Running Head: TRANSMISSION OF BELIEFS IN THE UNOBSERVABLE
Beliefs about unobservable scientific and religious entities are transmitted via subtle
linguistic cues in parental testimony
Niamh McLoughlin
1,2
, Ciara Jacob
1,3
, Petal Samrow
1
, & Kathleen H. Corriveau
1
*
1
Boston University,
2
University of Kent,
3
University of Bath
Acknowledgements: We thank the families and schools who participated in this research. We
also thank Jennifer M. Clegg, Yixin Kelly Cui, Telli Davoodi, Kathryn Leech, Ayse Payir, and
Wani Qiu for their help with collecting and coding the data and Paul L. Harris for valuable
comments on earlier drafts. This research was supported by a Large Grant #59820 from the John
Templeton Foundation awarded to KHC.
Manuscript accepted for publication. Journal of Cognition & Development
Corresponding author information:
Kathleen H. Corriveau
Wheelock College of Education and Human Development
Boston University
621 Commonwealth Avenue
Boston, 02215
Email: kcorriv@bu.edu
TRANSMISSION OF BELIEFS IN THE UNOBSERVABLE
2
Abstract
We explored the role of parental testimony in children’s developing beliefs about the ontological
status of typically unobservable phenomena. US parents and their 5- to 7-year-old children (N =
25 dyads) separately rated their confidence in the existence of scientific and religious
unobservable entities (e.g., germs, angels), and were invited to engage in an unmoderated dyadic
conversation about the entities. Both parents and children were more confident in the existence
of the scientific entities compared to the religious entities. Parental religiosity predicted the
strength of their belief in the religious entities, and these beliefs were positively associated with
their children’s judgements in the domain of religion. We coded parental testimony produced
during the unmoderated conversation for a number of subtle linguistic cues that convey their
confidence and prevailing beliefs in an entity’s existence. The results revealed consistent cross-
domain differences: parents expressed more uncertainty, were more likely to mention variation
in people’s beliefs and make explicit claims about the ontological status of the religious, as
compared to the scientific entities. However, with increasing religiosity, parents produced fewer
cues to uncertainty, mentioned belief variation less often, and were more likely to make claims
of endorsement when talking about the religious unobservables. Importantly, the pattern of
linguistic cues in parental testimony was significantly associated with children’s ontological
judgements. The present findings have implications for understanding the socio-cultural
mechanisms by which confidence in the existence of invisible agents and processes develops in
childhood.
Keywords: Concepts, parent-child interaction, social cognition
TRANSMISSION OF BELIEFS IN THE UNOBSERVABLE
3
Beliefs about unobservable scientific and religious entities are transmitted via subtle
linguistic cues in parental testimony
Although we are not able to directly experience or encounter many everyday causal
phenomena, beliefs in the existence of these phenomena can have a powerful influence on our
behavior. For example, in the months succeeding the declaration of the global COVID-19
pandemic by the World Health Organization (WHO, 2020), psychologists have sought to
understand the perceptions and motivations that lead to commitments to public health practices
that can halt the spread of the invisible, yet highly infectious, virus (see Van Bavel et al., 2020,
for a review). Similarly, belief in the causal efficacy of both scientific and supernatural remedies,
such as vaccines and prayer, influences the behaviors that adults and children engage in to
protect themselves (Legare, Evans, Rosengren, & Harris, 2012; Rutjens & Preston, 2020). An
open question is just how beliefs about typically unobservable entities develop in individuals
across the lifespan.
In the current study, we focus on the mechanisms through which confidence in the
existence of unobservables emerges in childhood. Two hypotheses seem plausible. On the one
hand, constructivist theories of cognitive development indicate that belief formation is heavily
influenced by children’s own experience with the world (Gopnik & Meltzoff, 1997; Gopnik &
Wellman, 2012; Piaget, 1957). On this hypothesis, because unobservable entities are by
definition not visible, children should ignore or even doubt their existence. On the other hand, a
second body of research proposes that young children learn about unobservable agents and
processes through the prevalent adult testimony in their community (Harris & Corriveau, 2014;
Harris & Koenig, 2006). On this hypothesis, subtle variation in discussions about different
unobservables influences the development of children’s conceptualization of these entities.
TRANSMISSION OF BELIEFS IN THE UNOBSERVABLE
4
To date, research on the core proposition of the second hypothesis has focused on the
strength of beliefs in the unobservable across different cultures, as well as the quality of
children’s epistemic justifications. One striking finding is the similarity in the types of
explanations that both adults and children provide when justifying the existence of endorsed
phenomena – namely scientific entities such as germs, as well as religious figures such as God,
and special beings such as Santa Claus (Harris, Pasquini, Duke, Asscher, & Pons, 2006;
Shtulman, 2013). For example, US adults and children frequently refer to the causal powers of
unobservable scientific and supernatural agents. Despite these cross-domain similarities in
patterns of justification, children profess more confidence in the existence of scientific as
compared to supernatural unobservables (among other endorsed entities), in most of the cultural
contexts studied thus far (Clegg, Cui, Harris, & Corriveau, 2019; Cui et al., 2019; Davoodi et al.,
2018; Davoodi et al., 2020; Harris & Corriveau, 2020; Harris et al., 2006; Guerrero, Enesco, &
Harris, 2010). One possible explanation for the differential level of confidence is that there is
important linguistic variation in the testimony surrounding each type of entity. The present
research explores this possibility by investigating the pattern of linguistic cues – that convey
confidence and prevailing belief in an entity’s existence - in adult testimony when discussing
endorsed unobservable entities in the domains of science and religion with young children.
Critically, we investigated the relation between variation in such linguistic markers and
children’s belief in the existence of the entities.
Children often have access to multiple sources of information in their environment, but
parents are likely to be an influential, familiar and even preferred source (Harris, Koenig,
Corriveau, & Jaswal, 2018; McLoughlin et al., in press), especially in early childhood (Corriveau
et al., 2009). Recent studies conducted in Mainland China highlight the importance of the home
TRANSMISSION OF BELIEFS IN THE UNOBSERVABLE
5
environment for children’s developing beliefs about unobservable phenomena. Cui et al. (2019)
showed that parents and their children (both 5- to 6-year-olds, and 9- to 11-year-olds) who
adopted the majority secular view of this culture were skeptical about the existence of religious
concepts. By contrast, Christian parents and their children (a minority group in China) were
confident that religious entities exist. Moreover, there was a positive association between the
ontological religious judgements of the parents and their children. Interestingly, this relation
continued to hold among the older children even though they had more extensive exposure to the
secular state curriculum. Indeed, Davoodi et al. (2020) found that children belonging to the
minority religious group in China frequently appealed to specific testimonial sources in their
immediate social circle when justifying their religious beliefs.
In a pioneering study of parental input, Canfield and Ganea (2014) identified subtle
variations in the language that US parents used to discuss endorsed special beings (e.g., Santa
Claus, Tooth Fairy, God) compared to entities in the domains of science and history with their 3-
to 5-year-olds. Parents were more likely to communicate doubt (e.g., use modulations such as “I
think” and “I believe”), acknowledge a lack of consensus in belief more generally (e.g., “Some
people think that God…”), and refer to the imaginary nature of the entity when talking about the
endorsed beings as compared to the scientific entities. These findings suggest that certain
linguistic cues are used more frequently to convey information about the ontological status of
unobservable religious entities as compared to scientific unobservables. However, Canfield and
Ganea (2014) did not explore the relation between parent and child beliefs, and, importantly,
whether the observed variation in adult testimony for the different type of entities was related to
children’s judgements (most likely in part because the families in this study were only asked to
discuss one of four potential topics in each domain).
