Children’s strategic theory of mind

Abstract

Human strategic interaction requires reasoning about other people's behavior and mental states, combined with an understanding of their incentives. However, the ontogenic development of strategic reasoning is not well understood: At what age do we show a capacity for sophisticated play in social interactions? Several lines of inquiry suggest an important role for recursive thinking (RT) and theory of mind (ToM), but these capacities leave out the strategic element. We posit a strategic theory of mind (SToM) integrating ToM and RT with reasoning about incentives of all players. We investigated SToM in 3-to 9-y-old children and adults in two games that represent prevalent aspects of social interaction. Children anticipate deceptive and competitive moves from the other player and play both games in a strategically sophisticated manner by 7 y of age. One game has a pure strategy Nash equilibrium: In this game, children achieve equilibrium play by the age of 7 y on the first move. In the other game, with a single mixedstrategy equilibrium, children's behavior moved toward the equilibrium with experience. These two results also correspond to two ways in which children's behavior resembles adult behavior in the same games. In both games, children's behavior becomes more strategically sophisticated with age on the first move. Beyond the age of 7 y, children begin to think about strategic interaction not myopically, but in a farsighted way, possibly with a view to cooperating and capitalizing on mutual gains in long-run relationships.practical reasoning | child development | game theory S trategic environments determine outcomes as a function of the decision of many players. Behavior in these environments is directed by a capacity we term strategic theory of mind (SToM), the capacity to infer other people's mental processes and predict their behavior on the basis of knowledge of their incentives and assumption of their rationality. SToM requires two more primitive capacities. The first is "ordinary" theory of mind (ToM), which is a person's ability to "impute mental states to himself and others" (1). The second is recursive thinking (RT), which is the ability to use the output of one step of a reasoning process as input to a following step. In addition, SToM requires that agents reason about the incentives of all involved.To take an example of the combination of ToM, RT, and reasoning about incentives that will be relevant in the context of this study, suppose that Ann and Bob play a game with rules known to both players, where the rules imply that Ann has an incentive to lie to Bob when Bob believes her. Ann can conclude that she should lie to Bob. Similarly, Bob can conclude that Ann has an incentive to lie to him, and hence he will not believe her. However, in the same manner, Ann concludes that Bob will perform exactly the same reasoning, so that Bob will be skeptical in view of Ann's incentive to lie. Ann will use this output of her reasoning about Bob's incentives as input for her next step of reasoning, and conclu...

Cortical thickness of the dorsolateral prefrontal cortex predicts strategic choices in economic games

“…Although these findings are inconsistent with the current lack of effect of strategic reasoning in the UG, the inconsistency may be due to age differences in the earlier study population (8,50,51) versus the current study. Cognitive control of normative demands in the DG may require the strategic use or goal-directed mobilization of theory of mind (52). Very young children, before acquiring cognitive theory of mind, do not make fair offers in either the UG or the DG (50, 51).…”

Delay of gratification and the role of defaults—An experiment with kindergarten children

“…In the kindergarten years, many important behavioral and economically relevant traits such as pro-sociality and strategic reasoning evolve (Fehr et al, 2008, Sher et al, 2014, Brosig-Koch et al, 2015, which makes this period of life particularly interesting to study whether time preferences -a key trait -are malleable through simple interventions.…”

Cognitive Skills and the Development of Strategic Sophistication

“…Economists and psychologists have begun to investigate how age influences strategic behavior in competitive games (Perner, 1979;Sher et al, 2014;Brosig-Koch et al, 2015;Czermak et al, 2016;. 5 However, we are not aware of any existing papers that investigate how psychometric measures of cognitive ability and theory-of-mind predict the strategic behavior of children in competitive games, although a handful of papers study related questions (Steinbeis et al, 2012;Sher et al, 2014;Geng et al, 2015;Czermak et al, 2016). 6 Nor are we aware of existing work that studies how children respond to information about the cognitive ability (or theory-of-mind) of their opponent in strategic environments.…”

“…The games were not played against another child, but against an experimenter who was not trying to win, and instead followed a non-equilibrium behavioral rule. Sher et al (2014) do not measure cognitive ability (in our sense of abstract or logical reasoning ability), but they do measure working memory, finding that better working memory predicts choices closer to equilibrium in the stickers game but not in the sender-receiver game. Geng et al (2015) study the behavior of adolescents rather than children, and find that their subjects do not play minimax in zero-sum games; since opponents do not play minimax, it is perhaps not surprising that they also find no correlation between cognitive ability and closeness to minimax play for adolescents (nor do they find a correlation with earnings).…”

“…The theory builds on the framework of psychological games introduced byGeanakoplos et al (1989). 5Perner (1979) finds that the ability to detect an opponent's dominant strategy improves with age Sher et al (2014). find that behavior tends to move toward equilibrium with age.…”