Development

Developmental Adaptations of Decision-Making
Taming Uncertainty Across Development
Primate and Human Decision-Making in the Face of Uncertainty and Risks
The Development of Deliberate Ignorance
The Struggle of Life Among Words: Cognitive Mechanisms Behind Language Change

Developmental Adaptations of Decision-Making

Anyone who has tried a skateboard trick and gotten injured knows that experience and learning shape how people interact with the world in the future. ARC's investigations into the mechanisms underlying the "long reach of childhood"—the durable influence of early experiences on later cognition, behavior, and health—suggest that the influence of early experiences is rooted in learning about and adapting to the statistics of the environment. Three lines of research at ARC illustrate this idea: how the statistical structure of risk shapes adolescent behavior, how learning and exploration across the lifespan resemble an optimization process, and how socioeconomic environments create distinct decision ecologies.

Adolescents and the Statistical Structure of Risk

Ciranka and Hertwig (2023) proposed a learning framework for analyzing how individuals' risk judgments and preferences change with repeated risk-taking over developmental timescales (Figure 1). This approach draws attention to the statistical structure of many of the risks adolescents typically take and, by analyzing actuarial data and various longitudinal household surveys, shows that negative experiences across many risk domains (e.g., in substance use, unprotected sex or reckless driving) are initially unlikely. Because these ostensibly safe experiences teach adolescents that their risky behavior is safe, risky behaviors can persist into adulthood. Rather than simply labeling adolescents as reckless, viewing their behavior through the lens of adaptive learning draws attention to the structure of the environment and calls for a greater understanding of individual learning experiences within a person's developmental trajectory.

Learning and Exploration Across the Lifespan

Giron et al. (2023) demonstrated that the developmental progression of feedback learning and exploration can be described as an optimization process. Examining learning and search in participants aged 5 to 55 years in a reward-learning experiment showed that across development, generalization and exploration become tuned toward greater efficiency. Computational modeling and simulations revealed that developmental differences in parameters governing learning resembled the trajectory of a near-optimal algorithm as it explored solutions in this experiment. This work empirically clarifies previously vague metaphors of artificial intelligence and optimization applied to human development. It raises the question of how different environments, for instance with more or less predictable rewards, influence the results of this developmental optimization process.

Socioeconomic Environments

People with different socioeconomic statuses grow up in environments that differ in probability and quality of rewards and risks they offer. As a consequence, socioeconomic differences in preferences and decision-strategies emerge. Ciranka et al. (2026) compared how people with different socioeconomic statuses used a risk–reward heuristic that exploits the knowledge that high rewards are unlikely. People with lower socioeconomic status used the heuristic more strongly, implying that they assume high rewards are less likely and low rewards are more likely than people with higher socioeconomic status. Simulating the foraging behavior of individuals who search for rewards in an uncertain world revealed that increasing scarcity and competition yield a stronger relationship between reward magnitude and the probability of foraging success, which aligns with the socioeconomic gradient in the strength of the risk-reward heuristic. This suggests that socioeconomic differences in decision heuristics may reflect concrete adaptations to the reward statistics people encounter when trying to make ends meet under different conditions of scarcity. 

In sum, an adaptive learning framework reveals how exposure to different statistical worlds gives rise to distinct decision strategies and behaviors, tracing the developmental roots of psychological differences among people. 

Key References 

Taming Uncertainty Across Development

Will children succeed in coercing their parents to buy candy if they throw a tantrum in the supermarket? Will adolescents live better lives if they quit school and start a band? These prospects are uncertain. Most decisions occur in a world of uncertainty, yet the development of strategies to cope with uncertainty is generally overlooked and understudied.

Children Explore and Adults Maximize in Repeated Decisions From Description

Thoma and Schulze (2025) asked children aged 3 to 7 years and adults aged 18 years and older to predict outcomes of various guessing games, such as the color of a randomly drawn ball from a bingo machine, where outcome probabilities were visually presented. Standard theory suggests that people diversify their choices according to success probabilities in repeated choice. Children's and adults' choices, however, followed different patterns: Adults consistently bet on the option with the highest probability of success, whereas children's choices were more explorative. Neither group's choices aligned with theoretical predictions, which may hold only in repeated choices with feedback. These results underscore the need for greater attention to the probabilistic structure of children's decision environments when examining how decision-making develops.

Adolescents, Uncertainty, and Social Information

Ciranka and van den Bos (2025) investigated the development of another strategy that can facilitate decision-making under uncertainty: the use of social information. Bayes's rule implies that the more uncertain a person is, the more they should update their beliefs in response to new information. This study tested whether social influence across adolescence (ages 12 to 24 years) follows these Bayesian principles. By manipulating uncertainty about lottery outcomes and the presence of social information favoring betting, and applying Bayes' theorem as a computational model (Figure 2), the study demonstrated that adolescents were more susceptible to social influence when they were more uncertain. This finding shifts the view of adolescence from a period of gullibility to one of uncertainty, and shows that what often distinguishes adolescents from adults is their greater uncertainty about what to do.

By formalizing the processes behind decision-making under uncertainty, these studies demonstrate the developmental diversity and appropriateness of children's and adolescents' decision strategies.

Key References

Primate and Human Decision-Making in the Face of Uncertainty and Risks

Risk preferences contribute to many of the decisions people make, including whether to look for a new mate, to compete for status, or to invest money. Risk preferences can influence the entire course of an individual's life and may have wide-ranging consequences for society. However, their evolutionary origins remain poorly understood. 

The Evolutionary Roots of Risk Preference

Haux et al. (2023) employed a comparative developmental approach to systematically investigate individual differences and average age trends in the propensity to take risks among chimpanzees, humans' closest evolutionary relatives. Through observer reports and experimentally investigating their decision-making, the study revealed that chimpanzee and human risk preferences share key similarities. First, chimpanzees' willingness to take risks manifests as a trait-like preference, consistent across domains and measures. Second, chimpanzees are ambiguity averse. Third, male chimpanzees are more risk-prone than females. And fourth, chimpanzees' appetite for risk shows an inverted-U-shaped relation to age that peaks during young adulthood (Figure 3). This work suggests that key features of risk preference may have evolutionary roots and emerge independently of human culture.

Key Reference

The Evolutionary Roots of Exploration

The evolutionary and developmental roots of human exploration—the process of gathering information to reduce uncertainty about the world—remain largely unknown. To investigate the evolutionary roots of human exploration, Haux et al. (2025) adapted the human decisions-from-experience paradigm, researching exploration under uncertainty in chimpanzees. Chimpanzees (N = 15) were simultaneously confronted with an uncertain option (with outcome variance) and a safe option (without outcome variance) and tested in both stable and changing environments. Again, there were striking similarities between primate and human exploration, particularly in their responses to environmental change. Like humans, chimpanzees explored more across trials in changing than in stable conditions. Chimpanzees also explored more when they experience variance in option outcomes. 

Taken together, these two studies on risk preferences and exploratory strategies indicate that humans and primates respond to environmental properties in similar ways, suggesting that core features of decision-making under uncertainty have deep evolutionary roots.

Key References

The Development of Deliberate Ignorance

Is there anything a child does not want to know? Caregivers of young children and many researchers may be inclined to say no, and for good reason: Children have little experience with the rules that govern the world, and they need more information to figure them out. At some point in development, however, human curiosity becomes more nuanced and may even reverse. Many adults come to prefer not knowing certain things (Hertwig et al., 2025; Offer et al., 2025), from the mundane, like a movie spoiler, to the deeply consequential, such as a genetic risk for a life-threatening illness. Choosing not to know information that is easy to obtain and potentially carries high utility is a phenomenon called deliberate ignorance. 

Kerbl et al. (2025) suggest that focusing on deliberate ignorance is essential to understanding the development of information search. Studying what children choose not to know offers a way of tracing the developmental shift from a young child's seemingly unquenchable thirst for knowledge to a growing awareness that knowledge can also carry psychological costs. Deliberate ignorance opens a window onto key developmental questions, such as how children regulate their emotions, how experiences shape adolescents' interactions with online worlds, or even how parents attempt to curate their children's information ecology (Figure 4). 

In conceptualizing the developmental origins of deliberate ignorance, ARC researchers treat deliberate ignorance as a skill whose use depends on cognitive building blocks. Each of these blocks develops on a different timescale, and together they enable a child to curate their information diet. Thus, deliberate ignorance offers a paradigm for studying how children learn to exercise agency over what they know—and what they choose not to know.

Key References

The Struggle of Life Among Words: Cognitive Mechanisms Behind Language Change

In the evolution of language, why do some words and grammatical forms endure while others fade away? Li et al. (2024) investigated how cognitive selection mechanisms operating at the individual level scale up to long-term language change. Combining experimental methods with large-scale historical corpus analysis, they identified which psycholinguistic properties (e.g., emotional valence and concreteness) systematically predict which words gain the upper hand in linguistic competition.

Study 1 captured micro-level selection pressures in language production and memory under ecologically valid communicative conditions by using a well-known serial reproduction paradigm. Participants read emotionally themed narratives and retold them across three generations of transmission. The authors modeled word preservation while controlling for frequency, word length, grammatical category, repetition, and individual and story-level variance. Earlier age of acquisition, higher concreteness, greater emotional arousal, and emotional extremity were found to reliably increase a word's likelihood of survival during interpersonal transmission, whereas emotional valence (positive vs. negative) did not.

Study 2 extended these findings to diachronic language change using two historical corpora (Corpus of Historical American English and two subcorpora from the Google Books Ngram corpus) spanning the period of 1800–2000. Again, age of acquisition, concreteness, and arousal predicted long-term increases in word frequency, even after controlling for properties such as baseline frequency (Figure 5). Divergences from Study 1—specifically the advantages of shorter and positively valenced words—are likely attributable to broader sociocultural shifts, underscoring the interaction between cognitive constraints and historical context.

Together, these findings support a view of language as a cognitively shaped adaptive system. Words that are learned earlier, easier to imagine, and emotionally engaging are more likely to survive. Language change thus reflects the constraints and preferences of the human mind, magnified over time.

Key Reference