an internal process of representation and computation.

And if we think about the two main ways MA proposed this internal representation happens, we have two classic really powerful theories.

Theory theory, or TT,

and simulation theory, ST.

We should probably clarify those since they're the main targets of the 4E critique.

They are absolutely the twin pillars of MA.

So theory theory, or TT, it suggests that we possess or develop a folk psychological theory.

Like a theory for everyday people.

Yeah, think of it as an intuitive internal physics engine, but for people.

We use these abstract laws or rules like people generally act to satisfy their desires given their beliefs about the world, to infer unseen mental states from observed behavior and the context.

So you're like a miniature scientist.

A miniature, highly intuitive scientist constantly running hypotheses about the internal causes of external action.

Right.

The classic example is seeing someone walking toward a vending machine.

My theory theory tells me, OK, they desire a snack, and they believe this machine contains that snack.

It's a deductive process based on general rules.

Precisely.

Now, simulation theory, on the other hand, suggests that instead of inferring based on these abstract rules, you achieve understanding by using your own mind as the model.

How does that work?

You internally simulate what you would think, feel or do if you were placed in that person's exact physical and informational situation.

You sort of project yourself, run the mental simulation, and then you attribute the resulting simulated state to the other person.

So both TT and ST, despite their differences in how they do it, are fundamentally internal.

They're representational.

And they're focused on attributing these hidden mental states to the other person.

That is the essential takeaway.

And the methodology reflects this internal focus.

The research leans so heavily on behavioral tasks designed to test this individual capacity for representation.

Like the false belief task.

The study of children's understanding of false beliefs is

When a child realizes someone can hold a belief that contradicts reality,

that's a huge milestone.

And then you have neuroscientists like Sachs and her colleagues back in 2004, who famously used neuroimaging to map the parts of the brain that light up when we perform these mental calculations.

They were trying to find where in the brain this happens.

They were essentially localizing where this internal capacity for belief attribution is physically realized.

So that's the baseline.

An internal, representational, individual capacity.

And that's what the 4E movement set out to challenge.

Okay, let's unpack the challenge.

Around the early 2000s, the criticisms against MA started crystallizing.

And they often aligned perfectly with these emerging 4E principles.

And this wasn't just methodological nitpicking.

No, it was a fundamental challenge across four major conceptual fronts.

It targeted the very scope of the MA project.

So what was the first critique?

The first critique was about phenomenological scope.

Critics argued that the traditional MnDreading account was fatally flawed because it focused way too narrowly on third -person detached observation.

Treating another person like an object to study, like a lab specimen whose internal workings you have to deduce.

Exactly, and they said this failed utterly to characterize the socio -cognitive processes that really dominate our lives.

Second person, unreflective, dynamic social interactions.

That distinction is so key.

The difference between passively watching a stranger on a street corner, treating them as a he or a she that's third person, versus being caught in a rapid fire argument with a friend, where they're an immediate you, that's second person.

When I'm in that argument, I'm not hitting pause to run my internal theory engine, it's too fast.

Absolutely, the speed and the fluidity of second person interactions simply don't seem to allow for constant slow representational calculations.

And this realization led directly to the second critique, expanding psychological processes beyond just the TT and ST monopoly.

So if interaction felt so different, there must be another way we do it.

Researchers started positing these non -inferential alternatives, and the most compelling one was the idea of direct perception.

Direct perception, okay, for a non -philosopher, that sounds almost magical, can you elaborate on that?

It's the idea that we can appreciate certain aspects of others' mental states, particularly emotions, intentions, or immediate action goals, through a non -inferential psychological process.

We don't deduce their anger based on a rule, we directly perceive the anger in their expressive movement, in their posture, in their tone of voice.

So the meaning is just visible, it's on the surface.

The meaning is visible in the physical dynamics themselves.

It bypasses the need for an internal propositional representation, like an abstract sentence in your head about their mind.

So the emotion isn't hidden inside their head, it's broadcast and perceivable on the surface of their behavior.

And this leads to the third, and maybe the most revolutionary claim, doesn't it?

It does.

The critique about deprioritizing mind -dreading.

If direct perception and other non -inferential, basic socio -cognitive processes are enough for most of social life things, like, you know, physical coordination, aligning your posture, mirroring someone.

The simple stuff.

Right.

If that's enough, then some advocates argued that mind -dreading, the full -blown attribution of beliefs and desires, is actually rarely used in everyday, routine life.

They said second -person interactions, guided by these fast, non -mentalistic processes, predominate.

Wow.

So MA is describing the exception, not the rule.

That's the stake.

The core capacity that defines social science for 30 years is suddenly relegated to a rare, slow backup system we only use when we're confused or analyzing a stranger.

Which is a huge claim.

It's a massive claim.

And the fourth critique supported this with an appeal to our subjective experience, the conscious experience mismatch.

If the traditional MA was correct, critics said,

much of our social existence would involve conscious, slow reflection on the causes of people's behavior.

We'd always be thinking, why did he say that?

What does she believe I know?

Which is exhausting and not what life feels like.

Right.

Our lived experience is overwhelmingly fast, unreflective and seamless.

So the direct perception view and other non -mind -dreading processes were thought to align much better with that immediate, unreflective experience of blowing through social life.

It's clear how these four critiques, scope, processes, priority and experience naturally fold into the whole 4 -E framework.

When we talk about these dynamic non -mind -dreading processes, we're talking about embodied cognition, emphasizing the body's role.

Absolutely.

The new rapid processes are fundamentally embodied.

They're rooted in our shared physical structure and our interaction capabilities.

And what's more, the strong version of the critique, which is exemplified by the inactivist approach, and we will discuss that with Diego, it adopts these four points entirely.

And what does inactivism argue?

Inactivism argues that social understanding doesn't happen before an interaction.

It actually emerges through the interaction itself.

OK, so we have the dominant MA and this powerful four -pronged 4 -E challenge.

But as the critique chapter notes, MA didn't just roll over and disappear.

Not at all.

Traditional cognitivists are resourceful.

They responded to these criticisms without converting to 4 -E.

How did they do that?

They accepted them as modest refinements of MA.

They acknowledged the validity of the scope and the speed complaints.

And this led to what is often called the dual process approach, pioneered by researchers like Aperly and Butterfill.

Tell us more about that.

How does a dual process model satisfy the critique while still maintaining that internalist MA commitment?

Well, it keeps the traditional cognitive architecture, but it splits the work.

It recognizes that humans use both fast, dynamic and non -reflective social processes and slow reflective thought about people's minds.

The key is that they characterize both of these processes within a traditional cognitivist framework.

So you say there are two systems?

They assign the faster processes to minimal or implicit mind -reading capacities that still involve representations, even if those representations are simpler or maybe unconscious.

So the MA defenders essentially said, OK, you're right, social life isn't always slow, but we don't need these radical 4 -E ideas to explain the fast parts.

Our brain just has two different interacting systems for representing beliefs, a slow explicit one and a fast implicit one.

That's the defense in a nutshell.

And the phenomenological critique, that difference between fast interaction and slow reflection,

that becomes the defining characteristic that separates the two systems.

And this actually drove a major revolution in how developmental psychologists now design their experiments to try and isolate that implicit system.

And what about the embedded part of 4 -E?

Did they concede anything there?

They did.

Traditional work also adopted a principle called minimal embeddedness.

Which sounds like a partial concession to the 4 -E idea of embedded cognition.

It is.

It's a small concession.

It means traditional researchers accept that an agent's interactions with their social environment are essential for accurately characterizing the internal cognitive resources being used.

The environment matters because it contextualizes the internal mechanism.

OK, so.

Think of it like this.

The computer inside your head is the mechanism and the social environment is the operating system and the constant stream of data it processes.

So the environment is causally essential, but it's context and input to the internal mechanism, not a compositional part of the mechanism itself.

Yeah.

It influences the calculation, but it isn't the calculator.

A perfect articulation of minimal embeddedness.

In this context, sets the stage for the critical notes main evaluation.

The author anchors the analysis of the five 4 -E focused chapters around two central defining issues.

OK, what are those two anchors?

First, pluralism and non -mind -dreading processes.

To what extent do the 4 -E accounts genuinely support the expansion of social cognition beyond just TT and ST and truly emphasize non -mind -dreading processes?

And the second.

Second,

theoretical specialness of social interaction.

Is there a unique fundamental form of social understanding that is inherently tied to the act of social interaction itself?

Or can it always be reduced back to internal individual processes?

Got it.

The rest of our deep dive will evaluate the chapters against those two demanding criteria.

So we start with the first question, pluralism.

We'll look at Galiz and Senegalia, Nguyen and De Bruijn.

Galiz and Senegalia focus on the concept of embodied simulation.

Yes.

How does their position measure up as a challenge to MA?

The critique assesses their account as a modest challenge.

While they invoke the embodied pillar of 4 -E, they remain firmly within a cognitivist framework.

They rely on cognitive processes involving neural representations.

It's just that these representations are bodily formatted.

They leverage the motor system.

So still internal, still representational.

But the content of the representation is physical or motoric rather than some abstract belief desire sentence.

Exactly.

Their breakthrough and the source of the challenge is that they articulate a way to understand others' action goals, the outcomes of their movements, without necessarily needing to attribute abstract internal beliefs or desires.

And this is all driven by embodied simulation.

Can you walk us through ES again, maybe with a simple example?

Of course.

Imagine watching someone reach for a glass of water.

According to ES, your understanding of their goal is achieved by reusing the processes and representations that would occur if the observer herself were planning to execute those actions.

So my brain acts like I'm about to reach for it too?

Precisely.

The mirror neuron system, which is the neural basis for this, fires in your brain as if you were preparing to execute that reach.

And this gives you immediate non -inferential access to the action's goal.

And this is deemed not mind -reading, crucially.

Why is that strong distinction made?

Because Galiz and Sinigaglia conceive of goals non -mentalistically.

They define a goal simply as the physical outcome or endpoint of an action, the hand grasping the cup, rather than the complex internal propositional mental state, like a desire to quench thirst or a belief that the cup contains water.

So they redefine the target?

By redefining the object of understanding from internal mental states to external action outcomes, they claim to bypass mind -reading.

Wait, I have to interrupt here.

If I define a key philosophical term goal in a way that makes my theory look non -mentalistic, isn't that a bit of a semantic trick?

I mean, the traditional M .A.

defender would say that you still need a belief or desire to fully explain the reasons for the action, not just the physical endpoint.

Is that part of the critique?

That is the essence of the critique's assessment.

The critical note finds that this E .S.

driven understanding is pretty basic.

It's maybe a fundamental precognitive module or a developmental precursor to full mind -reading, and it's perfectly compatible with it in adults.

So it's not a replacement?

No.

And Galiz and Sinigaglia failed to deliver a true revisionary challenge because they don't make any strong claims about the relative importance or frequency of E .S.

So they introduce a new tool but don't claim it should displace the old ones.

Exactly.

Radical 4E advocates argue that non -mentalistic processes like E .S.

are the norm in social life, but Galiz and Sinigaglia, they don't make that assertion.

Plus, E .S.

is just as useful for passively observing someone as it is for active social interaction.

Which means it doesn't support the 4E claim about the specialness of interaction.

Right.

And since they admit that genuine mind -reading still provides a richer understanding of the reasons behind actions,

their account is largely compatible with M .A., which is increasingly adopting these dual process models that include simple, non -mentalistic precursors.

It's an addition to the toolkit, not a structural demolition.

Okay, so if Galiz is modest, let's move to Nguyen, who explicitly embraces pluralism with his person model theory, BMT.

This is an attempt to integrate traditional M .A.

with 4E ideas.

This sounds like the textbook definition of enrichment.

Nguyen's framework really does attempt to create a complete catalog of the ways we understand others.

He includes the conscious, explicit use of TT and ST, but he heavily incorporates implicit strategies that are aligned with 4E.

Like what?

Like direct perception, situation models, which are scripts for conventional behavior, and even participatory sense making, which he views as a potential example of extended social cognition.

What makes PMT novel beyond just listing a bunch of different strategies?

The emphasis on specific idiosyncratic knowledge of individual persons.

Nguyen argues that after years of interacting with your spouse or your colleague or a close friend, you don't use abstract theory rules anymore.

You build a highly detailed, personalized model of that person's habits, tendencies, and predictable reactions, which vastly speeds up social comprehension.

That feels intuitively right.

I don't need a general theory to know my friend will order the same obscure coffee every time I just know them.

Right.

Now, the critique strongly endorses this pluralistic direction, but it raises serious conceptual questions about how Nguyen actually delineates these strategies.

The central confusion revolves around three clarity issues regarding his epistemic strategies or ways of knowing.

Okay, let's start with the first one.

The overlap issue between TTST and direct perception.

So Nguyen argues that TTST poorly capture that intuitive, fast social cognition and that direct perception is the alternative.

But the critique counters that modern MA is already sufficiently flexible.

Many MA researchers argue that TTST are not just theories of explicit reflection.

They also explain implicit intuitive understanding.

Ah, so MA already claims that territory.

It does.

So if MA is already covering intuitive processes, Nguyen has to do a better job of distinguishing his definition of direct perception as a unique non -MA strategy.

If he doesn't, traditional mind -rators can just absorb direct perception to their existing framework by saying, oh, our simulation theory is just so fast that it feels like direct perception.

The distinction becomes philosophical, not empirical.

Precisely.

The boundary needs to be clearer.

Okay, what's the second clarity issue?

The scope of information.

Nguyen isn't clear on whether all these different strategies, TTST, direct perception, situation models, are all capable of representing the same information.

So what can each tool do?

For example, he mostly discusses direct perception in the context of immediately observable basic mental states, like emotions.

Can direct perception handle complex, higher -order beliefs, such as, she believes that I know she lied to the boss last week.

Probably not.

That seems like a stretch.

If the answer is no, then the strategy has a hard limitation.

If direct perception is limited to basic emotions and intentions, it really restricts the utility of PMT as a true alternative to TTST, because TTST are precisely designed to handle those complex propositional beliefs.

It needs to clarify the boundaries of what each strategy can actually deliver.

Exactly.

And the third critique goes right to the heart of the four EMA attention, mental state priority.

Nguyen wants to incorporate non -mind -reading processes, aligning with the four E critique that we shouldn't overemphasize mental states.

However, when he's defining his own theory, he states that the PMT's main claim concerns the strategies humans use to access the mental states of others.

Oh, so it brings it right back to mental states.

The critique finds a real ambiguity there.

If the ultimate goal is still accessing mental states, even with these situation models, then how revisionary is the theory, really?

It's highly ambiguous.

Take situation models.

These are scripts for routine behaviors like how to behave at a funeral or order a coffee.

Do those scripts contain zero mental state representations?

That's the question.

When I use a script, do I completely bypass the thought that the barista intends to serve me?

Nguyen is vague on this.

If TNT ultimately reintroduces mental state representation, even implicitly, it just diminishes his endorsement of the four E criticism that we should deprioritize mind -reading.

So the critique suggests Nguyen has created a great pluralistic catalog, but he hasn't fully delineated the boundary between MA and four E.

Right.

It sounds like he's trying to please two different philosophical communities at once, and it results in a blurred line about what the output of social cognition really is.

Let's move to De Bruijn, who uses the very popular predictive processing paradigm, PPP, to tackle a specific developmental problem.

This is a very different kind of philosophical maneuver.

It is.

De Bruijn applies PPP to address the notorious developmental paradox in children's false belief understanding.

Right.

The paradox where young children often pass spontaneous response, false belief, tasks like looking, where an agent falsely believes an object to, years before they can pass elicited response tasks, where you explicitly ask them, where will Sally look for the marble?

Exactly.

Why do children demonstrate this implicit understanding but fail to answer the direct question?

It's a huge problem for MA, which assumes one singular internal capacity for belief representation.

So how does PPP solve this?

De Bruijn uses PPP, which is a powerful general framework in cognitive science.

In PPP, the brain is modeled as a hierarchical network of generative models constantly trying to predict sensory input.

And all cognitive work is the process of minimizing prediction error.

PPP suggests two strategies for minimizing error.

Perceptual inference, which is updating your internal representations or mental models, and active inference, which is interacting with the environment to minimize error.

So if I predict the door is locked and it's not, I can update my internal model, that's perceptual inference, or I can just push the door open to resolve the error that's active inference.

That's the mechanism.

And De Bruijn's solution to the developmental paradox hinges entirely on the task design constraining the child's preferred inference strategy.

What does he argue?

He argues that the spontaneous response tasks he studies, active helping tasks, allow children to use active inference.

They can physically interact with the target agent or the environment to resolve the prediction error immediately.

The child gets to do something to solve the problem.

Whereas the elicited response tasks, the ones that require the child to sit and answer the experimenter's question, they force the child to use perceptual inference.

They have to update their internal representational models of the agent's beliefs.

And young kids can't do that.

De Bruijn suggests younger children fail these elicited tasks because they lack the necessary recognition to model false beliefs accurately via internal representation.

They just lack the full internal mind -reading capacity.

That seems like a really neat division of labor, explained by PPP.

But the critical note strongly challenges both the empirical adequacy and the novelty of this solution.

What's the first empirical issue?

The conflict with existing spontaneous task evidence.

De Bruijn's model asserts that young children fail the elicited task because they incorrectly model the agent as having a true belief about the object's current location.

In other words, he says the young child doesn't truly grasp the false belief.

But the critique points out that we have extensive evidence from other types of spontaneous tasks, like violation of expectation, where children do show anticipatory looking behavior.

They correctly expect the agent to look in the falsely believed location.

Exactly.

The children's anticipatory behavior strongly suggests they possess a minimal implicit understanding of the false belief, which directly contradicts De Bruijn's claim that they lack the genuine understanding to model it accurately.

And what about the crucial philosophical critique?

Does PPP actually provide a novel insight here, or is it just new language for an old idea?

The critique finds that De Bruijn's core claim that belief representations are not required for spontaneous tasks only for elicited tasks is logically independent of PPP's core tenets.

That observation about task differences was already part of the dual process MA refinement.

So PPP isn't doing the real work here.

PPP, with its generative models, is merely a new theoretical vocabulary to describe the underlying psychological processes.

It offers no clear empirical reason to prefer De Bruijn's model over existing minimal mind -reading or non -mentalistic accounts already being debated in the literature.

So we've found that PPP here is a sophisticated conceptual framework, but applying it doesn't resolve the developmental paradox.

It just restates the problem in terms of different inference strategies.

And the implications for Fourier are therefore pretty modest, as De Bruijn himself admits.

PPP's inherent emphasis on rich internal generative models makes it very difficult to reconcile with the radical claims of extended cognition.

At best, it's compatible with embedded cognition and the modest side of embodied cognition.

A small non -revisionary win for the Fourier approach, then.

More about compatibility than revolution.

Okay, here's where it gets really interesting.

As we shift focus to the second core issue, the theoretical specialness of social interaction,

does the act of interaction itself create a unique form of social understanding that cannot be reduced to individual internal processes?

We look at Reddy and the Jaeger.

Let's start with Reddy's chapter.

Why engagement?

Reddy's contribution focuses heavily on the developmental necessity of social interaction.

She posits that children's second -person engagements with adults are developmentally crucial for establishing typical social cognition,

specifically the foundational understanding of attention and intention.

She seems to use the term engagement to capture something richer than just interaction.

What is her definition supposed to achieve?

She wants it to capture the unique emotional qualities and the person -ness of the other.

She defines engagement as an acknowledgement of or response to the person -ness of the other, which requires an emotionally charged mutual awareness of that person's mental states.

That sounds powerful.

It suggests engagement is inherently social.

But the critical note immediately finds conceptual holes in her definition based on the examples she uses.

What are the problems?

There are two main problems.

First, the problem of non -personal engagement.

Reddy surprisingly includes interactions with objects as counting for engagement, such as relishing the warmth of a hot shower or an infant's deep attraction to grasping objects.

Wait, a hot shower?

Yes.

So if interaction with a conscious person is not a necessary condition, the term engagement loses its specificity to the social domain.

It just starts sounding like a generic description of intense, effective sensory experience.

If the concept can apply to a person, an object, or a hot shower, it's too broad to ground a theory of social cognition.

Exactly.

And the second problem is the conflict with mutuality.

Reddy claims that emotional involvement without action, such as the intense, responsive sympathy you feel while watching a movie, also counts as engagement.

And this is where the vagueness about the second -person stance arises.

She defines the second -person stance as seeing the other as you and experiencing being addressed by another, coupled with a necessary feeling of mutuality and reciprocity.

But if mutuality is essential, how can you genuinely take this stance during a purely observational context, like watching a movie?

The character on screen is not recognizing or responding to you.

That's the issue.

This raises a fundamental conceptual question for Reddy.

Is an actual causal response from the other person required, or is simply imagining or projecting a response sufficient to generate the stance?

The critique argues that Reddy's concepts of engagement and second -person stance largely overlap and, crucially, both fail to adequately distinguish between the individual's internal cognitive stance, the subjective feeling or belief of interaction, and the actual causal properties of the relationship.

That's the core vulnerability.

If the subjective experience alone is driving the cognitive effects, if just believing you are interacting is enough, then the claim that there is something theoretically special about the actual social interaction itself is undermined.

The explanation gets pulled back into the MA camp where it's all about individual internal process.

Precisely.

But despite these conceptual issues, Reddy did identify a crucial developmental trajectory that is immensely valuable to 4E thinkers, even if it doesn't support the strongest philosophical claims.

Okay, what's that?

Yes.

She outlines compelling evidence that infant's awareness of others' minds develops in a specific mutualistic way.

The infant is first aware of the other's attention directed to themselves, then they notice the other's actions, and finally, they understand attention directed toward distal objects.

So awareness of minds is rooted in that reciprocal second -person engagement focused on the self.

It suggests that, yes, this supports the 4E idea that social interaction is developmentally crucial, providing the scaffolding for a later, more complex understanding.

But the philosophical question remains open.

Which is?

Does this mutualistic development require higher -order mental state representations early on?

Or can the development be fully explained by existing MA concepts?

For instance, maybe simulating another's actions and intentions is simply much easier and therefore develops earlier when the other person is directly attending to the same thing as the self.

So it could just be a difference of cognitive load, not a difference of mechanism.

It could be.

The question isn't settled.

This brings us to the most radical position we review today.

De Jager, who pitches her framework firmly within an activism and defends the highly controversial thesis of extended cognition.

She is the one making the boldest claim about the specialness of social interaction.

De Jager provides a robust, multi -level framework for inner subjectivity, correctly emphasizing the need to account for both individual subjectivity and the dynamics of social interaction, from the physiological level all the way up to the social group level.

The critique fully accepts this methodological necessity.

The problem arises when she claims that social interaction can constitute social cognition.

Let's clarify constitute.

In philosophy, that means the social interaction is a necessary compositional part of the cognitive mechanism.

It's not just a causal input.

It's a piece of the machine itself.

Exactly.

A good analogy is a computer program.

A causal factor is the power core.

If you unplug it, the program stops, but the cord isn't part of the program.

A constitutive factor would be a line of code or a specific register within the CPU.

De Jager claims the interaction is the line of code.

And the critique argues she makes this radical jump based on a confusion of compositional levels.

What does that confusion of levels look like in practice?

It manifests when discussing phenomena like interpersonal synchrony, the coordination of neurological, physiological, or behavioral processes between two or more individuals.

Synchrony is, by definition, a social level phenomenon.

De Jager contrasts the traditional view where interaction is context for an internal mechanism with her view, where the social interaction processes constitute the synchrony and therefore part of the social understanding that takes place.

Wait a minute.

If synchrony is a group phenomenon, of course the social dynamic constitutes the synchrony.

That's trivially true, isn't it?

But the critique is,

is social understanding referring to that social level synchrony, or is it referring to the individual level cognitive process of each agent?

That is the ambiguity that breaks the argument.

The critique asks, does the social interaction constitute the internal cognitive processes driving the individual agent's behavior?

The contingent perception system in person A cannot alone explain the synchrony, sure, but the interaction may only be input to that system, not a part of it.

So she hasn't shown how the boundaries of the individual process have expanded.

Not clearly.

And this issue comes into sharp focus with her key empirical example,

perceptual crossing experiments.

In perceptual crossing, two agents interact dynamically to search for a target, and the collective dynamics simplify the sensory task for each individual agent.

Dijager uses this as evidence for extended cognition.

However, the critique argues these experiments are plausibly better analyzed as highly sophisticated cases of embedded cognition.

Why embedded and not extended in this specific case?

The critique uses a mechanistic philosophy of science framework to analyze the system.

In this view, the cognitive task, the search behavior of each individual, is constitutively explained by internal neural mechanisms.

The social environment, the interaction, simply streamlines the complexity of the task.

It acts as context.

It reduces the processing load on the internal mechanism.

So the interaction is causally essential.

It makes the task possible or much easier.

But we can still draw a clear compositional boundary around the individual's cognitive process.

The interaction is the high -speed modem.

Not the operating system itself.

Precisely.

The individual agent takes sensory input from the interaction and performs actions that affect it.

These are causal interactions.

To claim that these external causal loops constitute the individual's internal process requires a far deeper philosophical defense regarding how we demarcate a cognitive system.

That brings us to the necessity of deeper philosophical tools, like mutual manipulability.

Can you explain that concept simply for the listener and why it's needed here?

Certainly.

Mutual manipulability is a concept from mechanistic philosophy of science used to define the boundaries of a mechanism.

You identify a mechanism's components by showing that if you intervene on component A, component B changes its activity.

And if you intervene on component B, component A changes its activity.

The components have to affect each other.

The components that are mutually manipulable constitute the mechanism.

So if I manipulate the social interaction, the external component, and the individual's internal cognitive process changes, that's just a causal link, like unplugging the power cord.

To prove the extended claim, de Jager would need to show that the individual internal mechanism, say a specific neural contingency system and the external social interaction, are so tightly integrated and mutually manipulable that you cannot isolate the internal system without losing the cognitive phenomenon altogether.

The critique finds that de Jager's current work lacks this necessary philosophical depth to support the radical claim of constitution.

Finally, de Jager, like other 4E critics, argues that inactivism implies social cognition is not in the first place a question of figuring out each other's mental states.

And the critique strongly pushes back on this conclusion.

It does.

The critical note firmly maintains that accepting the importance of non -mind -reading processes in a pluralistic view does not logically entail rejecting the priority of mind -reading.

Just because there are other ways doesn't mean the main way is gone.

Right.

While traditional MA may have been wrong to ignore interaction, recent MA refinements have successfully incorporated non -mind -reading systems and acknowledge direct perception.

The core argument is, identifying alternatives is not the same as displacing the core function.

The frequency of mind -reading is an empirical question, not a philosophical one settled by identifying alternatives.

Absolutely.

The critic finds that de Jager and others move too quickly from other processes exist to the conclusion that MA should be displaced from its central position.

For example, when social norms break down or when you're dealing with complex multi -agent systems, the slow internal capacity for propositional mind -reading may become essential again.

Suggesting it remains the ultimate complex tool for navigating difficult social reality.

Exactly.

So what does this all mean?

We've reviewed five distinct attempts to integrate, refine, or outright replace the traditional mind -reading account using the 4E framework.

What is the final consensus here regarding the revisionary power of 4E accounts of social cognition?

The overall finding is that the chapters reviewed most firmly established the importance of embedded cognition.

This is the definitive non -controversial win for the 4E movement.

So that's the big takeaway.

The appreciation that social context and environment are fundamentally vital for characterizing and understanding the socio -cognitive processes driving agents' behavior.

No one can ignore the social environment as mere background noise anymore.

And tracing the spectrum of embodied and extended claims.

The modest form of embodiment as endorsed by Galiz and Senegalia, the idea of embodied simulation, is empirically well supported but remains non -revisionary.

It fits easily into the MA supplemented pluralism of the dual process approach.

It's an accepted new layer.

And the radical claim.

The truly radical thesis of extended cognition, championed by de Jager, remains the most exciting but also the most conceptually murky.

Its current defense is hobbled by the lack of clarity on how to demarcate the compositional boundaries between individual and social levels.

The final verdict then is that the 4E accounts, as analyzed in this critical note, offer a powerful modest challenge that has irrevocably pushed traditional MA to broaden its scope and adopt pluralistic models.

But they have yet to achieve the fundamental goal of rejecting or displacing MA's core claims about the importance of internal, representational, mind -reading capacity.

The capacity for representing beliefs, desires, and intentions, the traditional MA target, is still very much the center of the debate.

4E has forced MA to share the spotlight with basic, non -mentalistic processes, but it hasn't proven that those processes constitute the primary or most sophisticated form of human understanding.

That brings us to our final provocative thought for you to mull over after this deep dive.

We've established that early social development is highly dependent on second -person engagement directed toward the self.

And we've also accepted that non -mentalistic processes exist in our adult interactions.

Given all this, the critical question for you to explore remains.

If those fast embodied non -mentalistic scripts break down, say, when you move to a completely new culture, or when you encounter a highly erratic social actor, how quickly does the mature brain abandon those 4E strategies?

And how much does it revert to the slow, effortful, internal, representation -hungry calculations that characterize the traditional mind -reading account just to navigate unfamiliar or challenging social environments?

We might need MA most when we think we need it least.