Merleau-Ponty on shared emotions and the joint ownership thesis

Abstract

In “The Child’s Relations with Others,” Merleau-Ponty argues that certain early experiences are jointly owned in that they are numerically single experiences that are nevertheless given to more than one subject (e.g., the infant and caregiver). Call this the “joint ownership thesis” (JT). Drawing upon both Merleau-Ponty’s phenomenological analysis, as well as studies of exogenous attention and mutual affect regulation in developmental psychology, I motivate the plausibility of JT. I argue that the phenomenological structure of some early infant–caregiver dyadic exchanges is best described as involving joint subjects. From birth, some experiences are constitutively social in that certain phenomenal states, such as the positive emotions that arise within these early exchanges, are jointly owned. Along the way, I consider a possible objection. I conclude by considering the explanatory significance of adopting JT.

1 Introduction

In “The Child’s Relations with Others,”¹ Merleau-Ponty challenges what he thinks are the cognitivist and individualist assumptions informing contemporary theories of current child development. In place of these theories, he offers what might seem, at first blush, to be a fairly radical view. Merleau-Ponty argues that self-consciousness and social awareness (i.e., awareness of others as inhabitants of a shared world) emerge from a more basic state of self-other interconnectedness. Put this way, the view is fairly prosaic. However, the nature of this interconnectedness is what ultimately makes the view quite radical. The core feature of Merleau-Ponty’s view here is the claim that certain early experiences are jointly owned, in that they are numerically single experiences that are nevertheless given to more than one subject (e.g., the infant and caregiver). Call this the “joint ownership thesis” (JT).

Drawing upon both Merleau-Ponty’s phenomenological analysis, as well as studies of exogenous attention and mutual affect regulation in developmental psychology, I motivate the plausibility of JT. I argue that the phenomenological structure of some early infant-caregiver dyadic exchanges is best described as involving joint subjects. From birth, some experiences are constitutively social in that certain phenomenal states, such as the positive emotions that arise within these early exchanges, are jointly owned. Along the way, I consider a possible objection. I conclude by considering the explanatory significance of adopting JT. Two larger aims of this paper, then, are to draw attention to a specific thesis of Merleau-Ponty that hasn’t received much attention while, second, providing an example of how this thesis—and, indeed, phenomenology more generally—can make a fruitful contribution to ongoing discussions in cognitive science.²

2 Early emotions and the “joint ownership thesis” (JT)

In “The Child’s Relations with Others,” Merleau-Ponty concerns himself with the character of early infant experience. He argues that, phenomenologically, these early experiences are characterized by the way that others, such as the infant’s caregiver, enter into the experience as a joint subject. The phenomenology of these experiences is a property of this relation and might therefore be ascribed to both subjects. The ontogenesis of these early experiences is significant, according to Merleau-Ponty. It lends insight into the developmental origins of self-consciousness as well as the nature of face-to-face social interaction and interpersonal understanding in mature adults.

Note first the strength of this claim. Surely everyone would accept that emotions can be shared expressively, that is, made public via various behavioral manifestations. When I am angry, say, I share this anger with others via frowning, shaking my fists, speaking loudly, etc. But my anger remains my own in the sense that I remain the sole subject of the episode of anger. Others therefore experience my anger as an object (when they see my expressive behavior or hear my rising voice). And surely the emotions I share in this way can impact others’ behavior and experience; my anger may cause them to become similarly angry or perhaps fearful. But Merleau-Ponty endorses a more radical claim: namely, within some early experiences, the phenomenology is such that two subjects (infant and caregiver) can be said to simultaneously share the same episode of emotion. This is a claim about the ontology of certain early experiences. When I speak of the joint ownership thesis (JT) in what follows, I will be referring to this idea.

What motivates Merleau-Ponty to defend such a radical claim? He sees JT as a corrective to what, according to him, are the pervasive individualistic and cognitivist assumptions governing “psychologists of the classical period.”³ According to Merleau-Ponty, the individualistic assumption is that

the psyche, or the psychic, is what is given to only one person…I alone am able to grasp my psyche—for example, my sensations of green or of red. You will never know them as I know them; you will never experience them in my place. A consequence of this idea is that the psyche of another appears to me as radically inaccessible…I cannot reach other lives, other thought processes, since by hypothesis they are open only to introspection by a single individual: the one who owns them.⁴

An undesirable consequence of this view, according to Merleau-Ponty, is that I can never have direct acquaintance with—and thus epistemic certainty with respect to—the existence of another’s mental life. Rather, “I seize the other’s psyche only indirectly, mediated by its bodily appearances.”⁵ In light of this indirectness, I am forced to rely on analogical reasoning to infer the existence of other minds. This reasoning is what allows me to “decode” another’s behavior and to posit the “hypothetical” existence of a hidden mental life behind it.⁶ According to Merleau-Ponty, then, this individualistic assumption leads to an overtly cognitivist portrayal of interpersonal understanding, one that frames this process in terms of analogical reasoning.

As Merleau-Ponty notes, this analogical process is comprised of four terms: (1) my own “psyche”; (2) the “introceptive image” I have of my own body from the inside, including a felt sense of self-initiated agency; (3) the other’s “visual body” (i.e., their body as an object for my perception); and finally (4) the other’s “psyche” as I “imagine or suppose it across the appearance of the other through his visual body.”⁷ Since I regularly experience a correlation between (1) and (2) in my own case—certain mental states cause characteristic patterns of overt behavior—I infer the existence of analogous correlations in others when I see their “visual body” engaging in certain behaviors.

There are a number of objections to this analogical view that go beyond present concerns.⁸ Rather than survey these objections, I want to instead focus on Merleau-Ponty’s alternative picture. First, Merleau-Ponty observes that this sort of inferential process appears far too cognitively sophisticated to be attributed to infants who nevertheless exhibit “a relative precociousness of the perception of others.”⁹ This precociousness is indeed well-supported by developmental research. For example, very young infants appear to perceive facial displays of emotions directly, without the need for analogical or inferential mediation. They show a preference for human stimuli, particularly faces.¹⁰ Newborns discriminate faces from other stimuli and preferentially track moving face stimuli.¹¹ They can even discriminate different faces, showing a clear preference for their mother’s face and attractive faces.¹² Finally, infants as young as 10 min old imitate facial expressions—but not inanimate objects that simulate these expressions—and do so in a way indicating sensitivity to their emotional significance.¹³

In light of their social precociousness, Merleau-Ponty insists, as John O’Neill puts it, that the young infant “short-circuits this game of [analogical] correspondences with a global body overlap, as when the infant responds to a smiling face with his own smile.”¹⁴ The infant’s context-sensitive responses emerge not from an inferential cognitive mechanism but rather from a “postural” or “corporeal schema” enabling the infant to respond to the conduct of others with similar conduct.¹⁵ As Merleau-Ponty says elsewhere:

A baby of fifteen months opens its mouth if I playfully take one of its fingers between my teeth and pretend to bite it. And yet it has scarcely looked at its own face in a glass, and its teeth are not in any case like mine. The fact that is own mouth and teeth, as it feels them from the inside, are immediately, for it, an apparatus to bite with, and my jaw, as the baby sees it from the outside, is immediately, for it, capable of the same intentions. ‘Biting’ has immediately, for it, an intersubjective significance.¹⁶

Infants directly perceive other’s expressions as soliciting their own kinaesthetic possibilities; they are in this way dynamically and bodily “coupled” to others.¹⁷ Merleau-Ponty argues that the infant’s early social sensitivity is a corporeal, situated responsiveness, and not an inferential or analogical process. We are bodily coupled to one another via “an internal relation which causes the other to appear as the completion of the system.”¹⁸

Secondly, and in light of this infant-caregiver coupling, Merleau-Ponty proposes that the rudiments of self-consciousness emerge from a more basic “state of pre-communication (Max Scheler,) wherein the other’s intentions somehow play across my body while my intentions play across his.” At this stage of the infant’s development, “there is not one individual over against another, but rather an anonymous collectivity, an undifferentiated group life.” Only later, “on the basis of this initial community, both by the objectification of one’s own body and the constitution of the other in his difference, there occurs a segregation, a distinction of individuals” whereby self-consciousness, understood as awareness of oneself as a distinct phenomenal subject, can be said to emerge. Merleau-Ponty thus concludes that “[c]onsciousness of oneself as a unique individual, whose place can be taken by no one else, comes later and is not primitive,” developmentally speaking.¹⁹ This initial “pre-communicative” unity of self and other helps to understand why even young infants are precociously responsive to others’ emotional expressions. Not only do they perceive emotional expressions directly. Moreover, they are quite literally part of the same shared experience.

There is much to unpack here. I look at the details of this claim, as well as empirical evidence appearing to support aspects of it, below. First, however, note that one way to block this view from getting any traction is simply to deny that infants are phenomenally consciousness at this stage of development, and thus are neither conscious in an individual nor a joint sense.²⁰ But this idea is not widely accepted; and it’s certainly not one that Merleau-Ponty would endorse. Indeed, it seems to be contradicted by a wealth of developmental evidence suggesting that from birth infants are genuine social subjects: agile perceivers and actors in meaningful social environments.²¹

We can thus plausibly assume that infants are phenomenally consciousness. Nevertheless, it’s still not clear that one must, as Merleau-Ponty seems to think, affirm JT in order to account for the infant’s early social precociousness. There are alternative accounts one might adopt.²² These views offer various accounts of early social sensitivity that don’t appeal to jointly owned phenomenal states. Of course, these explanations do involve theoretical constructs (e.g., Theory of Mind-style explanations) and experimental data unavailable in Merleau-Ponty’s day. I therefore leave it as an open question whether or not a defense of JT is ultimately necessary in order to account for early social precociousness; this is a larger issue beyond the scope of this paper. Instead, my intention in what follows is more restricted. I want to motivate the plausibility of Merleau-Ponty’s JT by summoning different streams of supporting empirical research. I will, however, conclude by indicating how JT sits alongside current views of infant social cognition (e.g., Theory Theory, Simulation Theory), and suggest how it might enrich these views—even if, ultimately, it doesn’t supplant them.

In sum, Merleau-Ponty appears to be endorsing two claims in the passages cited. First, he suggests that the perceptual world of the infant is not simply a chaotic constellation of simple sensations. Rather, it has some degree of coherence and structure—as it seemingly must for infants to be able to pick out and respond to human faces. As he writes in another lecture on child psychology: “In the child, thanks to the phenomenon of constancy, a nonchaotic and structured vision of the perceptual field exists (though this is not to say that the structuration is the same, or as perfect, as that of the adult).”²³ This structural coherence is rooted in a sense of the body’s location in space: a situated or ecological sense of being a perspective onto a perceivable environment.²⁴ Second, despite this rudimentary structure, the newborn is experientially coupled to the caregiver in a state of phenomenal undifferentiation. This is the phenomenological condition that makes JT possible.

Having clarified Merleau-Ponty’s motivation for and formulation of JT, it is time to look at different streams of empirical evidence that might be summoned to defend this idea. This is the task of the next section.

3 Attention, emotion, and social interaction in infancy

3.1 Regulating attention

We can begin by looking at some details of the relation between attention and perceptual consciousness in infants. This will help us understand how phenomenal states are, at this early stage of development, structurally poised to be shared with others in the deep sense that JT entails.²⁵

Newborn infants less than an hour old are capable of sustaining some degree of attentional control. Though visually impaired, this control allows them to look intently and scan salient features of a scene (within limits, of course) including segregated, distal objects. As we’ve seen, newborns can focus on the facial expressions of others, selectively imitate these expressions, and seemingly appreciate their emotional significance.²⁶ But attentional control manifests in others ways. For example, newborns can also discriminate between self versus externally caused stimulation (touching their own cheek versus someone else doing it), and are capable of making olfactory discriminations between maternal amniotic fluid and the amniotic fluid of a stranger.²⁷ Preterm infants orient toward pleasant sounding music as well as the sound of their mother’s voice.²⁸ Two month-old infants can sustain the auditory attention needed to memorize short melodies and later discriminate the remembered melody from other heard melodies.²⁹ This brief survey suggests that young infants exhibit a surprising degree of attentional and perceptual sophistication when it comes to attending to the faces, voices, smells, and touch of others. Intense periods of face-to-face interaction with caregivers create the shared context for refining these capacities.

Of course, there are significant developmental constraints on the character and degree of this attentional control. What is important for our purposes is that, unlike adult attention, early infant attention is primarily exogenous (bottom-up, involuntary).³⁰ External objects and events catch infant attention and determine both what infants look at and how long they look at it. Adult perceivers, in contrast, exhibit endogenous (top-down, voluntary) control of attentional focus. While writing a paper, for example, I may become momentarily distracted by a noisy conversation outside my office. But I can voluntarily shift my focus away from this distraction and resume my writing. While driving to work, I focus on the road despite the visual allure of local scenery. Skillfully managing attention this way is the mark of a mature perceiver; but this skill is not present in neonates and very young infants. So, while the quality of infant attention is surprisingly rich, as the previously-cited studies indicate, its inhibitory component is comparatively underdeveloped.³¹ This lack of attentional inhibition can be an adaptive advantage when we’re young since there is an evolutionary imperative to learn as much as possible as quickly as possible.³² Broadening the purview of attention, instead of narrowing in on task-specific information, is one way to go about doing this.

This lack of attentional control has phenomenological consequences. In light of underdeveloped inhibitory mechanisms, young infants are deeply dependent upon caregivers to regulate their attention for them. Via various forms of physical interaction—gestures, touch, patterns of holding the infant, sustained eye contact and gaze manipulation, etc.—the latter provide exogenous scaffolding that does just this. While coupled with the caregiver, this scaffolding allows the infant to exceed her developmental constraints and achieve a flexibility and stability of attention well beyond her current developmental level.³³ Together, infant and caregiver function as a coupled social system.³⁴

For example, breastfeeding—perhaps the earliest and most complex form of social interaction the infant engages in—has a structured turn-taking, attention-directing character shaped largely by the mother “jiggling” the infant as a prompt to resume feeding.³⁵ This jiggling organizes the infant’s attention and guides their behavior (i.e., they resume feeding). The infant is an active participant in this process, however; she reliably postpones her sucking until the mother ends her tactile behavior, which determines the timing of the mother’s subsequent responses.³⁶ Nevertheless, the general contour of this early interaction—as well as the direction and focus of the infant’s attention—is shaped by the mother’s external regulation. Likewise, within episodes of imitation, caregivers regulate the direction and intensity of the infant’s attention by holding the infant in an immobile posture and leaning in very close to the infant’s face.

This process of attention-sculpting continues as the child develops. For example, most infants point to objects by the time they are 9 months old; at this stage parents increase their own use of pointing when interacting with infants.³⁷ And while infants initially respond to head movements in adults, by 18 months they track eye movements as well, allowing adults to manipulate the infant’s attention via directed looking.³⁸ These early interactions have enduring long-term effects. For instance, it was found that infants of low-sensitivity mothers maintained lower frequencies of gaze monitoring and ability to coordinate attention with attuned strangers than did infants who had highly sensitive mothers.³⁹

The take-way point from this work is that “adults influence what in the continuous stream of sensory input infants are most aware of, become familiar with, and think most about.”⁴⁰ This is particularly so in the first weeks and months of life when infants have very little inner control of attention and thus are particularly vulnerable to environmental stimulation.

But there is another related point to be made, one which helps support the phenomenological import of Merleau-Ponty’s JT. Inner control of attention is a crucial constituent of self-awareness, the feeling of being a conscious subject. Recognition of the relation between attention and self-awareness is what motivates William James to characterize consciousness as first and foremost a “selecting agency” actively shaping the structure and content of experience.⁴¹ By internally controlling my attention and shaping my experience, I become aware that the experiences I am structuring are mine. I develop a basic form of self-awareness: awareness of myself as a phenomenal subject. But as we’ve seen, very young infants lack inner control of attention; the world largely determines what they look at and experience.

Phenomenologically, then, young infants at this early stage of development might be said to experience the world in a relatively nondual way. Lacking the capacity to self-regulate attention, the dualistic subject-object structure characteristic of mature perception—the first-personal “in here” of consciousness presenting things and events in the world as “out there”—has yet to firmly take hold.⁴² To be clear, I am not suggesting that a first-personal aspect is missing entirely within early experiences (I return to this point below). Again, as Merleau-Ponty notes, infant’s capacity for visual imitation seems to suggest that they possess at least a rudimentary sense of embodied ecological or situated selfhood, part of which, as Ulric Neisser observes, is “awareness of one’s situation in an independent, spatially extended environment.”⁴³ Nevertheless, modulation of attention (i.e., narrowing the perceptual field to focus in on one thing while excluding others) seems to be a central mechanism for establishing the sense of being a stable perspective on the world. Since young infants lack this attentional dexterity, they likely lack the feeling of stable perspectival selfhood that flows from it.

It is considerations such as these that lead Alison Gopnik to say that, in a sense, babies are more conscious than adults; for them, consciousness is an encompassing lantern instead of a focal spotlight (a la adult consciousness).⁴⁴ Accordingly, I suggest that the experiential character of early infant consciousness is structurally poised to be shared in an intimate manner. As radically open (exogenously determined), phenomenal states at this stage of our ontogeny are potentially capable of being entered into, shaped by, and shared with others. Highlighting the radical openness of early experience is a first step in supporting Merleau-Ponty’s JT. An example of precisely how this occurs will be the focus of the next section.

3.2 Regulating emotions

By considering the way that young infants allow others to regulate their attention, we are now in a position to see how external regulation makes them ready to take on and share the emotions of others. Speculating about the character of infant phenomenology is, of course, difficult. As Merleau-Ponty notes, we must be careful not to assume that early experience has exactly the same character and structure as it does in adulthood.⁴⁵

Nevertheless, we can begin by noting that, in virtue of their lack of endogenous attention, infants exhibit what we might term attenuated “emotional agency.” This expression refers to the degree of volitional control we normally have in modulating the phenomenology of various emotional experiences. For example, as mature perceivers, emotions don’t merely happen to us. To a certain extent, rather, they are things that we do. When in the grip of an angry episode, say, I can regulate the felt character of this episode by willfully assuming a particular posture or stance (clenching my fists and leaning forward aggressively) intended to intensify and sustain the episode; and I can further intensify my anger by focusing on specific anger-inducing aspects of the situation. Conversely, I can suppress my angry experience by willfully adopting a more placid posture, breathing slowly, and directing my attention to pleasant things.⁴⁶

This is not to deny that many aspects of emotional experience are automatic and involuntary. Rather, the point is that emotions harbor both involuntary (neurophysiological) and voluntary (attentional) aspects.⁴⁷ Felt awareness of their voluntary aspects—using attention to modulate the character of emotional experience—is one aspect of “emotional agency.” Thus, not only can I as a mature perceiver inhibit my attention. Additionally, I can inhibit (or elevate) my felt responses to things happening around me. Exertion of emotional agency is, I suggest, a form of self-world differentiation, an ontogenetically significant form of “boundary maintenance.”⁴⁸ Phenomenologically speaking, attentional and affective regulation goes hand in hand.

Very young infants lack emotional agency.⁴⁹ As predominantly exogenous attenders, they are, as we’ve seen, largely subject to the whims of their environment. Within an interpersonal context, this means that they tend to experience whatever their caregivers experience. Consider mutual affect regulation.⁵⁰ This phenomenon refers to the way that infants and caregivers together establish a synchrony of emotion and expression within sequences of interaction. Both partners perceptually pick up on and respond to the emotions of others by developing similar emotions. This sort of responsiveness is the bedrock of social engagement. However, it involves more than a simple isomorphic matching of feeling. Rather, via gesture, facial expression, touch, speech, and spatial proximity, caregivers regulate the character of these interactions in a way vital for the infant’s socio-affective development. They enable the infant to feel and experience in ways that would otherwise be beyond their abilities.

From birth, infants are socially sensitive; as we’ve seen, they respond differentially to the expressive “packages” of auditory-visual-tactile information that caregivers send their way.⁵¹ But it is important to tread cautiously. In stressing their innate or “primary” intersubjectivity,⁵² we must not overemphasize the infant’s social autonomy. For, it is caregivers who organize interpersonal contexts in specific ways, scaffolding infants within particular bodily and gestural “play frames characterized by exaggerated contours, marked changes of tempo, and systematic repetitions.”⁵³ Caregiver responses to infants call forth the next behavior and emotion within an interactive sequence.⁵⁴ Thus, while the infant enjoys some degree of agency—again, they are not completely inert, socially speaking—their agency is mediated by the physical interventions of the caregiver. The latter provides the bodily dynamics that shape the interactive context—what Merleau-Ponty refers to as the sphere of “intercorporeity”—in which these exchanges are organized and play out.

By regulating gaze behavior, caregivers regulate perceptual input and thus regulate infant arousal and affect.⁵⁵ This is clear when we look at positive affect in infants. As any parent knows, infants can from birth readily express negative affect. Even newborns leave little doubt when they’re distressed. But the crucial point for our purposes is that the experience and expression of positive affect in infants “require[s] the participation of an attuned adult who can both construct and coregulate the positive affect in a moment-by-moment process.”⁵⁶ Simply put, the origin of positive affect in early infant experience is inherently dyadic. There is, therefore, a class of emotional experiences young infants simply cannot access without the help of caregivers. The latter awaken these experiences in infants by directly engaging with them and continually optimizing the stimulus value of their auditory-visual-tactile packages, crafted to keep the infant in an “optimal zone for play” between over-stimulation and under-arousal.⁵⁷

For example, consider the following free play exchange between a mother and a 4- to 5-month-old.⁵⁸ Initially, both are relatively expressionless:

• The mother then opens with a low-keyed behavioral constellation of simultaneous behaviors: looking at baby; head leans toward him; faint smile; mouth a bit open; eyebrow up a bit; soft vocalization.

• The baby responds with a slight and transient smile, mild eye widening, and eyebrows lift. His arms start to move about in irregular circles.

• The mother then leans closer to the baby, progressively exaggerating the facial display already present (i.e., her expression grows toward one of “mock surprise”; the baby regards her unflinchingly, his eyes open more and then close some several times over as does his mouth, as he watches the intensity of her facial display grow.

• Finally, the baby breaks into a moderately intense mouth smile accompanied by an elongated vocalization, his eyebrows fly up, his head tilts back, his arms flap, and his hands open and close. This constellation last about 2 s.

• 250 ms after the above baby behavior starts, the mother bursts into a smile, imitating (unaware) that the baby’s facial expressions and vocalization—such that for the last 1 ¾ s of the baby’s behavior the mother’s behavior is being contoured to match his

• And so on.

Within this exchange, the mother utilizes various physical strategies (sometimes unknowingly) to elevate the infant into a state of elevated affect. Without these interventions, the infant would not reach this state on his own.

The still face paradigm provides a vivid experimental example of the essential role caregivers play in constructing and regulating positive affect.⁵⁹ After establishing a positive interaction, the caregiver (usually the mother) will suddenly look away and, a few moments later, look back at the infant with an expressionless “still” face. The effects on the infant’s mood are immediate and striking. They immediately detect a change in the interaction, become highly distressed, and their behavior becomes increasingly disorganized. However, when the mother breaks the still face—and with a bit of gentle coaxing—equilibrium is soon restored and the infant is once more elevated into a state of positive affect.

For the first weeks and months of life the experience of emotional agency—as well as positive affect—lie on the caregiver side of the infant-caregiver dyad. Experientially, infants feel themselves pulled into the dynamic character of these dyadic exchanges. Part of the phenomenology of these early exchanges is thus to experience them as controlled and regulated via an external agency. The infant plays some role in shaping their affective contour, of course; the mother adapts and refines her responses to the infant’s ongoing expressions. But what I suggest is that in virtue of the caregiver’s disproportionate influence on this process—coupled with a dramatically open or world-directed exogenous attention—infants lack a strong phenomenological sense of emotional selfhood and social differentiation. This dramatic openness is what allows caregivers to construct and regulate positive affect in the infants. For an important part of their affective development, infants require interpersonal evocation and regulation in order to register certain experiences.⁶⁰

In light of these considerations—and to return to Merleau-Ponty’s JT—it is sensible to speak of caregivers as entering into the infant’s experience as a joint subject. By coordinating her responses with the infant and elevating the infant’s arousal and affect, the caregiver’s positive affect expands to include the infant.⁶¹ Both are part of the shared emotion: the caregiver, because it originated with her; the infant, because she lacks the inhibitory resources to keep the caregiver out of her experience. This “relation of “reciprocity”” is thus the vehicle for the emergence of this shared emotion.⁶²

The emotion of the caregiver is thus a constituent part of the infant’s emotional experience. Crucially, the infant’s emotional experience stands in a different sort of dependency relation to the caregiver than, say, in cases of emotional contagion (e.g., where one person’s emotional experience triggers an affectively similar but numerically distinct emotion in another person or persons). This is because the caregiver’s emotion is part of the very process responsible for the realization of the emotion within the infant’s experience—along with, of course, things like the infant’s properly-functioning brain and central nervous system. So, it cannot be bracketed out of taxonomic characterizations of the infant’s experience. Of course, the infant is already conscious. The caregiver does not magically bring the infant to awareness by emotionally engaging with it. But without the immediate presence of the caregiver’s emotion within the interaction—along with the external regulatory resources the caregiver provides—the infant cannot register positive affect. Once more, this dependence relation results from the infant’s dramatic exogenous attention and lack of emotional agency (affective self-regulation), meaning that the infant’s consciousness is structurally poised to accept the caregiver’s positive affect as its own, and to incorporate it into the process of realizing that experience for itself. Yet the emotion simultaneously remains a constituent of the caregiver’s experience as well. In this sense is it jointly-owned.

4 An objection

Earlier I suggested that Merleau-Ponty appears to endorse (at least) two fundamental claims in “The Child’s Relations with Others” and his defense of JT. First, he argues that the perceptual world of the infant has some degree of coherence and structure. This structural coherence is rooted in a sense of the body’s location in space: a situated or ecological sense of being an embodied perspective onto a perceivable environment. Second, despite this rudimentary structure, the newborn is experientially coupled to the caregiver in a state of phenomenal undifferentiation.

It may be objected that these two claims are inconsistent—or at least that they stand in tension with one another.⁶³ This tension can be made explicit in the following way. If the child is born with a primitive sense of embodied selfhood—the imitation research, recall, suggests that the infant can feel coherent proprioceptive possibilities for imitating from within a body (her own) that she’s never seen—it would appear that the infant already possesses a minimal sense of separation from her environment. In other words, the body-schema—a coherent system of sensory-motor functions that govern movement and maintenance of posture—exists prenatally and is thus sufficiently present at birth to give the infant an articulate sense of being an individuated bodily self.⁶⁴ For, in order for the infant to imitate, she must be aware that (1) she has or is a body, and that (2) she can mobilize her body to do certain things in response to external events in the environment. In addition to being aware of her surroundings, therefore, the newborn is aware of herself as a separate physical agent.⁶⁵ If so, this would appear to put pressure on the idea that infant is ever coupled to the caregiver in a state of a subjective undifferentiation. Rather, from birth, the infant possesses a minimal sense of situated bodily subjectivity which includes a sense of independence from the world and others.

With a minor qualification, these two claims can be reconciled. First, as I’ve already argued, it does appear likely that since inner control of attention is lacking in infants, so, too, is a primary mechanism for establishing the sense of being a stable, unified, and enduring subject. Thus, we ought not to assume that whatever structure the infant’s experience has is identical to that of a mature perceiver. Nevertheless, we can grant that the infant’s primitive sense of being a bodily self establishes both a certain unity and structure to its experiences as well as a minimal sense of independence from its environment. So, if Merleau-Ponty is arguing that all of the infant’s experiences are undifferentiated (i.e., she has no sense of herself as a distinct subject whatsoever), various developmental studies appear to falsify this idea.⁶⁶

However, that the infant possesses a primitive sense of bodily selfhood is not inconsistent with its entering into certain experiences (i.e., episodes of positive affect) as a joint subject. Again, as we’ve seen, the infant requires the intervention of the caregiver to realize certain capacities and experiences, including endogenous control and positive affect. Within these scaffolded experiences, it appears that a new category of experience opens up for the infant. Daniel Stern helpfully suggests that, in characterizing these experiences, a third category is needed: the “self-with-a-self-regulating other.”⁶⁷ This option charts a middle way between two contrasting ideas: that the child is born into a state of total phenomenal undifferentiation; and that the child is born with a robust, enduring sense of being a self independent from its environment. As Stern puts it, this middle way instead describes a

category of self-experience in infancy in which there is no confusion between the physical self and other, but where certain self-functions (e.g., a high level of joy or felt security) depend upon the other’s invariant presence and interaction for their existence. In mutual regulation of joy, other affects, attachment, love, and many meaning system and beliefs, this entity of a self-regulatory-other becomes a large part of the interactive experience.⁶⁸

The exogenous scaffolding provided by the caregiver is what opens up this qualitatively new form of experience.

So, with Stern’s idea in place, we can now weaken one of Merleau-Ponty’s claims and in so doing reconcile it with the other one. First, it is unlikely that the infant is born into a state in which it experiences itself as completely undifferentiated from its environment. This claim needs to be weakened. For, as we’ve seen, much developmental evidence suggests the contrary; it is likely that newborns come into the world with a minimal sense of physical agency and self-world separation. Nevertheless, because they lack inner control of attention—and thus a central mechanism for establishing the enduring sense of being a conscious subject—it is quite possible that the boundaries of this sense of separateness are constantly in flux, shifting from context to context. In cases of visual perception (i.e., imitating faces), for example, this sense of separation may be more prominent. In extended cases of affect regulation, however, it may recede from awareness. This plasticity and malleability of self-consciousness is what, at this state of our ontogeny, renders some experiences “open” such that others can enter into them as a joint subject. Thus, contra Merleau-Ponty, it may only be in cases of positive affect (or similar high-arousal emotions) that the infant is coupled to the caregiver in a state of (emotional) undifferentiation. Most of the time, however, they experience themselves as independent embodied subjects.

But let’s make a further concession and grant that infants always possess a minimal sense of bodily self-world differentiation—including during episodes of affect regulation. Even now, I suggest, they may still be said to share emotions with caregivers in the deep sense JT entails. How so? This is because the physical and emotional interventions of the caregivers—the exogenous scaffolding they provide—is part of the apparatus the infant needs to realize a certain class of experiences. Within these interactions, the infant and caregiver create a coupled social system; the caregiver provides socio-cognitive resources (positive affect) and capacities (endogenous attention) that the infant incorporates into her own basic suite of sensorimotor skills to (jointly) realize the experience.

This incorporation is a functional integration. The infant’s native capacities for perceiving and feeling integrate with the external capacities of the caregiver. Crucially, only when these native capacities hook up with the caregiver’s external capacities does the infant realize certain processes (i.e., positive affect and heightened attentional control), much the way that the process of air conditioning is only realized when various spatially distributed parts of an air conditioning system (thermostat with electrical connections to a breaker box, a refrigerant, expansion valve, evaporator coil, compressor, fan, etc.) integrate their respective functions to cool the air. An evaporator coil cannot on its own realize the process of air conditioning. But when integrated with other appropriate (i.e., complementarity) components, it can. It becomes part of a more complex system, and new and more sophisticated form of functionality emerges. Likewise, the infant depends upon regulatory input from the caregiver that, when integrated with its own native sensorimotor capacities, expands the complexity and coherence of the infant’s experience.⁶⁹ Part of this expanded experience is the caregiver’s positive affect, which the infant—again, via regulatory input from the caregiver—takes on as her own. It becomes a constituent of her “self-with-a-self-regulating-other” experience. This new experience need not negate the infant’s sense of being a separate physical agent. But this experience does require the ongoing activity of the infant-caregiver system functioning as a whole.⁷⁰ Insofar as the infant’s newly-expanded experience only emerges within the confines of this system, infant and caregiver might thus be said to share the same episode of emotion. JT can thus be reconciled with the infant’s bodily sense of self-world differentiation.

5 Empirical consequences of adopting JT

I now briefly consider an empirical consequence of adopting JT. I suggest that JT helps specify a primary mechanism for the development of early social cognition: the dyadic emotional intimacy characteristic of instances of early interactions. This mechanism tends to be overlooked, or at least underemphasized, in many current approaches to social cognition.

If the philosophical problem of other minds is first and foremost an epistemological question⁷¹—How am I justified in believing that others are likewise minded?—there is nevertheless another important dimension to this problem. We can term this the “mechanisms” question. This question is concerned with specifying the mechanisms enabling us to attribute mental states to others—beliefs, desires, intentions, emotions, etc.—whether or not we’re epistemically justified in doing so (which remains an independent concern). A focus of ongoing work on social cognition in philosophy of mind and cognitive science is to identify the precise mechanisms responsible for this capacity.

Until recently, two approaches were dominant: Theory theory (TT) and simulation theory (ST). Both approaches accept the premise that we cannot perceive the mental states of others. Both therefore posit distinct extra-perceptual cognitive mechanisms that enable us to detect and respond to another’s mental states. According to TT, social cognitive capacities—our ability to “mindread”⁷²—depend upon the early development and refinement of a quasi-scientific theory of mind.⁷³ ST, on the other hand, posits that we use our own mental states as a model to imaginatively simulate what another person is thinking and feeling when we observe them behaving a certain way.⁷⁴ This simulation process may be personal-level and explicit, or sub-personal and implicit (invoking activation of the mirror neuron system).⁷⁵

The TT and ST paradigms have recently received criticism from multiple fronts. For example, a number of phenomenologically motivated theorists—drawing upon Merleau-Ponty, as well as Husserl, Scheler, and Schutz, among others—have argued that, prior to or independently of theorizing and/or simulating, we directly perceive another’s thoughts, intentions, emotions, etc., within their behavior, which is overtly expressive of their mental life.⁷⁶ Another line of criticism emerges from those who argue that the reciprocal dynamics of the interactive process itself are sufficient on their own to give us relatively rich social understanding, even from an early age.⁷⁷ The extent to which these alternative models are truly incompatible with either TT or ST or even with each other remains an open question.

What is salient to present concerns is that JT suggests a central social-cognitive mechanism that tends to be overlooked, or at least underemphasized, in these approaches (including those explicitly motivated by phenomenology): specifically, the affective intimacy of early dyadic exchanges, and the jointly-owned emotions that characterize them. To head off an immediate objection: none of the approaches mentioned above are strictly speaking incompatible with the claim that affective intimacy is a central mechanism responsible for basic social cognitive capacities. Despite a tendency of some critics to paint them in simplistic caricatures suggesting otherwise, both TT and ST stress the importance of social interaction—including engaging with our own emotions and those of others—in providing the “data” for our other-directed theories or simulations, respectively.⁷⁸ Likewise, both perception-based and interactive accounts of social cognition concede a place for affect attunement. Perception and interaction are processes that enable us to detect another’s emotional expressions and calibrate our own emotional responses accordingly. So, none of these approaches would deny a role for some degree of affective intimacy.

In stressing the primacy of affective intimacy and jointly-owned phenomenal states, what, then, is genuinely unique about taking Merleau-Ponty’s JT seriously? First, Merleau-Ponty explicitly inverts the guiding question behind both TT and ST—Which mechanisms enable us to overcome the original self-other gap?—and asks instead, How does subjectivity arise out of elemental intersubjectivity? Merleau-Ponty’s JT thus urges, contra TT and ST, that subjectivity is organically and ontologically intersubjective. Self and other arise from the same elemental “flesh of the world” to use a provocative phrase from his later work, interwoven via a complex dynamic of separation (écart) and intimacy-making “reversibility.”⁷⁹

Second, the explanatory significance of this inversion is that the intersubjective matrix from which subjectivity arises (i.e., the common flesh of the world of which we are all “counterparts” or “cohabitants”)⁸⁰ becomes the target explanandum. It is within this context that the infant becomes part of a dyadic system more complex than either the infant’s or the caregiver’s endogenous state(s), and thus first develops the self-regulatory capacities that enable it to negotiate this system—capacities that are simultaneously a crucial part of the individualizing process.⁸¹

While both TT and ST concede a causal-explanatory role for the social world, neither insists that it is necessary for taxonomic considerations of basic social cognitive capacities and individualizing processes. Yet via this shift of emphasis, the affective intimacy made possible via jointly-owned experiences is seen to be the developmental ground of various individualistic capacities such as self-consciousness, language, self-regulation, and representation (including theorizing and simulating).⁸² Accordingly, this move suggests the need to further refine our concepts and models of intersubjectivity, moving away from mentalistic approaches emphasizing inside-out mental state attribution to more dynamic models emphasizing the affective primacy of relational equilibrium—that is, a reciprocal negotiation of the elemental bodily dynamics that link us to others.⁸³ For this model, intersubjectivity is not characterized as two hidden minds looking to escape their intracranial confines to make contact. Rather, as Merleau-Ponty notes, “[w]e have here a dual being [un être à deux], where the other is no longer a mere bit of behavior in my transcendental field, nor I in his; we are collaborators in a consummate reciprocity. Our perspectives merge into each other, and we co-exist through a common world.”⁸⁴ While critics of both TT and ST have offered a similar call, PT emphasizes that this move is not simply theoretically desirable but explanatorily necessary.

Some recent approaches have made a move in this direction. For example, interactionist approaches tend to focus on the formal dynamics of the interaction process itself, appealing to concepts like “autonomy,” “sense-making,” “co-regulation,” etc., to highlight systems-level properties that purportedly given social encounters their coherence and significance.⁸⁵ But the emotional character of our encounters isn’t granted nearly the same emphasis. And while the shift of emphasis called for here is, once more, consistent with both the perception-based and interactionist approach to social cognition, PT again departs from both in explicitly foregrounding the importance of shared affect—in arguing, that is, that our perceptual and interactive skills are only refined against a pre-existing backdrop of deep affective sharing in the sense here articulated. More specifically, Merleau-Ponty’s JT urges that the affective quality of early interactions—their warmth, reciprocity, coordination, etc.—specifically influences the ontogenesis of fundamental socio-cognitive skills such as proximity seeking, social initiating, gaze monitoring, imitative learning, symbolic functioning, and even the attribution of intentionality to other’s action.⁸⁶

6 Final thoughts

To conclude, I have summoned different strands of work in developmental psychology to render plausible Merleau-Ponty’s claim that some early experiences are jointly owned (i.e., what I termed the “joint ownership thesis”). In virtue of the infant’s lack of attentional control and ability to self-regulate their emotions, certain phenomenal states—such as experiences of positive affect—are, for the infant, structurally poised to be shared with others in the deep sense that JT entails. Extended periods of face-to-face interaction within infant-caregiver dyads are the context in which this sharing emerges. And while this view need not supplant current models of social cognition and interpersonal understanding, it does suggest that the intimacy of shared affect within early exchanges may play a more important ontogenetic role than is often acknowledged. Additionally, it provides a vivid example of how phenomenological approaches to intersubjectivity continue to offer important resources for furthering debates within the cognitive sciences.

Many open questions remain. One is whether the experience of jointly owned emotions continues into adulthood. Since adults can regulate both attention and emotion on their own (i.e., endogenously), one might conclude that their experiences lack the structural openness of early infant (externally-regulated) experiences—and thus others cannot participate in their experiences in the manner possible with young infants. But things are not quite that simple. It may well be that, as with young infants, there are certain emotions we simply cannot generate on our own without the regulatory input provided by certain environmental structures—including the presence of other people.

Consider the following experiences: attending a powerful musical performance; enthusiastically cheering for the home team at a sporting event; listening to a skilled orator give an impassioned speech; being swept up into a mystical fervor at a religious ceremony; being part of a highly-charged political event or protest rally. Within these cases (and many more like them), the environmental structures and bodily presence of others supply input that allows us to experience intense emotions impossible in their absence. In other words, these cases may be similar to the infant-caregiver dyadic experiences considered previously: examples of “phenomenal coupling” in which we are drawn into group-wide emotions generated only in the presence of the relevant external dynamics.⁸⁷ It might therefore be sensible to speak of the emergence of a single group-wide experience of elation at a sporting event, say, or collective grief at a funeral. Remove the context and external structures, and the associated emotion similarly dissolves.

Merleau-Ponty would likely accept this idea. It’s not at all clear to me that we should reject it. The ability to experience forms of uniquely shared emotions is seductive, and further emphasizes the crucial role that others play in shaping our subjectivity. But a further consideration of this idea will have to wait for another occasion.