The effect of language on prosocial behaviors in preschool children

Abstract

The present study investigated how linguistic group membership influences prosocial behaviors, namely helpfulness and cooperation, in preschool children. Whilst research indicates that children preferentially direct their prosocial behavior towards members of their own groups, the influence of perceived linguistic group membership on actual helpfulness and cooperation has not been investigated. We presented an experimenter to 4- and 5-year-olds either as a foreigner, who did not speak the local language or as a native person. Children were then given the opportunity to help or cooperate with this experimenter in a series of nonverbal playful tasks. Whilst 4-year-olds helped and cooperated equally with the foreign and the native experimenter, 5-year-olds required significantly more cues and prompts in order to help or cooperate in the foreign condition. We also found that children were overall more reluctant to respond prosocially in the cooperation tasks than in the helping tasks. We tested children in two European countries (France and Hungary) and found the same pattern of responses in the two locations, suggesting that our findings are not specific to the local culture. Our results extend the findings of earlier research that showed selectivity according to the language spoken by the partner for sharing and imitation. Studies that looked at helpfulness or cooperation used the minimal group paradigm to induce group membership (based on arbitrary cues) and used indirect measures of prosociality, such as different forms of reasoning about the partner. In our study, we used language, a natural cue for group membership (versus arbitrary cues or cues based on social conventions) and directly observed children's helpful and cooperative behaviors toward the experimenter. Our results also confirm previous results indicating that with age, children become selective in their prosocial behaviors as they acquire new means of social evaluation and categorization. We conclude that the language associated with a potential social partner is not only a cue for affiliation and shared knowledge but also a cue mediating children's prosocial acts.

Fig 1. Illustration of the five tasks that were proposed to children to assess their prosocial behaviors.

The Clothespin- and the Two Buttons tasks measured helping and the Elevator-, Trampoline- and Tube tasks measured cooperation. Three of these were adapted versions of tasks designed by Warneken and Tomasello [5], the others were designed for the current study.

Fig 2. Effect of experimenter language and age on prosociality scores in preschoolers.

Experimenter language (Native vs. Foreign) had a significant effect on 4- and 5-year-old preschoolers’ prosociality scores (Wald χ2 = 7.94, df = 1, p = .005), with overall lower scores in the Foreign condition than in the Native condition (U = 16684, p = .001). The interaction between Partner language and Age was also significant (Wald χ2 = 16.10, df = 3, p = .001), as Experimenter language had a significant effect in 5-year-olds but not in 4-year-olds (Wald χ2 = 11.32, df = 1, p = .001; Wald χ2 = 2.05, df = 1, p = .15, respectively). This shows that 5-year-old preschoolers, unlike 4-year-olds, were significantly more reluctant to help or to cooperate when the experimenter was foreign as compared to when she was a native person.

Fig 3. Effect of experimenter language and task type on prosociality scores in preschoolers.

Task type (Helping vs. Cooperation) had a significant effect on preschoolers’ prosociality scores (Wald χ2 = 15.44, df = 1, p = .001), with overall lower scores in the Cooperation than in the Helping tasks (U = 14398, p = .001). We found no interaction between Experimenter language and Task type, indicating that Experimenter language had a similar effect for both Task types.

Fig 4. Frequencies of the four prosocial response categories across experimenter language conditions and age groups.

Experimenter language had a significant effect in 5-year-olds but not in 4-year-olds (Wald χ2 = 11.32, df = 1, p = .001; Wald χ2 = 2.05, df = 1, p = .15, respectively). Five-year-olds’ responses were categorised as ‘no request required, helps or cooperates upon expression of E1’s state of need’ (score 3) significantly less frequently in the Foreign (21%) than in the Native condition (34%, Wald χ2 = 6.76, df = 1, p = .009). Also, they ‘helped or cooperated upon verbal request by E2’ (score 1) more frequently in the Foreign (33%) than in the Native condition (17%, Wald χ2 = 9.07, df = 1, p = .003). These results indicate that 5-year-olds, but not 4-year-olds, were selective in their prosocial behaviors.

Fig 5. Frequencies of the four prosocial response categories in the two task types.

In cooperation tasks, 4- and 5-year-old preschoolers’ responses were significantly more frequently categorised as ‘does not help or cooperate’ (score 0) and ‘requires verbal request by E2’ (score 1) than in helping tasks (p = .004 and p = .02). Children also ‘helped or cooperated upon expression of E1’s state of need’ (score 3) significantly less frequently (p = .001) in cooperation tasks. These results indicate that children were overall more reluctant to respond prosocially in the cooperation tasks than in the helping tasks.

Fig 6. Preschoolers’ mean response latencies across experimenter language conditions and age groups in helping and cooperation tasks combined.

Experimenter language had a significant effect on mean response latencies (F(3, 84) = 10.66, p = .002, partial η² = .12). The effect was stronger at 5 years, when preschoolers showed significantly greater response times in the Foreign condition (M_RLForeign5y = 25.14, SD_RLForeign5y = 6.53) vs. the Native condition (M_RLNative5y = 15.94, SD_RLNative5y = 6.29, p = .004).