Three-year-olds’ spontaneous lying in a novel interaction-based paradigm and its relations to explicit skills and motivational factors

https://doi.org/10.1016/j.jecp.2021.105125Get rights and content

Highlights

  • Employed a novel interaction-based lying paradigm without verbal instruction.

  • Three-year-olds spontaneously informed a friend and misinformed a competitor.

  • Children lied with egocentric and prosocial motivations.

  • They were less proficient to lie in an analogue explicit version.

  • Passing an explicit false belief task was related to spontaneous lying.

Abstract

Previous research has investigated children’s lying and its motivational and social-cognitive correlates mostly through explicit tasks. The current study used an anticipatory interaction-based paradigm adopted from research with preverbal infants. We investigated 3-year-olds’ spontaneous lying within interaction and its motivational basis and relations to explicit skills of lying, false belief understanding, inhibitory control, and socialization. Children interacted with puppets to secure stickers that were hidden in one of two boxes. Either a friend or a competitor puppet tried to obtain the stickers. Nearly all children helpfully provided information about the sticker’s location to the friend, and about half of the sample anticipatorily provided false information to the competitor. Children misinformed the competitor significantly more often than the friend, both when the reward was for themselves and when it was for someone else. Explicitly planning to lie in response to a question occurred significantly less often but predicted spontaneous lying, as did passing the explicit standard false belief task. Thus, by 3 years of age, children spontaneously invoke false beliefs in others. This communicative skill reveals an interactional use of false belief understanding in that it requires holding one’s perspective to pursue one’s goal while providing a different perspective to distract a competitor. Findings support the view that practical theory of mind skills emerge for social coordination and serve as a basis for developing explicit false belief reasoning.

Introduction

Theory of mind (ToM), the ability to predict others’ behaviors by imputing mental states to others (Premack & Woodruff, 1978; see also Flavell, Green, Flavell, & Lin, 1999), enables one to flexibly adjust behaviors in the course of an interaction in order to coordinate meaningfully with others (Dennett, 1978). Mental state attributions are best revealed when they involve false beliefs because the represented content of a false belief can be distinguished from represented reality. It has remained contested, however, whether young children spontaneously represent false beliefs when interacting with other persons (Baillargeon et al., 2010, Bloom and German, 2000, Poulin-Dubois et al., 2018, Wellman et al., 2001).

Standard verbal false belief tasks assess whether children impute mental states by asking them directly (Perner and Roessler, 2012, Perner and Wimmer, 1985), with positive results around 4 or 5 years of age (Wellman et al., 2001). Thus, these tasks measure elicited reflective responses outside of direct interactions but do not tap spontaneous modifications of others’ behavior within interactions. Until relatively recently, comparably less research had adopted Dennett’s (1978) criteria to investigate how children spontaneously modify others’ behaviors within interactive situations by anticipating others’ behaviors based on false belief attributions. Lying is a natural test case of ToM given that it requires the implementation of false beliefs in others. However, as we review below, research on lying heavily rests on verbal task instructions and explicit questions and thus may be less able to capture early spontaneous use of ToM. The current study employed a novel paradigm to measure spontaneous, less explicit skills of lying within social interaction.

Liszkowski and colleagues found that 1-year-olds anticipatorily and flexibly adapt their nonverbal communication within interactions, supporting the interpretation that they anticipate a person’s action based on mental state attributions. For instance, when an adult was about to retrieve an object but was mistaken about its location, 1-year-olds spontaneously provided true information; they intervened helpfully before the adult would commit the mistake. In control conditions, when the adult knew about the location or did not intend to retrieve the object, infants intervened significantly less (Knudsen and Liszkowski, 2012a, Knudsen and Liszkowski, 2012b, Knudsen and Liszkowski, 2013). Subsequent studies have suggested that 1.5-year-olds (Buttelmann et al., 2009, Knudsen and Liszkowski, 2012a, Southgate et al., 2010) and 3-year-olds (Király, Oláh, Csibra, & Kovács, 2018) will indeed make appropriate inferences about a person’s action, even when the person holds a false belief, and react on the basis of the person’s belief, not reality. These interaction-based paradigms then provide insights into an early-emerging use of ToM within interaction before explicit tasks are mastered (Liszkowski, 2013).

However, on empirical grounds, several findings on infants’ implicit ToM skills have been found to be difficult to reproduce (Crivello and Poulin-Dubois, 2018, Dörrenberg et al., 2018, Dörrenberg et al., 2019, Poulin-Dubois et al., 2018, Priewasser et al., 2018). Crucially, on conceptual grounds, they are amenable to a leaner interpretation; infants may act in these tasks by tracking an adult’s knowledge state and determining the adult’s goal (they understand ignorance but not false belief; Tomasello, 2018). Thus, infants may understand whether another person does or does not share a perspective with them but might not understand whether a person has a different perspective (Liszkowski, 2018); that is, they do not represent two conflicting perspectives at the same time (Moll and Meltzoff, 2011, Perner, 1991).

Lying is a much clearer case of ToM use in interaction. When lying, one does not want to share one’s perspective with the other person but instead wants to provide a different (nonfactual) perspective to the other person. Thus, one intends to implant a nonfactual perspective in a recipient (i.e., a “false” belief) in order to make the recipient behave in an anticipated way so that it does not impede pursuing one’s own goal (based on one’s own perspective). Most research converges to show that children begin to lie around the same age when they pass the verbal standard false belief task, that is, around 4 years of age (Talwar & Crossman, 2011). However, just like verbal standard false belief tasks, evidence on lying rests heavily on paradigms that use verbal task instructions, use elicited responses, and often require verbal lies. As with verbal standard false belief tasks, thus, these methods may be less apt to reveal spontaneous, less explicit skills of lying within social interaction. To date, paradigms for assessing spontaneous lying analogous to interaction-based tasks of spontaneous informing are still lacking.

Distinct paradigms have been used to measure lying behavior. In transgression paradigms, children are verbally instructed not to peek at a desirable object in the experimenter’s absence. When being asked upon the experimenter’s return, they typically deny their transgression (they answer “no”; Evans & Lee, 2013). In the disappointing gift task (Talwar and Crossman, 2011, Talwar and Lee, 2002, Talwar et al., 2007), children receive an undesirable gift, such as a bar of white soap. When being asked by the gift giver whether they like the gift, they typically deny their disappointment (they answer “yes”). Findings reveal an age-related increase of these so-called antisocial and prosocial lies, respectively, with most children lying not before 4 years of age. Antisocial lies seem to appear earlier than prosocial lies (Talwar and Crossman, 2011, Talwar et al., 2017), perhaps reflecting socialization toward socially accepted behavior (Lavoie, Yachison, Crossman, & Talwar, 2017). Although compatible with a false belief interpretation, it is also quite possible that performance in these tasks rests on a pragmatic understanding that one should not transgress social norms and conventions. Children may attempt to undo, or withhold, a socially unfavorable perspective rather than provide a specific novel nonfactual perspective (i.e., implant a false belief). Thus, they may rather conceal factual information than provide false information, which is compatible with the interpretation that younger children’s ToM involves an understanding of ignorance, not false belief (Tomasello, 2018).

Another common way of measuring lying in children is the hide-and-seek paradigm. In this paradigm, the child is typically involved in a hiding process and then explicitly is asked, or sometimes even instructed, to mislead another agent who is searching for the hidden object. With this paradigm, initial research suggested deceptive abilities in some 2- and 3-year-olds (Chandler, Fritz, & Hala, 1989), although subsequent research revealed that once appropriately controlled and compared across ages, most children have difficulties with lying in such tasks before 4 years of age (Bigelow and Dugas, 2009, Hayashi, 2017, Mascaro et al., 2016, Sodian, 1991, Sodian et al., 1991). Interestingly, younger children are able to physically prevent the competitor from obtaining the object (Peskin, 1992, Sodian, 1991); that is, they do understand the situation and act appropriately. Yet, they seem to be deficient at communicating a different perspective to the competitor. Similarly, Carlson, Moses, and Hix (1998) found that children under 4 years of age had difficulties with deceptively pointing to a false location when instructed to trick an opponent. Hala and Russell (2001) found that 3-year-olds had difficulties with employing a strategy of deceptive pointing even after a series of differentially rewarding feedback in which participants would lose a reward if they did not point deceptively. Although these studies showed that children did not lie communicatively when being explicitly instructed to do so, both studies found that children were able to use associative strategies of placing markers that would make competitors search in the marked (but false) location. Similarly, Harvey, Davoodi, and Blake (2018) found that 5-year-olds would circle a location on a map where they wanted a thief to search (falsely) for a target object. Although the response measure was nonverbal, children were verbally instructed and questioned, and they provided false information only when the hypothetical narrative verbally emphasized harm and transgression. Taken together, findings in the respective explicit tasks suggest that children younger than 4 or 5 years do not spontaneously communicate differing perspectives to implant false beliefs even when they are able to hinder a competitor from winning. Recently, Ding and colleagues (Ding, Heyman, Fu, et al., 2018, Ding et al., 2015) found that 2-year-olds can discover deceptive strategies when playing the same game over a 10-day period even when they initially do not know how to deceive. In this task, children’s performance was scaffolded by the experimenter and based on instructional feedback. It is less clear whether the trained competence in the task yielded a transfer such that children would spontaneously lie to others in novel situations.

Taken together, prior research has relied on paradigms that require verbal instructions or vignettes as well as explicit questions in order to elicit lying in children. There is less experimental research on a spontaneous anticipatory use of lying within an interactional flow. Although spontaneous informing paradigms have been employed with much younger infants, there is a gap in analogous knowledge about children’s spontaneous misinforming, that is, lying. Furthermore, several conventional lying paradigms are amenable to an interpretation of perspective withholding (corresponding to ignorance understanding) rather than provision of a different nonfactual perspective (corresponding to false belief understanding). In the current study, therefore, we developed an interactive lying paradigm and tested whether young children intentionally and spontaneously provide false information as indicated by unelicited, spontaneous nonverbal lying.

Our general paradigm was based on the structure of previous informative pointing paradigms (Knudsen & Liszkowski, 2012b), which refrain from behavioral instructions, hypothetical narratives, and explicit questions and allow even preverbal infants to spontaneously interact and provide information by pointing to an object’s location in anticipation of an adult’s goal. Our paradigm made use of puppets because protest paradigms have shown that children are not afraid to correct and stand up against competitor puppets (Rakoczy et al., 2008, Schmidt et al., 2013). Furthermore, Hala and Russell (2001) found that children are less prone to lie in the presence of an adult authority figure. In our interaction paradigm, children interacted with a protagonist puppet while either a competitor puppet or a friend puppet (relative to both children and the protagonist puppet) was looking for a hidden sticker. Whereas the competitor puppet aimed at stealing the sticker from children or the protagonist puppet (depending on the motivation condition; see below), the friend puppet aimed at cleaning or providing the sticker, respectively. Children knew where the sticker was hidden and could spontaneously provide truthful or false information to the puppets, withhold information, or hinder the puppets physically.

Based on previous findings with 1-year-olds (Knudsen & Liszkowski, 2012b), we expected that children would understand the friend puppet’s goal and spontaneously inform the friend puppet more than misinform her about the sticker’s location to help her find it. That is, we expected that children would inform the friend puppet already when it approached the scene on an arbitrary anticipation path, before it asked about the sticker’s location and before it was evident from its behavior where it would look for the sticker. For the competitor puppet, we reasoned that, if children had a practical implicit ToM understanding, including the understanding of false beliefs and different perspectives, they should spontaneously provide false information to the competitor puppet, again in the anticipation phase before seeing the puppet approach and steal the sticker. Thus, children should misinform the competitor more than the friend. Alternatively, it could be that children operate with an understanding of ignorance (shared/not shared perspective). In that case, children would not be able to lie. However, they should withhold information about the sticker’s location more often in the competitor condition compared with the friend condition; that is, they should inform the competitor less often than the friend. Finally, children could also have a less epistemic understanding of the situation and simply understand the competitor’s goal. In that case, they should physically hinder the competitor to achieve his goal more so than the friend puppet. The null assumption was that children would not differentiate between the two conditions at all and would point equally often to the stickers, perhaps out of interest or in an imperative manner to obtain these.

In addition, we manipulated the motivational context of children’s behaviors in the competitor and friend conditions. To this end, children in the egocentric motivation conditions could obtain the stickers for themselves (to collect them in a sticker book), whereas children in the prosocial conditions never obtained the stickers but could help the protagonist puppet obtain the stickers and collect them for herself. For the friend condition, this manipulation was not central because previous research had already shown that infants inform an ignorant person both egocentrically for their own benefit to get a toy (O’Neill, 1996) and prosocially for the benefit of the person to help her find something (Knudsen and Liszkowski, 2012b, Liszkowski et al., 2008). Misinforming, however, may vary as a function of motivational context, and thus the motivational manipulation was central to the competitor conditions. Egocentric lies are often the first ones observed by parents (Newton et al., 2000, Talwar et al., 2017), which could suggest that initially young children spontaneously lie for egocentric rather than prosocial purposes. In that case, we would expect young children to misinform the competitor more than the friend only in the egocentric motivation condition. On the other hand, young infants readily help others to achieve their goals in various situations such as by retrieving out-of-reach objects and opening a door for an adult when the adult’s hands are full (Warneken & Tomasello, 2006) or by providing missing information (Liszkowski et al., 2006). Based on this natural tendency to help others, we reasoned that young children could also help someone by misinforming someone else so that a friend profits from the lie. For example, Harvey et al. (2018) found that 5-year-olds lied instrumentally to prevent a moral transgression on behalf of a third person. On this view, thus, children should misinform the competitor more than the friend also in the prosocial condition when only the protagonist, but not the children, will benefit from the lie by receiving stickers.

Furthermore, to better understand the spontaneous, practical, interaction-based response measure, we investigated it in relation to explicit verbally stated plans of action. To this end, in a subset of trials we elicited future hypothetical verbal responses to a verbal question about what children thought to do when the puppet appeared. That is, we asked them before they could spontaneously intervene. According to Sodian (1991), children younger than 4 years have difficulties with verbally explicating a lying strategy, although they are able to physically prevent a competitor from attaining his goal. Similarly, Rhodes and Brandone (2014) found that children spontaneously interacted appropriately based on anticipating a person’s behavior, but struggled when they needed to explicitly state what the person would do next. Therefore, we expected that children have difficulties with talking about their lying, although they might nevertheless be able to spontaneously lie in the interaction paradigm. Such a dissociation would support two-system accounts that suggest a distinction between implicit and explicit ToM skills (Apperly and Butterfill, 2009, Low et al., 2016).

Finally, we collected several correlational measures to relate children’s spontaneous misinforming to established measures of explicit cognitive ToM processes and inhibitory control. Regarding ToM processes, we used a verbal knowledge–ignorance task from a standard ToM scale and an explicit standard false belief task (Hofer and Aschersleben, 2007, Perner and Wimmer, 1985, Wellman and Liu, 2004). As reviewed before, one proposition is that lying in the sense of denying rule violations requires a less complex ToM that pertains to understanding whether a person is knowledgeable or ignorant (Leduc et al., 2017, Ma et al., 2015). In contrast, intentional lying in the sense of providing a false perspective requires the ability to implant false beliefs in others (Bigelow and Dugas, 2009, Talwar et al., 2007, Williams et al., 2016) and has been related to explicit false belief understanding (Ding, Heyman, Sai, et al., 2018, Ma et al., 2015). If performance on a standard false belief task would correlate with spontaneous anticipatory misinforming in the interaction paradigm, this would support the interpretation that lying involves representing conflicting perspectives.

Regarding inhibitory control, we used the bear–dragon task (Kochanska et al., 1996, Reed et al., 1984). Lying has been suggested to require inhibiting the truth in order to provide false information, and executive functions—in particular inhibitory control—have been related to the development of lying in several studies (Carlson and Moses, 2001, Carlson et al., 1998, Ding, Heyman, Sai, et al., 2018, Evans et al., 2011, Hala and Russell, 2001, Talwar et al., 2017, Talwar and Lee, 2008). Furthermore, executive functions have been related to a number of conceptual ToM skills, including children’s ability to distinguish reality from fantasy (Davoodi, Corriveau, & Harris, 2016). If the bear–dragon task correlated with spontaneous lying in the current paradigm, this would support the interpretation that lying depends not just on conceptual advances in ToM but also on domain-general processes more generally. Alternatively, spontaneous lying may be different from instructed, elicited explicit lying and may be unrelated to inhibitory skills. For example, spontaneously providing a false perspective may depend less on inhibiting one’s own perspective because after all, pursuing one’s own perspective is often the very reason to belie someone else.

Finally, spontaneous lying may be a function of social interactional experiences. Because it is difficult to obtain direct observational data, we used a parent questionnaire (the German extended version of the Alabama Parenting Questionnaire (GEAPQ-P-ES; Frick, 1991, Reichle and Franiek, 2009), which asked parents about their parenting styles. In addition, we administered an informal exploratory questionnaire about children’s frequency of lying in daily life. Previous research has suggested a positive relation between children’s development of antisocial lying and authoritarian parenting (Baumrind, 1971), that is, exposure to a harsh and physical disciplinary parental style (Stouthamer-Loeber, 1986, Talwar and Lee, 2011). Furthermore, control parenting (e.g., criticism, behavioral control) and firm but responsive parenting (i.e. authoritative parenting) (Baumrind, 1971) are associated with a lower propensity to lie (Ma et al., 2015, Talwar et al., 2017). In the current study, we continued to explore possible relations between parenting behavior and spontaneous lying. If social-interactional experiences, reflected in the form of parenting style, were related to the spontaneous use of lying, we expected positive relations between spontaneous lying and authoritarian parenting and expected negative relations between spontaneous lying and positive responsible parenting.

Section snippets

Participants

The sample consisted of 3-year-old Caucasian children (mean age = 42.48 months, SD = 3.68) who were recruited from the department’s database (n = 48) of parents who had agreed to participate in child studies and from day-care institutions (n = 34) in a metropolitan city in Germany. Participants were tested in the research lab and day-care centers, respectively, balanced across conditions. No socioeconomic status (SES) measures were obtained, but the estimated level was middle to high SES, and

Communication and physical intervention

Fig. 2 presents the mean proportion of trials with the different behaviors in the competitor and friend conditions. A 3 (Behavior: informing, misinforming, or hindering) × 2 (Condition: competitor or friend) × 2 (Motivation: egocentric or prosocial) ANCOVA with age in months as a covariate revealed a significant interaction between condition and behavior, F(2, 77) = 22.98, p < .001, η2 = .23, and among the three factors, F(2, 77) = 4.05, p = .019, η2 = .05. Age had no effect as a covariate.

Discussion

The 3-year-olds in the current study spontaneously communicated information to others that was either true or false. The paradigm was analogous to an interaction-based paradigm, which had shown that younger infants communicate spontaneously truthful information in anticipation of helping others (Knudsen and Liszkowski, 2012a, Knudsen and Liszkowski, 2012b). Of interest for the current study was whether children also communicate spontaneously false information to someone in order to provide that

Acknowledgments

This research was partially funded by the German Research Foundation (Deutsche Forschungsgemeinschaft) (Project: FOR 2253; Grants: LI 1989/3-2). We thank all parents and their children, as well as the day-care institutions, for participating in our study. We thank all students who helped in recruiting participants and thank Charlotte Mannstein for coding our data. We also thank Leon Li for his helpful comments on a previous version of the manuscript.

References (76)

  • J. Perner et al.

    From infants’ to children’s appreciation of belief

    Trends in Cognitive Sciences

    (2012)
  • J. Perner et al.

    “John thinks that Mary thinks that”: Attribution of second-order beliefs by 5- to 10-year-old children

    Journal of Experimental Child Psychology

    (1985)
  • D. Poulin-Dubois et al.

    Do infants understand false beliefs? We don’t know yet—A commentary on Baillargeon, Buttelmann and Southgate’s commentary

    Cognitive Development

    (2018)
  • B. Priewasser et al.

    Helping as an early indicator of a theory of mind: Mentalism or teleology?

    Cognitive Development

    (2018)
  • P. Rubio-Fernández

    Perspective tracking in progress: Do not disturb

    Cognition

    (2013)
  • M.F.H. Schmidt et al.

    Young children understand and defend the entitlements of others

    Journal of Experimental Child Psychology

    (2013)
  • M. Stouthamer-Loeber

    Lying as a problem behavior in children: A review

    Clinical Psychology Review

    (1986)
  • V. Talwar et al.

    From little white lies to filthy liars: The evolution of honesty and deception in young children

  • V. Talwar et al.

    The role of executive functioning and theory of mind in children’s lies for another and for themselves

    Early Childhood Research Quarterly

    (2017)
  • V. Talwar et al.

    Influence of social factors on the relation between lie-telling and children’s cognitive abilities

    Journal of Experimental Child Psychology

    (2017)
  • S. Williams et al.

    The role of executive functions and theory of mind in children’s prosocial lie-telling

    Journal of Experimental Child Psychology

    (2016)
  • H. Wimmer et al.

    Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children’s understanding of deception

    Cognition

    (1983)
  • I.A. Apperly et al.

    Do humans have two systems to track beliefs and belief-like states?

    Psychological Review

    (2009)
  • D. Baumrind

    Current patterns of parental authority

    Developmental Psychology

    (1971)
  • A.E. Bigelow et al.

    Relations among preschool children’s understanding of visual perspective taking, false belief, and lying

    Journal of Cognition and Development

    (2009)
  • P. Bloom et al.

    Two reasons to abandon the false belief task as a test of theory of mind

    Cognition

    (2000)
  • S.M. Carlson et al.

    Individual differences in inhibitory control and children’s theory of mind

    Child Development

    (2001)
  • S.M. Carlson et al.

    The role of inhibitory processes in young children’s difficulties with deception and false belief

    Child Development

    (1998)
  • M. Chandler et al.

    Small-scale deceit: Deception as a marker of two-, three-, and four-year-olds’ early theories of mind

    Child Development

    (1989)
  • T. Davoodi et al.

    Distinguishing between realistic and fantastical figures in Iran

    Developmental Psychology

    (2016)
  • D. Dennett

    Beliefs about beliefs

    Behavioral and Brain Sciences

    (1978)
  • X.P. Ding et al.

    Young children discover how to deceive in 10 days: A microgenetic study

    Developmental Science

    (2018)
  • X.P. Ding et al.

    Theory-of-mind training causes honest young children to lie

    Psychological Science

    (2015)
  • A.D. Evans et al.

    Emergence of lying in very young children

    Developmental Psychology

    (2013)
  • A.D. Evans et al.

    When all signs point to you: Lies told in the face of evidence

    Developmental Psychology

    (2011)
  • J.H. Flavell et al.

    Development of children’s knowledge about unconsciousness

    Child Development

    (1999)
  • Frick, P. J. (1991). The Alabama Parenting Questionnaire. Unpublished Instrument. Alabama: University of Alabama,...
  • H. Hayashi

    Young children’s difficulty with deception in a conflict situation

    International Journal of Behavioral Development

    (2017)
  • Cited by (0)

    View full text