The Moral Life of Babies (Puppet Studies) – Paul Bloom and Karen Wynn

For centuries, the moral character of the human infant was largely relegated to the realm of speculative philosophy. Thinkers across the classical and Enlightenment traditions debated whether human beings enter the world trailing clouds of innate virtue, as romanticized by Jean-Jacques Rousseau, or arrive fundamentally corrupted by primal, self-interested appetites that demand state subjugation, as asserted by Thomas Hobbes. In the twentieth century, developmental psychology appeared to settle this debate in favor of a modified tabula rasa: infants were conceptualized as premoral, sensorimotor entities whose ethical faculties could only assemble slowly through linguistic socialization, cognitive maturation, and the internalization of cultural taboos. Morality, within this constructivist and behaviorist paradigm, was an achievement of later childhood—a complex intellectual superstructure erected over an amoral biological foundation.

This long-standing consensus was radically disrupted in the early twenty-first century by research emerging from the Yale Infant Cognition Center, affectionately known as the “Baby Lab.” Under the leadership of developmental psychologists Paul Bloom and Karen Wynn, alongside foundational collaborators such as J. Kiley Hamlin, researchers designed ingenious non-verbal experimental paradigms to probe the sociomoral evaluations of preverbal infants. Utilizing dynamically choreographed puppet shows featuring geometric wooden characters and whimsical puppets, Wynn, Bloom, and Hamlin posed a deceptively simple question: Can babies, prior to acquiring language or undergoing formal parental indoctrination, discern the difference between benevolent and malevolent social behavior, and do they exhibit an innate preference for the good?

The empirical findings emanating from these studies have revolutionized cognitive science, evolutionary psychology, and moral philosophy. Preverbal infants as young as three to six months old do not merely perceive physical collisions and spatial trajectories; they parse the social world into valenced intentions, displaying robust preferences for prosocial “Helpers” over antisocial “Hinderers,” expecting equitable resource distribution, and endorsing retributive justice against bad actors. However, this innate moral foundation is not an unalloyed engine of universal benevolence. The same cognitive apparatus that enables infants to applaud altruism also harbors the evolutionary roots of parochial altruism, in-group favoritism, and tribal suspicion. This article presents an exhaustive investigation of Bloom and Wynn’s puppet studies, dissecting their historical background, experimental paradigms, philosophical implications, replications, and ongoing debates surrounding the origins of the moral mind.

1. Introduction to Infant Moral Cognition and the Yale Infant Cognition Center

1.1 The Epistemological Shift in Developmental Psychology

The emergence of infant moral cognition research represents one of the most profound epistemological shifts in developmental science. Throughout much of the twentieth century, developmental psychology was anchored by the constructivist frameworks of Jean Piaget and the stage-based moral developmental theories of Lawrence Kohlberg. These dominant paradigms asserted that morality was fundamentally an explicit, reflective, and linguistic enterprise. Piaget posited that young children dwell in a state of heteronomous morality characterized by moral realism, wherein rules are regarded as immutable decrees handed down by adult authorities, and actions are evaluated strictly by their material consequences rather than the underlying intentions of the actor. Kohlberg extended this paradigm into an elaborate six-stage developmental ladder, placing young children in the preconventional stage where ethical decisions are dictated entirely by the avoidance of physical punishment or the crude pursuit of self-interest.

Under these traditional architectures, the preverbal infant was viewed as cognitively incapable of sociomoral evaluation. Lacking language, formal operational logic, and socialization, infants were classified as amoral organisms driven strictly by physiological homeostasis and primitive perceptual reflexes. However, the closing decades of the twentieth century witnessed the rise of evolutionary psychology and cognitive nativism. Pioneered by scholars such as Elizabeth Spelke and Susan Carey, researchers discovered that preverbal infants possess rich “core knowledge” systems—specialized, innate mental modules designed to parse physical mechanics, animate agency, and numerical relationships long before any formal education takes place.

The conceptual leap executed by the Yale Infant Cognition Center was the realization that core knowledge might extend beyond physics and mathematics into the sociomoral domain. Rather than treating morality as a late-emerging cultural veneer, researchers began to conceptualize the moral sense as an evolutionary adaptation. In this view, our human ancestors could not survive without the capacity to rapidly identify cooperative partners, punish free-riders, and navigate complex coalitionary dynamics. Thus, the moral mind was re-theorized not as an esoteric philosophical construct acquired through schoolbooks and sermonizing, but as an ancient biological guidance system hardwired into human ontogeny to ensure social survival.

1.2 Paul Bloom and Karen Wynn: Collaborative Foundations

The intellectual synergy that catalyzed this empirical revolution emerged from the complementary research backgrounds of Paul Bloom and Karen Wynn. Karen Wynn had already established an international reputation for her groundbreaking work on infant core computational systems. In her landmark 1992 study published in Nature, Wynn demonstrated that five-month-old infants possess an innate capacity for basic mathematical operations, calculating the addition and subtraction of small quantities of physical objects hidden behind an occluder. This discovery shattered the dogma that numerical abstraction depends entirely on linguistic symbolic systems. Wynn established a sophisticated methodological framework showing that preverbal infants use looking time and expectancy-violation metrics to reveal surprisingly complex abstract representations of their environment.

Paul Bloom, concurrently, was examining the foundations of human thought through the dual lenses of cognitive psychology, linguistics, and philosophy of mind. His work explored how children acquire language, how they construct theory of mind—the ability to attribute mental states, desires, and beliefs to other agents—and how emotional adaptations such as empathy and disgust shape adult ethical frameworks. Bloom recognized that the missing link in developmental moral theory was an empirical investigation of moral origins. If infants could compute mathematical differences without language, could they not also compute the valence of social transactions? Together with their then-doctoral student J. Kiley Hamlin, Bloom and Wynn unified these paradigms to inaugurate the systematic investigation of preverbal moral cognition.

Their central hypothesis departed fundamentally from the classical blank-slate models. Bloom and Wynn posited that humans enter the world equipped with an evolved “moral sense”—a foundational, non-linguistic capacity to judge actions as good or bad, to distinguish helpful agents from obstructive ones, and to mete out rudimentary retribution. While acknowledging that adult moral reasoning is heavily scaffolded by culture, explicit philosophy, and institutional laws, they asserted that this sophisticated superstructure rests on a foundation of innate, biologically prepared sociomoral instincts.

1.3 Deconstructing the Preverbal Research Dilemma

Investigating the internal cognitive states of human subjects who cannot speak, follow verbal commands, or write questionnaires presents an imposing methodological challenge. This preverbal research dilemma has historically forced developmentalists to walk a tightrope between over-interpreting simple motor reflexes and underestimating profound cognitive capacities. In traditional animal and developmental psychology, behavioral outcomes were frequently attributed to simple operant conditioning: an organism behaves in a particular manner because it has been reinforced with a physical reward such as food, milk, or maternal warmth. Demonstrating genuine sociomoral assessment required designing experimental procedures that were entirely free from operant reinforcement, training trials, or overt linguistic cuing.

The central dilemma lay in constructing an experimental theater that was perceptually transparent yet socially meaningful. If the stimuli were visually overly intricate, an infant’s unrefined visual cortex and limited working memory would succumb to cognitive overload, yielding uninterpretable, noisy gaze patterns. Conversely, if the stimuli were stripped down to sterile geometric movements devoid of intentional cues, the infant would perceive them merely as mechanical physics problems rather than social dilemmas. The Yale team resolved this tension by adopting the medium of puppet theater, employing stylized, anthropomorphic physical agents with large, expressive googly eyes performing goal-directed, teleological actions.

To access the infant’s underlying cognitive evaluations, researchers had to rely on indirect behavioral proxies: the direction and duration of an infant’s visual gaze, and the spontaneous physical reaching of their hands toward presented objects. By linking these subtle, non-linguistic behaviors to rigorous experimental designs, Bloom, Wynn, and Hamlin unlocked a methodological portal into the infant mind. Their innovative paradigms allowed researchers to isolate the precise variables that trigger moral approval or disapproval long before the child possesses the vocabulary to articulate words like “kindness,” “meanness,” “fairness,” or “justice.”

2. Historical Frameworks of Moral Development: From Blank Slate to Innate Capacity

2.1 The Tabula Rasa Tradition and Radical Behaviorism

To fully grasp the disruptive nature of Bloom and Wynn’s work, one must contextualize it within the philosophical and psychological frameworks that preceded it. The intellectual heritage of Western child psychology was profoundly shaped by seventeenth-century philosopher John Locke. In his monumental work, An Essay Concerning Human Understanding (1689), Locke vigorously rejected Cartesian notions of innate ideas, proposing instead that the human mind is at birth an “unscribed tablet”—a tabula rasa. Locke argued that all knowledge, moral values, and understanding derive exclusively from sensory experience, reflection, and empirical association. Within Locke’s formulation, children are born neither inherently moral nor intentionally immoral; they are impressionable raw material to be molded through parental education, discipline, and habituation.

In the twentieth century, Locke’s empirical doctrine was radicalized by the advent of behavioral psychology, championing figures such as John B. Watson and B.F. Skinner. Radical behaviorism systematically banished internal mental states, desires, representations, and innate concepts from scientific discourse. Watson famously proclaimed that given a dozen healthy infants, he could take any one at random and train them to become any type of specialist—doctor, lawyer, artist, or thief—regardless of their ancestral talents or biological predispositions. Skinner extended this mechanistic logic directly into the moral sphere, asserting that so-called moral behavior is simply behavior that has been positively reinforced by society, while immoral behavior is that which has been extinguished via aversive conditioning or punishment.

Under the behavioral lens, an infant is simply an organism governed by primary homeostatic drives: hunger, thirst, thermal comfort, and pain avoidance. Prosocial impulses, such as sharing a toy or comforting a crying peer, were conceptualized as learned behaviors acquired via secondary conditioning. The infant learns that smiling at a parent or returning an object yields smiles, verbal praise, and physical affection. The idea that a six-month-old infant could harbor an endogenous preference for altruism, completely devoid of any direct personal reinforcement, was deemed an unscientific impossibility by the behaviorist orthodoxy.

2.2 Constructivist Models: Piaget and Kohlberg

While Jean Piaget broke with the strict environmentalism of the behaviorists by granting children an active role in constructing their own knowledge, his constructivist paradigm remained deeply skeptical of early moral capabilities. Piaget’s seminal 1932 work, The Moral Judgment of the Child, argued that morality emerges through a sequence of qualitative cognitive stages, heavily mediated by peer play and cognitive maturation. For Piaget, children under the age of six or seven reside in a “premoral” or “heteronomous” stage. In this phase, young children lack the capacity to take the perspective of another person—a cognitive limitation he termed egocentrism.

Piaget famously demonstrated this phenomenon using his clinical interview method, presenting children with contrasting moral vignettes. In one classic scenario, a child accidentally breaks fifteen expensive cups while opening a door, unaware that the cups were behind it. In another scenario, a child breaks a single cup while attempting to sneak jam from a cupboard against their mother’s strict orders. Piaget observed that young children overwhelmingly judged the child who broke fifteen cups as “naughtier” and deserving of harsher punishment than the child who broke one, despite the second child’s malicious intent. From this, Piaget concluded that young children are “moral realists” who fixate entirely on the material, objective consequences of an act while remaining utterly blind to subjective intentionality.

Lawrence Kohlberg subsequently expanded Piaget’s model into a lifelong developmental taxonomy, structured across three levels and six distinct stages of moral reasoning:

• Level I: Preconventional Morality (Common in young children)
  • Stage 1: Obedience and Punishment Orientation (Morality is defined strictly by the direct physical consequences of an action; an act is wrong simply because it leads to punishment).
  • Stage 2: Individualism and Exchange (Right behavior is defined instrumentally as whatever satisfies one’s own immediate, hedonistic desires and reciprocal needs).
• Level II: Conventional Morality (Common in adolescents and adults)
  • Stage 3: Interpersonal Relationships (Good behavior is driven by the desire for social approval and living up to interpersonal expectations).
  • Stage 4: Maintaining Social Order (Morality is defined by upholding societal laws, duty, and respect for established social authority).
• Level III: Postconventional Morality (Attained by a minority of adults)
  • Stage 5: Social Contract and Individual Rights (Laws are understood as flexible instruments designed to protect human rights and collective welfare).
  • Stage 6: Universal Ethical Principles (Morality is guided by internalized, self-chosen abstract principles of justice, human dignity, and equality that transcend statutory laws).

The fatal flaw shared by both Piagetian and Kohlbergian frameworks, as Bloom and Wynn pointed out, was their heavy reliance on verbal clinical interviews. By requiring children to listen to complex linguistic narratives and formulate articulate justifications for their judgments, these theorists conflated moral reasoning with moral competence. Just as a child may fully grasp the deep grammar of their native language years before they can consciously recite the formal rules of syntax, an infant may possess intuitive moral evaluations long before they can deliver an analytical defense of their ethical choices.

2.3 Nativist and Evolutionary Moral Psychology

The conceptual alternative to constructivism emerged through the lens of evolutionary psychology and cognitive nativism. In evolutionary terms, the human species is characterized by an extreme degree of obligate sociality and cooperative breeding. Over evolutionary timescales, hominin survival depended entirely on the internal cohesion of the hunter-gatherer band. In his seminal 1971 paper, Robert Trivers formulated the theory of reciprocal altruism, demonstrating that cooperation among non-kin can evolve if individuals possess the cognitive architecture to reward altruists, punish cheaters, and avoid interactions with free-riders.

If social survival hinges on avoiding cooperative defection, relying entirely on trial-and-error experiential learning would be biologically disastrous. A child who had to spend years learning from scratch that deceptive, aggressive, or uncooperative individuals are hazardous would likely not survive to pass on their genes. Consequently, evolutionary psychologists hypothesized that natural selection must have favored the emergence of an innate, biologically prepared sociomoral evaluation module. Just as humans do not need to be taught to experience fear when encountering an abrupt drop-off or a venomous predator, they should possess an innate affective attunement to prosocial and antisocial cues.

This perspective aligns with the concept of Universal Grammar pioneered by Noam Chomsky, later extended to the ethical domain by philosophers and cognitive scientists such as John Mikhail and Marc Hauser under the rubric of “Universal Moral Grammar.” This hypothesis posits that just as human infants possess an innate computational toolset for parsing syntactic structures from raw acoustic data, they also possess a specialized cognitive engine for parsing actions into categories of right and wrong, permissible and impermissible. Bloom and Wynn adopted this nativist foundation, proposing that human infants possess an innate “moral core”—a foundational bedrock of sociomoral evaluations that operate independently of cultural socialization, providing the raw material upon which mature moral culture is subsequently constructed.

3. Methodological Architecture: How Experimental Puppet Shows Test Preverbal Minds

3.1 Violation-of-Expectation and Habituation Paradigms

To circumvent the linguistic constraints of infancy, researchers at the Yale Baby Lab harnessed one of the most reliable and validated methodologies in developmental science: the violation-of-expectation (VoE) paradigm, coupled with infant habituation dynamics. This methodology capitalizes on a fundamental quirk of infant visual attention: infants, like adults, quickly become bored when exposed to repetitive, familiar events—a psychological process known as habituation. Once an infant habituates to a stimulus, their looking time drops significantly. However, if the experimenter introduces a novel event that contradicts the infant’s expectations or mental models of the world, the infant experiences cognitive dissonance, causing their visual attention to spike dramatically. This resurgence in looking time is known as dishabituation.

Historically, this paradigm had been applied to prove that infants possess an understanding of basic physics, such as object permanence and solid-body mechanics. For example, if a solid wooden block appears to pass straight through another solid wooden wall, an infant will stare at the event significantly longer than at a physically possible event, indicating an expectation violation regarding physical solidity. Bloom, Wynn, and Hamlin realized that this visual fixation metric could serve as an exquisite probe of social and normative expectations. Rather than merely tracking physical laws, researchers could assess whether infants hold normative expectations about how social agents ought to interact with one another.

In a typical social violation-of-expectation experiment, infants are first habituated to a social drama—such as an agent struggling to attain a goal, followed by an intervention by a second agent who either helps or hinders that goal. Once this social dynamic is established, researchers present test events where the original agent initiates an approach toward either the helpful agent or the obstructive agent. If infants view social interactions through an amoral, purely random lens, they should display equivalent looking times to both approach events. However, if infants possess an internalized expectation that prosocial behavior invites positive social affiliation while antisocial behavior triggers social avoidance, approaching the hindering character should register as a violation of social reality, resulting in systematically longer looking times.

3.2 The Preferential Reaching and Choice Technique

While looking-time studies illuminate an infant’s cognitive expectations, they do not definitively answer whether the infant harbors an affective, valenced preference for one character over another. Staring longer at an entity might indicate surprise, curiosity, or confusion, but it does not necessarily denote moral approval or personal affinity. To overcome this interpretative ambiguity, Wynn and Bloom augmented their looking-time methodologies with an innovative tactile paradigm: the preferential reaching and choice technique. This methodology transforms passive infant observers into active behavioral decision-makers.

In the preferential reaching protocol, after watching an experimental puppet show unfold multiple times on a stylized wooden stage, the infant is presented with the actual physical characters mounted on a presentation tray. An experimenter, seated directly in front of the infant, holds the characters equidistant from the infant’s center of mass, slightly out of immediate reach, to ensure the infant visually scans both objects. The experimenter then advances the tray forward to within reach, holding their own gaze firmly fixed on the ceiling or closed behind dark sunglasses, entirely oblivious to which character is positioned on which side of the tray. This rigorous double-blind protocol prevents subconscious experimenter cuing via eye gaze, breathing alterations, or subtle physical prompts.

The primary dependent variable is the infant’s first purposeful, unambiguous physical reach and touch. To ensure scientific rigor, high-definition video recordings of the interaction are captured from multiple angles and independently evaluated by trained research coders who are blind to the experimental condition and the specific identities of the puppets. Coders assess multiple behavioral parameters:

• The exact trajectory of the infant’s arm extension and the character first contacted by their fingers or palm.
• The duration of physical contact and whether the infant grasps, strokes, or pulls the chosen character toward their body.
• Visual fixation during the reach, ensuring the choice is intentional rather than an accidental, sweeping motor spasm.
• Secondary affective expressions, such as smiling at the chosen puppet or averting gaze and displaying distress toward the unchosen puppet.

By establishing rigorous statistical thresholds for non-random reaching, Bloom and Wynn demonstrated that an infant’s tactile choice represents a reliable, robust read-out of their subjective preference for the character, providing direct behavioral evidence of sociomoral valence.

3.3 Apparatus Engineering and Experimental Controls

Conducting reliable psychophysical experiments with preverbal infants requires immense mechanical precision. To ensure that an infant’s preference for a particular puppet is genuinely driven by the sociomoral valence of its actions rather than irrelevant low-level perceptual artifacts, Bloom and Wynn engineered experimental setups with meticulous controls. A persistent danger in infant research is that babies might naturally prefer the color red over blue, circles over triangles, or rapid zigzagging trajectories over slow linear movements. If an experimental design happens to assign the “helpful” role to a bright yellow bouncing ball and the “hindering” role to a dull brown static cube, the infant’s subsequent choice might reflect trivial sensory aesthetics rather than moral appreciation.

To eliminate these perceptual confounds, the Baby Lab instituted exhaustive counterbalancing protocols across every experimental cohort:

• Counterbalancing Identity and Role: The physical identity, shape, and color of the characters were systematically rotated across participants. If Participant A observed a yellow triangle acting as the Helper and a blue square acting as the Hinderer, Participant B observed the exact inverse, with the blue square providing assistance and the yellow triangle causing obstruction.
• Counterbalancing Spatial Orientation: The lateral position of characters was strictly alternated. The Helper ascended from the right-hand side of the stage in half of the trials and from the left-hand side in the other half. Similarly, during the choice presentation phase, the Helper was placed on the infant’s left side for 50% of the presentations and on the right side for the remaining 50%, extinguishing any natural reaching biases related to right- or left-handedness.
• Standardization of Kinematics and Movement: The physical movement of the puppets was meticulously choreographed. Puppet manipulators were concealed behind heavy black curtains, operating the wooden agents via thin, matte-black rods attached to the rear of the characters. Movement velocities, collision durations, and trajectory arcs were timed to fractions of a second using calibrated metronomes or digital cueing lights invisible to the infant.
• Acoustic and Linguistic Neutrality: All interactions took place either in complete silence or accompanied exclusively by standardized, emotionally neutral synthetic chimes designed to mark the onset of an action without imparting emotional cues. The parents seated behind the infants were instructed to remain entirely silent, keep their eyes firmly closed, and refrain from touching, nudging, or interacting with the infant throughout the entirety of the experimental session.

Through these painstaking mechanical and environmental controls, the Yale researchers ensured that the only variable systematically differentiating the characters was the socio-intentional meaning embedded within their behaviors.

4. The Core Helper vs. Hinderer Paradigm: The Climber on the Hill

4.1 Narrative Mechanics of the Hill Scenario

The empirical breakthrough that catapulted the Yale Baby Lab into the international spotlight was the seminal “Climber on the Hill” experiment, published by J. Kiley Hamlin, Karen Wynn, and Paul Bloom in Nature in 2007. The experimental apparatus featured a small, three-dimensional wooden stage displaying a vivid green hill with two distinct inclines: a shallow lower slope and a steep upper crest. The protagonist of the drama was an anthropomorphized wooden circle—affectionately designated as the “Climber”—painted a bright color and fitted with large, expressive plastic googly eyes to unambiguously denote agency, intentionality, and gaze direction.

The experimental narrative unfolds in a carefully structured three-act sequence:

1. Act I: Establishing the Goal: The Climber enters at the base of the hill and begins to ascend the shallow slope. Upon reaching the midpoint where the slope becomes steep, the Climber stalls, rocks gently from side to side, and attempts to surge upward toward the plateau. The Climber makes two unsuccessful attempts to crest the summit, sliding back down to the midpoint each time, before coming to a complete rest while gazing upward at the peak. This sequence clearly establishes a teleological, goal-directed narrative: the Climber possesses an internal desire to reach the top of the hill but lacks the physical capacity to achieve it alone.
2. Act II: The Prosocial Intervention (The Helper): During the Helper condition, a second wooden character—for instance, a yellow triangle—enters the stage from the bottom of the hill. The triangle approaches the Climber from behind, pauses briefly, and then gently and rhythmically nudges the Climber from below, pushing it upward past the steep crest until the Climber successfully lands on the flat, stable plateau. Once the goal is achieved, the Climber dances gleefully in place, and the Helper quietly slides back down the hill and exits the stage.
3. Act III: The Antisocial Intervention (The Hinderer): During the Hinderer condition, the Climber begins again at the midpoint, making its typical stalled attempts to ascend. This time, a third character—such as a blue square—enters the stage from the top of the hill. Instead of providing upward momentum, the square charges aggressively downward into the Climber, violently colliding with it from above and knocking the Climber tumbling down the incline until it crashes to the very bottom of the hill. The Hinderer then triumphantly pauses at the crest before sliding off-stage, leaving the Climber lying motionless at the base.

Each infant participant observed these alternating scenarios multiple times—typically six to eight habituation trials—alternating systematically between helpful and unhelpful encounters until the social valence of the agents was thoroughly demonstrated.

4.2 Infant Choice Responses and Empirical Findings

Following the presentation of the puppet shows, Hamlin, Wynn, and Bloom transitioned the infants to the preferential reaching choice apparatus to assess their personal evaluations of the Helper and Hinderer characters. The results were extraordinary. When presented with a tray offering both the helpful triangle and the obstructive square, six-month-old and ten-month-old infants exhibited an overwhelming, statistically robust preference for the prosocial agent.

Across multiple cohorts, the preference for the Helper was remarkably high: approximately 80% to 100% of the tested infants reached directly for the character that had assisted the Climber, reaching past or actively turning away from the character that had knocked the Climber down. The statistical probability of such an outcome occurring purely by chance across these cohorts was p < 0.001. The effect was immediate, decisive, and consistent across both the six-month and ten-month age groups.

This finding carried profound implications for the developmental timeline of sociomoral cognition. Six-month-old infants are pre-linguistic; they cannot speak words, nor can they comprehend complex grammatical instructions. Most six-month-olds are not yet independently locomotive; they have not spent years crawling or walking through their physical environment negotiating social agreements or enduring formal discipline from peers. Yet, despite their sensory and motor immaturity, these infants demonstrated an unmistakable capacity to parse an ambiguous visual scene, extract the underlying intentional goals of an actor, evaluate a third-party interaction, and generate an affective alignment toward the prosocial actor. They preferred the good guy.

4.3 Perceptual Control Conditions

To defend these revolutionary findings against skeptical claims that infants were merely reacting to physical mechanics, Hamlin, Wynn, and Bloom devised a series of rigorous perceptual control conditions. Skeptics hypothesized that infants might not be evaluating morality at all, but rather responding to raw physical vectors: perhaps babies simply possess an inherent aesthetic preference for objects that move upward rather than downward, or perhaps they were captivated by the physical “launching” kinematics described in Michottean perceptual physics.

To definitively address these concerns, the researchers stripped the characters of their intentionality while retaining the exact physical movements and collision kinematics. In one crucial control experiment, the googly eyes were removed from the wooden characters. Without eyes, the central round object ceased to appear as an intentional, animate creature with desires and sightlines; it became merely a lifeless wooden disc. The disc rolled up the hill, paused, and was either bumped further upward by the triangle or bumped downward by the square.

When six- and ten-month-old infants were presented with the eyeless Helper and Hinderer blocks after observing these identical physical movements, their preferential reaching vanished completely. The infants chose between the two blocks at pure chance levels (roughly 50/50). The upward-pushing motion was no longer preferred, and the downward-colliding motion was no longer avoided. This demonstrated that infants do not possess a low-level, hardwired bias for upward physical trajectories. Their preference for the Helper in the original experiment was contingent upon the attribution of agency and teleological intentionality: they only favored the upward-pushing character when that push served the subjective goals of an intentional agent who wanted to climb the hill.

5. Evaluating Prosociality and Antisociality: Preferential Reaching and Looking Metrics

5.1 The Valence Question: Valence Attraction vs. Aversion

The discovery that infants prefer the Helper over the Hinderer introduced a fundamental theoretical question: What is the primary psychological driver of this preference? Is the infant’s choice propelled by a positive affective attraction toward the benevolent Helper, an active aversion and moral condemnation toward the malevolent Hinderer, or a simultaneous combination of both valences? A simple binary choice between a Helper and a Hinderer cannot resolve this question, as a preference for Puppet A over Puppet B could mean that the infant loves Puppet A, hates Puppet B, or is indifferent to Puppet A while actively despising Puppet B.

To resolve this question, Hamlin, Wynn, and Bloom (2007) introduced a baseline: a Neutral character. The Neutral character was introduced onto the hill stage alongside the Climber, but instead of interacting with the Climber, it simply traversed the stage passively, rolling past the Climber without helping or hindering its ascent. Following habituation to the drama, infants were presented with two distinct types of choice pairings:

• Helper versus Neutral: When presented with a choice between the character that helped the Climber and the character that did nothing, infants overwhelmingly selected the Helper. This demonstrated an active positive valence: infants do not merely seek out non-harmful agents; they positively appreciate and are drawn to agents that perform altruistic acts.
• Hinderer versus Neutral: When presented with a choice between the character that knocked the Climber down and the passive Neutral character, infants overwhelmingly chose the Neutral character, actively avoiding the Hinderer. This demonstrated an active negative valence: infants harbor an authentic social aversion to agents that impede others’ goals.

Crucially, Hamlin and colleagues tested these dynamics across different developmental stages. They discovered that while both six-month-olds and ten-month-olds reliably preferred the Helper over the Hinderer, three-month-old infants (tested via looking-time preferences rather than motor reaching) already displayed an aversion toward the Hinderer, while the robust positive attraction to the Helper became more pronounced slightly later. This implies that avoiding dangerous, uncooperative social agents may be the primary evolutionary baseline, with active prosocial admiration developing as cognitive and social capacities mature.

5.2 The Box-Opening and Ball-Dropping Paradigms

To establish that infant sociomoral evaluation is a domain-general cognitive capacity rather than a narrow response idiosyncratic to the hill-climbing scenario, the Yale researchers engineered entirely novel narrative paradigms utilizing distinct characters, goals, and thematic settings.

In the Box-Opening Paradigm (Hamlin & Wynn, 2011), the drama featured a central protagonist—a plush dog or duck puppet—struggling to achieve a transparent mechanical goal. The protagonist stands in front of a clear plastic box containing an attractive rattle. The protagonist repeatedly grasps the lid of the box and attempts to pull it open, but the lid repeatedly falls shut, leaving the toy trapped inside. Once this unfulfilled desire is established, two secondary puppets intervene:

• The Helpful Box-Opener: A puppet bounds onto the stage, grasps the opposite side of the box lid alongside the protagonist, and pulls the lid completely open, allowing the protagonist to happily retrieve the rattle.
• The Obstructive Box-Closer: A puppet jumps onto the box, forcibly slams the lid shut, and sits aggressively on top of it, completely thwarting the protagonist’s attempts to retrieve the toy before sprinting away.

In the Ball-Dropping Paradigm, the narrative tested reciprocal social play. A central protagonist puppet plays with a bright, bouncing ball in the center of the stage. During the play session, the ball accidentally slips from the protagonist’s grasp and rolls across the stage toward one of two secondary puppets standing at the perimeter:

• The Prosocial Returner: When the ball rolls to this puppet, it catches the ball and deliberately rolls or passes it directly back to the original protagonist, allowing the play session to continue.
• The Antisocial Thief: When the ball rolls to this puppet, it seizes the ball in its paws and immediately bolts off-stage, stealing the toy and leaving the protagonist abandoned and empty-handed.

When tested using the preferential reaching paradigm, infants as young as three to five months systematically reached for the puppet that helped open the box and the puppet that returned the ball, while rejecting the box-slammer and the toy-thief. The successful replication across these radically different physical and narrative landscapes confirmed that infants are not responding to hill-climbing mechanics; they are deploying an abstract, domain-general socio-evaluative metric capable of parsing diverse third-party actions into fundamental categories of helpfulness and harmfulness.

5.3 Looking Time Asymmetries and Normative Expectations

Beyond evaluating characters for themselves, do infants expect other individuals to behave in morally coherent ways? To investigate whether infants possess an implicit normative framework regarding how social relationships should unfold, researchers returned to the violation-of-expectation looking-time paradigm.

In these studies, ten-month-old infants were habituated to the standard hill-climbing narrative until they were fully accustomed to the identities of the Helper and the Hinderer. Once habituated, the infants were shown two novel test events that depicted the original Climber taking action, rather than the secondary agents:

1. The Congruent Event: The Climber, resting at the bottom of the hill, initiates an approach trajectory toward the Helper, resting peacefully next to it, while ignoring the Hinderer.
2. The Incongruent Event: The Climber initiates an approach trajectory toward the Hinderer, sidling up to and affiliating with the character that had previously knocked it down, while ignoring the Helper.

The empirical findings revealed a stark asymmetry: ten-month-old infants stared significantly longer at the incongruent event where the Climber approached its abuser. Within the methodological logic of the violation-of-expectation paradigm, this prolonged visual gaze reveals that infants found the Climber’s affiliation with the Hinderer to be deeply surprising—a violation of their mental model of social reality. They expected the victim to align with its benefactor and shun its tormentor. Interestingly, when this exact experiment was conducted with six-month-old infants, the looking-time asymmetry was less distinct, suggesting that while six-month-olds can evaluate characters for their own reaching choices, the sophisticated expectation that third parties will selectively affiliate based on past moral history consolidates between six and ten months of age.

6. Third-Party Evaluations and the Emergence of Retributive Justice

6.1 Evaluating Actions Directed at Prior Actors

In the real social world, actions do not occur in isolated, two-agent vacuums. Social life is a continuous, interconnected tapestry where individuals interact with others who themselves have a rich history of good and bad deeds. A mature moral sense does not treat all aggression as universally evil, nor does it treat all assistance as universally virtuous. If a police officer tackles an armed robber, an adult observer does not judge the officer as a violent villain; the aggression is contextualized as a socially necessary, retributive intervention. Conversely, if an individual helps a fugitive escape justice, that “helpful” act is condemned as an antisocial obstruction of law.

To investigate whether the preverbal infant mind is capable of this type of multi-tiered, hierarchical social calculation, J. Kiley Hamlin, Karen Wynn, Paul Bloom, and Neha Mahajan devised an ingenious experimental architecture that placed infants in the role of third-party evaluators judging actions directed at prior actors. Rather than evaluating an agent who helps or harms an innocent protagonist, the infant was presented with a scenario where the target of the assistance or harm was an individual who had already established a clear moral track record as either a Helper or a Hinderer.

This required infants to compute a complex, higher-order social calculus: they had to synthesize the valence of a primary interaction, store the moral reputation of that actor in working memory, observe a secondary interaction visited upon that actor, and subsequently evaluate the moral valence of the secondary actor based on whether the target deserved reward or punishment. This paradigm tested whether infants are simplistic pacifists who blindly adore anyone who is nice and despise anyone who is mean, or whether they possess the foundational roots of desert-based justice.

6.2 Early Intuitions of Retribution and Punishment

The experimental setup for investigating retributive justice utilized a two-phase dramatic sequence combining the Box-Opening and Ball-Dropping paradigms. In Phase One, infants watched a standard puppet show establishing the moral reputation of two characters: one puppet acted as a prosocial Helper (helping open a box to retrieve a toy), while another acted as an antisocial Hinderer (slamming the box shut). In Phase Two, these newly established characters were placed in the center of the stage playing with a ball, which they accidentally dropped.

New secondary actors then intervened, creating four distinct experimental conditions:

• Condition A (Rewarding a Helper): A secondary actor returns the dropped ball to the puppet who had previously been a prosocial Helper.
• Condition B (Punishing a Helper): A secondary actor steals the ball away from the puppet who had previously been a prosocial Helper.
• Condition C (Rewarding a Hinderer): A secondary actor returns the dropped ball to the puppet who had previously been an antisocial Hinderer.
• Condition D (Punishing a Hinderer): A secondary actor steals the ball away from the puppet who had previously been an antisocial Hinderer.

When Hamlin and colleagues (2011) offered infants the choice between the secondary actors using preferential reaching, the results overturned conventional assumptions about child psychology. When judging actions directed toward the Helper, infants behaved predictably: they chose the puppet that was nice to the Helper (Condition A) and rejected the puppet that was mean to the Helper (Condition B). But when evaluating actions directed toward the Hinderer, the infants’ preferences underwent a profound reversal. Rather than favoring the puppet that returned the ball, eight-month-old infants overwhelmingly chose the puppet that punished the Hinderer by stealing its ball (Condition D). They actively rejected the puppet that attempted to be nice to the bad guy (Condition C).

This remarkable finding demonstrated that infants do not possess a simple-minded, blanket affinity for nice behavior. They do not view an act of aggression in a vacuum. When an agent inflicts harm upon an antisocial individual, infants perceive that harm not as an act of malicious violence, but as justified retribution. The desire to see wrongdoers punished—the psychological bedrock of retributive justice and institutional penal law—appears to be hardwired into human cognition, operating long before children are exposed to cultural doctrines of crime and punishment.

6.3 Valence Reversal: When Harming Is Regarded as ‘Good’

The cognitive flexibility required to execute this valence reversal is conceptually extraordinary for an eight-month-old infant. In formal philosophical terms, this phenomenon represents a transition from first-order moral assessment to second-order moral evaluation. In a first-order framework, physical actions possess fixed valences: giving an object is always “good,” while taking an object away is always “bad.” If infants operated strictly within this low-level affective heuristic, they would unconditionally reward any agent that handed over a ball and condemn any agent that snatched one away.

Instead, the infant mind deploys a contextual moral algorithm. The valence of the action (positive or negative) is multiplied by the valence of the recipient’s moral status (positive or negative). In a manner strikingly analogous to the mathematical rule where multiplying two negative numbers yields a positive number, the infant recognizes that inflicting a negative action upon a negative actor yields a positive moral outcome:

[Negative Action: Ball Theft] × [Negative Recipient: Box Slammer] = [Positive Sociomoral Valence: Justified Retribution]

This valence reversal provides compelling evidence that preverbal infants are not simply running on associative conditioning or primitive emotional empathy. Empathy would dictate that an infant feel distressed whenever any puppet is deprived of its toy, prompting them to comfort the victim regardless of its history. Instead, infant judgment overrides primitive empathy in favor of an abstract moral calculus of desert. The bad actor has forfeited their claim to moral protection, and the agent who administers punishment is elevated to the status of a moral hero. This dynamic offers a powerful developmental antecedent to the adult “just-world” phenomenon and explains why human societies throughout history have universally embraced the moral legitimacy of punitive justice against those who violate social contracts.

7. In-Group Favoritism and the Evolutionary Roots of Parochial Altruism

7.1 Minimal Group Paradigms in Infancy

The findings demonstrating that infants possess an innate sense of justice and a preference for kindness paint an undeniably optimistic portrait of human nature. However, evolutionary biology warns against viewing the human animal as an indiscriminate wellspring of universal benevolence. In ancestral environments, cooperation was dangerous if extended blindly to members of competing out-groups. Survival necessitated a sharp cognitive boundary between the cooperative “us” and the threatening “them.” To determine how deep these tribal instincts run, Karen Wynn, Paul Bloom, and their post-doctoral researcher Neha Mahajan designed experiments using adapted versions of social psychology’s famous minimal group paradigm.

In adult social psychology, Henri Tajfel historically demonstrated that adults will display in-group favoritism based on the most trivial, arbitrary distinctions imaginable, such as flipping a coin to create groups or sorting people based on whether they overestimate or underestimate dots on a screen. To replicate this in the preverbal realm, Mahajan, Hamlin, and Wynn (2011) introduced a minimal group distinction rooted in sensory preference: food selection.

In this experiment, an infant participant is seated at a table and presented with two small bowls containing different snacks: one containing sweet graham crackers, and the other containing savory green beans. The experimenter allows the infant to taste both options and explicitly declare a personal preference (e.g., the infant enthusiastically points to or consumes the graham crackers). Once the infant’s preference is established, two animal puppets—such as an orange cat and a gray bunny—are placed on the stage. Each puppet is offered the snacks:

• The Similar (In-Group) Puppet: Tastes the graham crackers, rubs its belly, and declares with delight, “Mmm, yum! I love graham crackers!” while ignoring the green beans.
• The Dissimilar (Out-Group) Puppet: Tastes the green beans, rubs its belly, and declares, “Mmm, yum! I love green beans!” while rejecting the graham crackers.

When infants were subsequently presented with both puppets in a preferential reaching test, the results were unequivocal: infants overwhelmingly reached for the puppet that shared their specific food preference. A trivial, non-moral dietary choice was sufficient to instantaneously divide the puppet world into an “in-group” ally and an “out-group” alien.

7.2 The Dark Side of Infant Morality: In-Group Bias and Out-Group Derogation

The critical question for moral psychology was not simply whether infants prefer individuals who are similar to themselves, but how this tribal categorization intersects with their evaluations of justice and harm. Does an infant’s moral framework remain universally objective when applied to an out-group member, or does their moral compass bend to favor the tribe?

To test this, Hamlin, Mahajan, Liberman, and Wynn (2013) combined the food-preference minimal group paradigm with their classic Helper and Hinderer scenarios. After the infant witnessed one puppet declare an identical food preference (In-Group) and another puppet declare an opposing food preference (Out-Group), these puppets were placed into the Ball-Dropping drama, where they were either assisted by a secondary Helper puppet or victimized by a secondary Hinderer puppet.

When watching actions directed toward the In-Group puppet, infants adhered to standard moral expectations: they strongly preferred the character that helped the In-Group puppet and rejected the character that harmed it. But when observing actions directed toward the Out-Group puppet—the character that liked the “wrong” food—the infants’ moral preferences underwent a disturbing inversion:

• Infants preferred characters that harmed, mistreated, and stole from the Out-Group character over characters that were kind to it.
• Infants viewed prosocial assistance toward an out-group member with active moral disdain, treating the Helper of the dissimilar puppet as if it had done something wrong.

These findings uncovered the profound evolutionary limits of innate morality. The preverbal human mind is not an engine of universal human rights; it is hardwired for parochial altruism—the evolutionary pairing of internal group altruism with external hostility toward out-groups. Long before an infant is exposed to cultural racism, xenophobia, or nationalistic propaganda, their cognitive architecture spontaneously generates tribal boundaries. Harm visited upon an outsider is not merely tolerated; it is actively savored—an early developmental precursor to schadenfreude.

7.3 Bloom’s Dual-System Perspective: Empathy Versus Reason

The discovery of pervasive in-group bias in infants led Paul Bloom to articulate a critical philosophical and psychological argument, synthesized extensively in his book Against Empathy: The Case for Rational Compassion (2016) and Just Babies: The Origins of Good and Evil (2013). Bloom forcefully argues that unreflective emotional empathy—frequently celebrated as the ultimate moral virtue—is actually a deeply flawed, parochial emotion. Empathy is an evolutionary spotlight: it shines intensely on those who look like us, talk like us, and share our tastes, while leaving everyone else in the cold dark.

Bloom posits a dual-system model of human morality:

1. System 1: The Innate Moral Sense: A fast, automatic, emotionally driven core morality that evolved in ancestral tribal environments. This system generates automatic feelings of compassion for friends, instantaneous moral outrage at direct physical harm, an obsession with retributive punishment, and a fierce, reflexive bias favoring the in-group over the out-group.
2. System 2: Rational Moral Capacities: A slow, deliberate, culturally scaffolded intellectual apparatus that utilizes formal logic, perspective-taking, philosophy, and institutional laws. This system enables humanity to recognize that an out-group member possesses the same fundamental human rights as an in-group member, despite our primal instincts screaming otherwise.

Bloom concludes that the greatest moral achievements of human civilization—the abolition of slavery, universal human rights, international humanitarian law, and the eradication of tribal blood feuds—are not the triumphs of our innate infant moral intuitions. Rather, they are the triumphs of reason over instinct. Infant morality gives us the raw emotional materials of kindness and justice, but it also burdens us with tribal hatred and xenophobia. It is only through deliberate philosophical education, cultural evolution, and institutional scaffolding that humanity expands the moral circle beyond the narrow boundaries of the crib.

8. Developmental Trajectories: How Innate Moral Rudiments Evolve Across Childhood

8.1 The Transition to Active Prosocial Behavior

While the Yale puppet studies primarily isolated passive socio-evaluative choices—measuring where an infant looks or which puppet they reach out to grasp—a complete account of moral development must examine the bridge between these receptive evaluations and active, costly prosocial behavior. How does an infant transform from an armchair moral critic into an actively altruistic agent?

Pioneering work by developmental psychologists Felix Warneken and Michael Tomasello at the Max Planck Institute for Evolutionary Anthropology revealed that by twelve to fourteen months of age, human infants spontaneously initiate instrumental helping without any parental prompting or material reward. In these classic experiments, an adult experimenter accidentally drops an object (such as a clothespin or a marker) out of reach while attempting to complete a task, or struggles to open a cabinet door with their hands full. Upon witnessing the adult’s predicament, toddlers will immediately crawl or waddle across the room, retrieve the dropped item, and hand it back to the adult, or push the cabinet door open for them.

Crucially, Warneken and Tomasello demonstrated that toddlers engage in this instrumental helping even when doing so requires navigating physical obstacles, leaving behind engaging toys, and receiving no praise or physical treats. In fact, subsequent studies revealed an “overjustification effect”: when toddlers were given material rewards (such as toys or stickers) for helping, their intrinsic motivation to help in future trials actually decreased compared to toddlers who received no reward at all. This confirmed that the drive to assist others in achieving their physical goals is an intrinsically motivated, biologically prepared adaptation that blossoms organically as motor competence expands.

8.2 Fairness, Resource Distribution, and Inequity Aversion

As infants cross the threshold into toddlerhood and early childhood, their sociomoral cognition expands into the complex domain of distributive justice and resource allocation. Do children possess an inherent sense of fairness, and how do they balance this ideal against their own selfish desires?

Experimental studies spearheaded by Jessica Sommerville and colleagues demonstrated that by fifteen months of age, infants hold implicit expectations of equal resource distribution. Utilizing violation-of-expectation looking-time paradigms, researchers presented infants with an adult distributor dividing milk or cookies between two recipients. When the distributor divided the resources equitably (e.g., two cookies to Recipient A and two cookies to Recipient B), infants habituated normally. However, when the distributor enacted an unequal distribution (e.g., three cookies to Recipient A and one cookie to Recipient B), fifteen-month-olds looked significantly longer, indicating that asymmetrical allocation violates their core baseline expectation of fairness.

However, the active behavioral enforcement of fairness reveals a fascinating, asymmetrical developmental progression characterized by two distinct psychological phenomena:

• Disadvantageous Inequity Aversion: The intense psychological aversion to receiving less than an interaction partner. This emerges robustly around three to four years of age. If a child receives one sticker while their peer receives four, the child reacts with visceral distress and moral outrage, frequently rejecting their own single sticker to protest the unfairness.
• Advantageous Inequity Aversion: The much more sophisticated, morally demanding aversion to receiving more than an interaction partner. This phenomenon does not reliably emerge until approximately seven to eight years of age. In early childhood, children are perfectly delighted to hoard ten candies while their peer receives one. It requires years of cognitive development, executive inhibitory control, and social enculturation before older children willingly surrender their own unearned surplus to ensure an equitable distribution.

Studies adapting economic games—such as the Dictator Game and Ultimatum Game—for children reveal that by age eight, children will actively pay a personal cost to destroy resources entirely rather than allow an unfair distribution to stand, demonstrating the consolidation of mature, egalitarian distributive justice.

8.3 Linguistic Enculturation and Moral Reasoning

The acquisition of language fundamentally transforms the architecture of the moral mind, constructing a sophisticated linguistic interface over the preverbal core intuitions revealed by the Yale puppet studies. Language provides children with a symbolic taxonomy—words such as “mean,” “fair,” “stealing,” “lying,” and “promise”—that allows them to categorize, debate, and modulate their automatic sociomoral reactions.

This transition illuminates why early theorists such as Piaget and Kohlberg drew such dramatically pessimistic conclusions regarding young children’s moral capabilities. When an adult asks a four-year-old child why an action was wrong, the child’s linguistic repertoire is fundamentally inadequate to convey the computational sophistication of their intuitive evaluations. The child may default to crude, concrete explanations, such as “Because mommy will yell at me” or “Because the cup broke into a million pieces.” These responses led Kohlberg to infer that the child’s internal moral universe was governed exclusively by a preconventional fear of punishment.

In reality, developmental cognitive science now recognizes that speech-based moral justifications are frequently post-hoc rationalizations. As social psychologist Jonathan Haidt demonstrated in his Social Intuitionist Model, moral evaluations are primarily generated by fast, automatic affective intuitions (the modern evolutionary equivalent of Bloom and Wynn’s infant core morality). The slow, conscious, verbal reasoning mind operates not as an objective judge discovering truth, but as a skilled lawyer fabricating verbal justifications after the intuitive verdict has already been rendered. Language does not generate our fundamental moral sentiments; rather, it formalizes, debates, and culturally refines the primal moral evaluations that have been operating quietly since infancy.

9. Methodological Critiques, Perceptual Biases, and Replication Efforts

9.1 The Perceptual Attractiveness Challenge (Scarf et al.)

Despite the immense global impact of the Yale Baby Lab’s findings, the methodology was not without fierce academic contention. The most prominent methodological challenge emerged in 2012, when psychologist Damian Scarf and colleagues published a provocative critique in the journal Developmental Science. Scarf asserted that Hamlin, Wynn, and Bloom had fundamentally misread their data, falling prey to a subtle, low-level perceptual confound: the physical phenomenon of movement kinematics and bouncing.

Scarf and his team pointed out a seemingly innocuous detail in the choreography of the original Hill Climbing experiment. In the prosocial Helper scenario, when the Helper successfully pushed the Climber to the summit, the Climber celebrated by engaging in a rhythmic, happy “bounce” up and down on the plateau. Conversely, in the antisocial Hinderer scenario, when the Climber was knocked to the bottom of the hill, it came to a static, motionless halt without any bouncing. Furthermore, the Helper puppet itself frequently bounced while pushing. Scarf hypothesized that infants possess a low-level, hardwired visual fascination with bouncing objects. Through simple Pavlovian associative conditioning, infants might have associated the Helper character with the attractive, joyful bouncing movement, leading them to reach for the Helper purely as a perceptual preference for a “bouncing-associated object,” completely divorced from any comprehension of helping or hindering.

To substantiate this critique, Scarf and colleagues replicated the experiment while manipulating the presence of the bounce:

• When they altered the animation so that the Climber bounced at the bottom of the hill after being knocked down by the Hinderer, infants reversed their preference and systematically selected the Hinderer!
• When all bouncing movements were completely removed from both conditions, infants selected between the Helper and Hinderer at chance levels.

Scarf concluded that the “moral life of babies” was an illusion—an artifact of unconstrained kinematic variables masking simple visual salience preferences.

9.2 Replication Debates and the Reproducibility Project

The Scarf critique ignited a fierce methodological debate. Kiley Hamlin, Paul Bloom, and Karen Wynn issued a detailed rebuttal, identifying severe methodological flaws in Scarf’s counter-experiments. The Yale researchers argued that Scarf had inadvertently dismantled the social narrative of the scenario. In the original Yale choreography, the Climber bounced only after its goal was achieved, functioning as a vital behavioral cue of emotional satisfaction and success. By forcing the Climber to bounce at the bottom of the hill after being struck, Scarf had created a bizarre, incoherent social interaction where an agent appeared delighted to be violently assaulted, confusing the infant’s ability to discern whether the target’s goal had been fulfilled or thwarted.

To definitively put the kinematic critique to rest, Hamlin and colleagues conducted new, highly controlled studies. They demonstrated that even when all bouncing was completely eliminated—or when the Helper and Hinderer engaged in identical physical kinematics—infants continued to robustly prefer the Helper, provided that the characters maintained clear social intentionality. Furthermore, they emphasized that the Scarf critique applied exclusively to the Hill Climbing setup and could not explain the identical moral preferences observed in the Box-Opening and Ball-Dropping paradigms, where no bouncing movements occurred.

Nevertheless, the debate underscored the acute fragility of infant empirical research. Infant studies inherently suffer from high attrition rates—frequently ranging from 20% to 40% of participants excluded due to crying, fussiness, distraction, parental interference, or procedural failure. In response to the broader “replication crisis” in psychology, large-scale multi-laboratory collaborations, such as the ManyBabies Consortium, have mobilized to conduct massive, pre-registered replications of foundational developmental paradigms. While the core finding that infants evaluate agents based on social actions has retained broad empirical support, multi-lab replications have revealed that effect sizes can fluctuate considerably depending on precise environmental parameters, including display sizes, puppet textures, and the precise timing of behavioral cues.

9.3 Experimenter Bias and Blind Presentation Challenges

Another major methodological frontier in infant moral cognition involves the elimination of subconscious human bias. In physical puppet theater, the experimental actors are human beings operating rods behind curtains, and the final choice presentation is administered by a human experimenter holding a tray. Even the most rigorous blind protocols—such as experimenters wearing sunglasses, looking at ceilings, or keeping their eyes closed—leave theoretical room for skepticism. Could an experimenter subconsciously tilt the tray by a fraction of a millimeter toward the Helper? Could their breathing shift slightly, or could their muscle tension signal an expectation to the hyper-attuned infant?

To resolve these epistemological concerns, the field has increasingly shifted toward automated eye-tracking and digital 2D/3D animations. In these modern paradigms, infants sit in front of high-resolution monitors equipped with infrared eye-tracking cameras that calculate corneal reflections at 300 to 1000 Hz. The social dramas are rendered via computer-generated animations, eliminating any possibility of human actor variability. The infant’s preference is measured via precise gaze fixations, pupil dilation metrics, and anticipatory saccades.

While digitized animations offer absolute experimental standardization, they introduce their own ecological dilemmas. Infants process three-dimensional physical objects in their immediate physical space with far greater cognitive engagement and emotional salience than flat, pixelated representations on a screen. Comparative analyses reveal that physical puppet interactions consistently evoke stronger affective choices and higher reach-engagement rates than automated 2D video screens. Navigating the delicate boundary between technological precision and real-world ecological validity remains one of the central methodological challenges in contemporary developmental cognitive neuroscience.

10. Theoretical Interpretations: Nativism, Core Knowledge, and Evolutionary Psychology

10.1 The Core Social Cognition Hypothesis

How does the human mind translate the visual perception of moving shapes into rich, valenced moral evaluations? Developmental nativists interpret the findings of the Yale Baby Lab through the framework of Core Knowledge Systems. Popularized by cognitive scientists Elizabeth Spelke and Susan Carey, core knowledge theory posits that human infants do not enter the world with a general-purpose, undifferentiated learning engine. Instead, natural selection has equipped the human brain with distinct, evolutionarily ancient, domain-specific computational modules:

• Core Object Mechanics: Dedicated to tracking physical cohesion, continuity, solidity, and contact mechanics of inanimate objects.
• Core Number: Dedicated to representing discrete numerical quantities and ratios.
• Core Geometry: Dedicated to navigating spatial environments via geometric layouts.
• Core Social Cognition: A dedicated mental module designed specifically to track animate agents, identify subjective goals, compute social cost-benefit ratios, and assign moral valence.

The Core Social Cognition module operates via a specialized computational architecture. When an infant views an entity in their environment, the module first runs an “Agency Check.” Does the entity possess eyes? Does it exhibit self-propelled motion? Does it change its velocity contingently in response to other entities without physical contact? If the answer is yes, the module switches the entity from the category of “Inanimate Physical Object” to “Animate Social Agent.”

Once classified as an agent, the infant’s brain ceases to analyze the scene using standard Newtonian physics (mass, momentum, gravity). Instead, it deploys a teleological stance: it infers the agent’s internal mental states, desires, and goals. When a second agent intervenes, the module computes whether the second agent’s action reduced or expanded the energetic cost required for the first agent to attain its goal. If Agent B reduces Agent A’s cost to achieve a desired state, Agent B is tagged with a positive social valence (Helper). If Agent B increases the cost or makes goal attainment impossible, Agent B is tagged with a negative social valence (Hinderer). This computation occurs automatically, pre-linguistically, and modularly.

10.2 Evolutionary Pressures Driving Innate Morality

From an evolutionary standpoint, why should such an elaborate computational system exist in a helpless, preverbal infant who is months away from hunting, foraging, or defending the tribe? The answer lies in the intense selective pressures of partner choice and coalitionary living. In highly cooperative hominin societies, choosing the wrong social partner—such as an untrustworthy, aggressive, or parasitic individual—carried fatal consequences. An individual who could not quickly distinguish altruistic cooperators from selfish defectors would be vulnerable to starvation, violence, and reproductive failure.

In evolutionary game theory, models of biological altruism demonstrate that cooperative strategies like Tit-for-Tat or Indirect Reciprocity can only stabilize in a population if cooperative individuals can preferentially interact with one another while isolating and punishing free-riders. The ability to monitor third-party interactions and construct a mental database of individual social reputations is the vital cognitive infrastructure that enables indirect reciprocity to function. An individual does not need to be personally burned by a defector; if they can observe that defector mistreating a third party, they can proactively avoid them.

Infant morality represents the ontogenetic emergence of this survival architecture. The preverbal baby is not engaging in abstract philosophical contemplation; they are running an evolutionary partner-screening program. By showing a powerful behavioral preference for Helpers, infants demonstrate an innate drive to form attachments with, affiliate with, and learn from cooperative, benevolent individuals. By avoiding and endorsing punishment against Hinderers, they deploy an ancient defense mechanism designed to insulate themselves and their kin from social toxicity.

10.3 Reconciling Nativism with Cultural Relativism

If human morality is rooted in an innate biological core, how do we account for the breathtaking diversity of moral codes across human history and culture? How can a nativist theory of infant morality explain why some cultures practice female genital mutilation, human sacrifice, or public executions, while others embrace universal human rights, radical pacifism, and animal liberation?

The resolution to this paradox lies in understanding that innate moral architecture is not a rigid, fully formed set of explicit adult laws. Rather, it operates like Jonathan Haidt’s metaphor of the moral mind as a “first draft” written by natural selection, which is subsequently revised, edited, and elaborated by culture. Haidt’s Moral Foundations Theory posits that humans possess innate psychological foundations—such as Care/Harm, Fairness/Cheating, Loyalty/Betrayal, Authority/Subversion, and Sanctity/Degradation. Different cultural environments act as moral tuners, dramatically amplifying some foundations while suppressing others.

Paul Bloom synthesizes nativism with cultural relativism by separating morality into its biological core and its cultural intellect:

• The Biological Core: Universal across all healthy human infants. It consists of an aversion to unprovoked suffering, a preference for helpful over obstructive actions, a rudimentary drive for fairness in simple resource allocation, an endorsement of retributive desert, and an instinctive bias toward the in-group.
• The Cultural Intellect: Highly variable and historically contingent. Culture dictates who belongs in the moral circle. In a xenophobic warrior society, the in-group boundary is drawn strictly around the immediate tribe, and members of competing tribes are categorized as non-human pests. In a modern cosmopolitan democracy, the moral circle expands outward to encompass all human races, nationalities, sexualities, and even non-human animals.

Thus, cultural relativism does not refute nativism. Rather, culture operates upon the innate parameters of infant sociomoral cognition, determining whether our ancient tribal switches are flipped on or systematically dialed down in favor of universal humanism.

11. Philosophical and Ethical Implications: Defining the Nature of Human Goodness

11.1 Rousseau versus Hobbes: Revisiting the Philosophical Classic

The empirical revelations of the Yale puppet studies inject hard behavioral data into one of the most enduring philosophical duels in Western thought: the conflict between the political philosophies of Jean-Jacques Rousseau and Thomas Hobbes.

In his Discourse on Inequality (1755), Rousseau articulated the vision of the “noble savage,” arguing that human beings are naturally good, compassionate, and endowed with an innate capacity for pitié (compassion or empathy) in the state of nature. For Rousseau, cruelty, greed, and social hierarchy are not innate human diseases, but artificial corruptions introduced by civil society, private property, and institutional tyranny. If humans could only be stripped of social indoctrination, their natural benevolence would reign.

Conversely, Thomas Hobbes argued in Leviathan (1651) that the natural state of humanity is a terrifying anarchic nightmare—a “war of all against all” where life is famously “solitary, poor, nasty, brutish, and short.” Hobbes asserted that humans are innately governed by egoism, vanity, and violent competition. Genuine morality, justice, and social peace are impossible without an all-powerful sovereign—the Leviathan—to impose order through the terror of legal punishment. Left to our natural instincts, we would destroy one another.

Bloom and Wynn’s puppet studies demonstrate that both classic philosophers were half-right and half-wrong:

• Rousseau was vindicated regarding the existence of an original, innate human goodness. Human infants are not cold, selfish, hedonistic automatons as claimed by radical behaviorists and extreme Hobbesians. Long before any schoolmaster or priest preaches a sermon, babies applaud kindness, display spontaneous compassion, and prefer altruistic agents. Human beings possess an authentic biological predisposition toward prosociality.
• Hobbes was vindicated regarding the innate darkness of human nature. The same infant mind that applauds altruism is fundamentally tribal, xenophobic, and vengeful. Babies do not enter the world loving all humanity. They instantaneously divide the world into “us” versus “them,” they relish the suffering of those who are different, and they enthusiastically endorse retributive violence. Without civilization, culture, and rational law, the natural infant mind would remain trapped in a dark, parochial tribalism.

Thus, the true nature of human goodness is not an absolute state of grace or a condition of total depravity, but an ongoing, internal evolutionary tug-of-war between our innate capacity for universal compassion and our hardwired impulse toward tribal exclusion.

11.2 The Epistemic Status of Infant Moral Judgments

A rigorous philosophical analysis must confront a vital meta-ethical problem: Can an infant’s behavioral preference for a puppet legitimately be designated as a moral judgment, or is it merely an aesthetic or social preference? When an adult states, “Torturing an innocent child is evil,” they are not merely expressing a personal sensory dislike analogous to “I do not care for broccoli.” They are making a normative truth claim that implies moral obligation, objective impermissibility, and a demand that any rational agent act accordingly.

Can a six-month-old infant reaching for a wooden triangle be credited with holding an objective normative judgment? Critics suggest that what the Yale Baby Lab measures is simply affective valence: the infant finds the Helper puppet visually or socially “pleasing” and finds the Hinderer “unpleasant.” In traditional Kantian deontology, an action possesses true moral worth only when it is undertaken from a conscious sense of duty dictated by universal moral law, parsed through rational categorical imperatives. Under this strict rationalist standard, an infant who cannot conceive of universal laws cannot be considered a moral agent.

However, the infant data provides profound empirical support for the philosophical tradition of Moral Sentimentalism, championed by eighteenth-century Scottish philosophers David Hume and Adam Smith. In A Treatise of Human Nature (1739), Hume famously asserted that “Reason is, and ought only to be the slave of the passions.” Hume argued that moral distinctions do not derive from abstract intellectual deductions, but from immediate moral sentiments: an internal “moral sense” that experiences an involuntary feeling of approbation when witnessing benevolence, and an involuntary feeling of disgust when witnessing cruelty.

Bloom and Wynn’s findings show that Hume was scientifically prescient. The infant’s evaluations are the raw, unadulterated manifestation of Humean moral sentiments in their purest ontogenetic form. While these infant preferences lack the philosophical meta-language of adult ethics, they represent the direct developmental foundation of adult morality. Normative ethics does not emerge from a cold, disembodied vacuum; it is the intellectual formalization of primal sentiments that were already active when we were sitting on our parents’ laps watching puppets on a stage.

11.3 Implications for Legal and Moral Responsibility

The realization that infants possess an innate moral sense introduces profound questions regarding the legal and ethical boundaries of human accountability. In contemporary jurisprudence, the concept of mens rea (a guilty mind) is foundational to criminal culpability. Legal systems universally establish a minimum age of criminal responsibility, typically ranging between seven and fourteen years of age depending on the jurisdiction. Children beneath this statutory threshold are classified as doli incapax—incapable of committing a legal crime—based on the historical presumption that they lack the cognitive competence to discern right from wrong.

The Yale research demands a nuanced conceptual refinement of this legal doctrine. Developmental cognitive science demonstrates that we must rigorously separate the status of being a Moral Evaluator from the status of being a Morally Responsible Agent:

• Moral Evaluator: An entity that can observe third-party interactions, comprehend goals, and evaluate an action as socially good or bad. As Bloom and Wynn demonstrated, humans achieve this status within the first six months of life.
• Morally Responsible Agent: An entity that possesses the fully mature executive control, emotional self-regulation, impulse inhibition, and long-term counterfactual foresight required to govern their own actions in accordance with moral principles. This capacity develops extraordinarily slowly, maturing only as the prefrontal cortex fully myelinates in early adulthood.

Thus, while an infant or young child understands that hitting is wrong and actively condemns others who hit, their underdeveloped prefrontal cortex leaves them frequently incapable of stopping themselves from striking a sibling when overwhelmed by a surge of primal frustration. The Yale findings do not justify lowering the age of criminal responsibility; if anything, they confirm why education and rehabilitation must take precedence over retribution when dealing with young minds. Education’s primary duty is not to inject an alien moral sense into an amoral biological vessel, but to empower the emerging executive prefrontal brain to master and overcome the dark, tribal, and impulsive excesses of our innate evolutionary heritage.

12. Future Directions in Developmental Moral Neuroscience and Epistemology

12.1 Neural Correlates of Infant Moral Evaluation

The contemporary frontier of infant moral cognition has transitioned from behavioral reaching trays to the cutting edge of developmental cognitive neuroscience. While behavioral choices provide clear macroscopic evidence of infant preferences, neuroscience seeks to uncover the precise electrophysiological and neurochemical circuitry that executes these sociomoral evaluations in real time.

Two non-invasive neuroimaging technologies have become central to this endeavor:

• Functional Near-Infrared Spectroscopy (fNIRS): Utilizing lightweight, fiber-optic headcaps that emit near-infrared light through the infant’s thin skull, fNIRS measures localized hemodynamic responses (changes in oxygenated and deoxygenated hemoglobin) in the cerebral cortex. Researchers can map which specific cortical regions fire when infants observe Helper versus Hinderer puppet interactions. Preliminary fNIRS investigations reveal that observing antisocial hindering events elicits heightened activation in the medial prefrontal cortex (mPFC) and the temporoparietal junction (TPJ)—regions known in adult neuroscience as the core hubs of Theory of Mind and social evaluation.
• High-Density Event-Related Potentials (ERP): Measuring microvolt fluctuations in brain electrical activity via dense arrays of scalp electroencephalogram (EEG) sensors. Electrophysiological studies have identified distinct neural signatures in infants exposed to prosocial versus antisocial actions. Specifically, observing a Hinderer or an unjust resource distribution evokes an amplified Negative central (Nc) component—an ERP wave associated with involuntary attentional allocation to novel or threatening events—followed by a prolonged Late Positive Potential (LPP), a recognized electrophysiological marker of evaluative emotional processing.

By mapping these neural time-courses with millisecond precision, developmental neuroscientists are constructing a granular, mechanistic blueprint of how the human infant brain transforms optical photons bouncing off a puppet stage into profound neurochemical verdicts of moral praise and condemnation.

12.2 Cross-Cultural Developmental Paradigms

A profound epistemological critique of modern psychology is that it has historically suffered from severe demographic insularity. The vast majority of psychological experiments have been conducted on what Joseph Henrich and colleagues termed WEIRD populations: individuals from societies that are Western, Educated, Industrialized, Rich, and Democratic. When the Yale Baby Lab conducted its landmark studies, its infant participants were drawn almost exclusively from suburban Connecticut and surrounding urban areas—primarily the children of university-educated, middle-class Western families.

To establish that the “moral life of babies” is a true human universal rather than an artifact of Western parenting, modern researchers have expanded the Helper/Hinderer paradigms to radically diverse cultures across the globe. Experimental teams have traveled to traditional, small-scale non-Western societies, including:

• Indigenous hunter-gatherer and horticultural communities in the Amazon Basin.
• Pastoralist communities in rural Kenya and Namibia.
• Rural agricultural villages in Vanuatu and the South Pacific.

These cross-cultural replications have generated profound insights. In small-scale, traditional societies where infants are raised in extensive communal cooperative breeding structures (alloparenting) and have never watched a television screen or entered an academic laboratory, preverbal infants still display an overwhelming, statistically significant reaching preference for the prosocial Helper over the antisocial Hinderer. While minor variations exist in the precise age at which complex retributive reversals occur, the baseline attraction to prosocial assistance and aversion to malevolent obstruction appears invariant across humanity, cementing its status as an evolved biological adaptation of the human species.

12.3 Artificial Intelligence, Robotics, and Evolutionary Modeling

The empirical findings of the Yale Baby Lab have crossed disciplinary boundaries into the bleeding edge of computer science, artificial intelligence, and autonomous robotics. Modern machine learning has mastered narrow technical domains, from chess and Go to natural language processing. However, current AI architectures remain famously deficient in common-sense social cognition. A state-of-the-art Large Language Model can generate an essay analyzing Kant’s categorical imperative, yet it lacks the fundamental sociomoral intuition that an eight-month-old infant uses when deciding which puppet to touch.

To address this gap, AI researchers at institutions such as MIT’s Center for Brains, Minds and Machines are utilizing computational models of infant sociomoral cognition, specifically using Bayesian Theory of Mind (BToM) frameworks:

• Inverse Planning and Cost-Benefit Algorithms: Algorithms are programmed to observe video feeds of moving entities, infer their hidden teleological goals, and compute the physical utility functions (energy expended versus goal attained) of each agent. By treating agents as intuitive utility-maximizers, the AI can independently generate the identical “Helper” and “Hinderer” labels that preverbal infants assign.
• Robotic Social Partner Choice: Autonomous robotics researchers are implementing infant-inspired moral choice algorithms into elder-care and collaborative robots. Before an autonomous robot selects an unfamiliar human or robotic partner to assist in a delicate physical task, it runs an infant-inspired partner-screening program, evaluating historical third-party data to calculate the likelihood of cooperative success.
• Evolutionary Game Theoretic Simulations: Computational biologists are building multi-agent computational simulations where digital organisms evolve over millions of generations. By manipulating parameters such as group competition, resource scarcity, and defector frequency, these simulations independently replicate the exact developmental trajectories observed in human ontogeny: the simultaneous, co-evolutionary emergence of within-group altruism, retributive punishment of cheaters, and out-group parochial hostility.

Through the synthesis of developmental psychology, evolutionary theory, and computational neuroscience, the pioneering work initiated by Karen Wynn and Paul Bloom continues to expand. By gazing into the crib and unraveling the moral life of babies, cognitive science is deciphering not merely the origins of the infant mind, but the deep evolutionary algorithms that govern the past, present, and future of the human species.

Conclusion: Synthesizing the Moral Architecture of the Preverbal Mind

The pioneering investigations conducted by Paul Bloom, Karen Wynn, and their colleagues at the Yale Infant Cognition Center have forever dismantled the classical dogma that human beings enter the world as amoral blank slates. Through the ingenious use of simple puppet dramas, violation-of-expectation metrics, and preferential reaching paradigms, these developmental researchers proved that the preverbal infant possesses a surprisingly rich, sophisticated, and modular moral core. Long before they learn to speak, read, or sit in a classroom, human infants are equipped by natural selection to extract social meaning from raw physical kinematics, distinguish intentional benevolence from malevolence, display an active affection for prosocial Helpers, and avoid antisocial Hinderers.

Moreover, this innate moral architecture is not a crude, two-dimensional reflex. As demonstrated by the Box-Opening, Ball-Dropping, and retributive justice paradigms, the infant mind is capable of navigating hierarchical, multi-agent social networks. Infants expect social actors to affiliate with their benefactors, and they actively endorse justified retribution against wrongdoers, reversing their typical preferences to applaud when an antisocial character is punished. These findings establish that the fundamental roots of desert-based justice, indirect reciprocity, and institutional penal law are hardwired biological adaptations forged by evolutionary pressures to sustain cooperation in our ancestral past.

Yet, Bloom and Wynn’s empirical legacy is profoundly balanced: it resists romanticizing the infant as a flawless angel. The same evolutionary apparatus that gifts the infant with compassion and retributive justice also anchors them in the dark waters of parochial altruism. Preverbal infants spontaneously divide the social universe into in-groups and out-groups based on arbitrary cues, actively favoring those who share their tastes while relishing the mistreatment of those who differ. The moral life of babies is thus an empirical mirror reflecting the fundamental paradox of human nature: an evolved synthesis of breathtaking altruism and fierce tribalism.

Ultimately, these revelations redefine our understanding of education, culture, and ethical progress. Morality is neither an artificial cultural veneer pasted over a savage beast, nor an innate perfection that society corrupts. Instead, human morality is a dynamic, lifelong dialogue between our biological core and our cultural intellect. Infant morality provides the vital, intuitive foundation—the raw emotional bricks of kindness, fairness, and justice. It is the noble responsibility of human culture, education, and reason to take those raw materials, dismantle the innate walls of tribal prejudice that surround them, and expand the circle of moral compassion to embrace all of humanity.

References

• Bloom, P. (2013). Just babies: The origins of good and evil. Crown Publishers.
• Bloom, P. (2016). Against empathy: The case for rational compassion. Ecco/HarperCollins Publishers.
• Carey, S. (2009). The origin of concepts. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780195367638.001.0001
• Haidt, J. (2001). The emotional dog and its rational tail: A social intuitionist approach to moral judgment. Psychological Review, 108(4), 814–834. https://doi.org/10.1037/0033-295X.108.4.814
• Haidt, J. (2012). The righteous mind: Why good people are divided by politics and religion. Pantheon Books.
• Hamlin, J. K., & Wynn, K. (2011). Young infants prefer prosocial to antisocial others. Cognitive Development, 26(1), 30–39. https://doi.org/10.1016/j.cogdev.2010.09.001
• Hamlin, J. K., Wynn, K., & Bloom, P. (2007). Social evaluation by preverbal infants. Nature, 450(7169), 557–559. https://doi.org/10.1038/nature06288
• Hamlin, J. K., Wynn, K., & Bloom, P. (2010). Three-month-olds show a negativity bias in their social evaluations. Developmental Science, 13(6), 923–929. https://doi.org/10.1111/j.1467-7687.2010.00951.x
• Hamlin, J. K., Wynn, K., Bloom, P., & Mahajan, N. (2011). How infants analyze third-party social interactions: The role of intention and valence. Cognitive Science, 35(6), 1184–1204. https://doi.org/10.1111/j.1551-6709.2011.01183.x
• Hamlin, J. K., Mahajan, N., Liberman, Z., & Wynn, K. (2013). Not like me = bad: Infants prefer those who harm dissimilar others. Psychological Science, 24(4), 589–594. https://doi.org/10.1177/0956797612457785
• Hauser, M. (2006). Moral minds: How nature designed our universal sense of right and wrong. William Morrow/HarperCollins Publishers.
• Hobbes, T. (1651). Leviathan or The matter, forme and power of a common wealth ecclesiasticall and civil. Andrew Crooke.
• Hume, D. (1739). A treatise of human nature. John Noon.
• Kohlberg, L. (1981). The philosophy of moral development: Moral stages and the idea of justice. Harper & Row.
• Locke, J. (1689). An essay concerning human understanding. Thomas Basset.
• Mahajan, N., & Wynn, K. (2012). Origins of ‘us’ versus ‘them’: Prelinguistic infants prefer similar others. Cognition, 124(2), 227–233. https://doi.org/10.1016/j.cognition.2012.05.003
• Mikhail, J. (2007). Universal moral grammar: Theory, evidence and the future. Trends in Cognitive Sciences, 11(4), 143–152. https://doi.org/10.1016/j.tics.2006.12.007
• Piaget, J. (1932). The moral judgment of the child. Kegan Paul, Trench, Trubner & Co.
• Rousseau, J.-J. (1755). Discours sur l’origine et les fondemens de l’inégalité parmi les hommes. Marc-Michel Rey.
• Scarf, D., Imuta, K., Colombo, M., & Hayne, H. (2012). Golden rule or golden coin? Rethinking the foundations of infant moral cognition. Developmental Science, 15(6), 751–760. https://doi.org/10.1111/j.1467-7687.2012.01177.x
• Schmidt, M. F., & Sommerville, J. A. (2011). Fairness expectations and altruistic sharing in 15-month-old human infants. PLoS ONE, 6(10), e23223. https://doi.org/10.1371/journal.pone.0023223
• Spelke, E. S., & Kinzler, K. D. (2007). Core knowledge. Developmental Science, 10(1), 89–96. https://doi.org/10.1111/j.1467-7687.2007.00569.x
• Tomasello, M. (2009). Why we cooperate. MIT Press. https://doi.org/10.7551/mitpress/8470.001.0001
• Trivers, R. L. (1971). The evolution of reciprocal altruism. The Quarterly Review of Biology, 46(1), 35–57. https://doi.org/10.1086/406755
• Warneken, F., & Tomasello, M. (2006). Altruistic helping in human infants and young chimpanzees. Science, 311(5765), 1301–1303. https://doi.org/10.1126/science.1121448
• Wynn, K. (1992). Addition and subtraction by human infants. Nature, 358(6389), 749–750. https://doi.org/10.1038/358749a0
• Wynn, K. (2008). Some innate foundations of social and moral cognition. In P. Carruthers, S. Laurence, & S. Stich (Eds.), The innate mind: Volume 3: Foundations and the future (pp. 330–347). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780195332834.003.0019