Development of Conceptual Reasoning versus Understanding of Children’s Theory of Mind and Extraordinary (Supernatural) Minds during Middle Childhood

Round 1

Reviewer 1 Report

The aim of this article was to exmaine when a significant distinction appears, the conceptual differentiation of the ordinary mind from the extraordinary mind.

The study has some strong points (clear introduction, accurate statistical analysis, constructive discussion).

 However, I would like to ask the authors to address some points in order to improve the paper:

1) The structure should be: Intorduction, Method, Results, Discussion.

2) Can you describe the concept of religiosity more thoroughly in Introduction.

3) What are the limitations of the study?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

The study that is presented in this article makes a useful contribution to the literature surrounding the formation of children’s religious concepts. It builds upon previous studies focused on theory of mind development and whether children’s understanding of superhuman minds is best captured by the anthropomorphism hypothesis. A sample from Poland and the analyses in relation to conceptual reasoning of children are valuable additions to this area. Nonetheless, the study will have much greater impact if the presentation of the findings are improved in a number of ways. 

First, I recommend introducing the study through attention to the real-world problem of conceptualizing unusual, nonhuman intentional beings instead of by way of cognitive developmental theory. The readership of Religions will be more interested in the lived religious experience of children and their families than various stage theories and the like in child developmental psychology. This theoretical framing is especially unnecessary given how little contemporary psychology, let alone cognitive approaches to the study of religious cognition, attend to these Piagetian frameworks. I appreciate that the focus on domain-general Piagetian theory helps motivate the use of the conceptual reasoning scale, but given that it proved to be fairly unimportant in understanding the results, the degree of attention to this theory seems unnecessary here. I suspect a more effective way of motivating the study would be to observe that children face what looks like an interesting conceptual challenge in thinking about the mind of God (and other superhumans); it seems reasonable to assume that they reason analogically from human minds (i.e., anthropomorphize); but it turns out that even reasoning about other human minds is no trivial task. Previous focus has been largely on domain-specific cognitive systems (theory of mind), but perhaps in addition to theory of mind maturation, such reflection on the mental properties and states of superhumans is aided by more general conceptual strengths. Etc. 

Second, it would be helpful if some of the measures and hypotheses are at least mentioned in some summary fashion before the results are presented. Perhaps the choice of order (e.g. placing methods toward the end) was driven by the requirements of the journal. Nonetheless, the introduction should still take us to the point of understanding the general contours of the study and presently it doesn’t do that.

Third, more use of descriptive statistics would be helpful. That is, it would be good to know how children at various ages tended to answer on the various measures because this information is critical for interpreting the findings. 

Fourth, because this is an interdisciplinary journal that includes a readership that does not necessarily have statistical expertise, I recommend being clearer about what is being reported in the results section. For instance, repeatedly a ‘dependency’ is being reported between two variables, but it isn’t clear to me (someone with quantitative methods experience) what this means. Is this meant to say that there is an association between the two variables? A (partial) correlation? I note that the word ‘dependency’ could be misconstrued to mean a certain type of causal connection (e.g., a deterministic one). 

Finally, the interpretation of the results concerning the anthropomorphism hypothesis doesn’t obviously follow from the results as presented. Indeed, just focusing on the false belief task, it appears that children at age three and at age seven are responding similarly for Superman and for God, with very irregular curves in between. These curves do not tightly correspond to the Mum or Girl curves and so it is hard to see why this is supportive of anthropomorphism. Indeed, at age 3, children appear to be treating all of the agents as approximately the same AND as likely to have true beliefs. Though the God curve is irregular, concerning Superman (at least), children take longer to become accurate in thinking about Mum and Girl than Superman. And yet, the manuscript suggests (pp. 7-8) that these extraordinary minds are more complex for children than human minds. The data presented do not seem to support this conclusion. Indeed, the results presented for the false belief task look much more aligned with the preparedness hypothesis.

I also have some relatively minor suggestions. (1) For instance, the manuscript repeatedly refers to the children in this study as being in ‘middle childhood’, but 3- to 5-year-olds (at least) are commonly referred to being in early childhood in the developmental psychology literature with which I am familiar. (2) It would be helpful to indicate more clearly what the religious knowledge scale amounted to and why it is important here. (3) I have a minor concern that showing children an image of God may have made them think more anthropomorphically about God than they might have done otherwise. Perhaps this methodological decision led to noisier results concerning God and increased the demands on children to conceptually disentangle their concept of God from the more anthropomorphic concept that the protocol triggered. Such possibilities should be considering in interpreting the results. (4) The presentation of the ‘preparedness’ hypothesis isn’t entirely clear in the introduction and at least some elements are, technically, in error. My understanding is that the authors who have written in support of ‘preparedness’ do not rely on innateness or no role at all for the environment to shape development, but only that the theory of mind system tends to default to over-estimating the knowledge of others, to extend to whatever it is that children know themselves could be known. This is different than ‘omniscience’. Some more recent writing in this space than the Barrett and Richert pieces in the early 2000s may be clearer on these points.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

The study claims to support the anthropomorphic hypothesis about the development of the ability to think about extraordinary minds. First, we need to gain competence in understanding ordinary minds to conceptualize the extraordinary mind. In particular, researchers examine when a conceptual distinction between ordinary and extraordinary minds appears in development. Additionally, they want to assess whether the distance is related to understanding religious concepts. For this purpose, they use three methods: The Deceptive Box Task, The Conceptual Reasoning Scale from IDS-P and a test of understanding religious concepts and compare their results.

My comments mainly refer to the presentation of the study area in such a way that it motivates the approach to the problem and the choice of methods.

Firstly, an important general point: a precise reference to the field of cognitive developmental psychology and, in particular, to the subfield of Theory of Mind (ToM) is missing in the introduction of the paper. And this is striking because it is assumed there is a connection between the development of the concept of God and mindreading.

In the paper, the anthropomorphic hypothesis presented as its first version (Piaget 1929) did not refer to the most recent formulations. The most recent versions postulate a relationship between the concept of extraordinary minds (e.g., God) and egocentrism or reality bias, the observed performance of pre-representational children in false beliefs tasks. Mentioning this would partly remedy the general issue referred to previously. And also suggest why the false beliefs tasks (such as the Deceptive Box Task) are chosen for the study. The election of the method is not understood based only on Piaget's version of the anthropomorphic concept of an extraordinary mind. In other words, the introduction needs to make explicit the connection between TOM development and the study of the acquisition of the concept of extraordinary minds (e.g., God), which is not adequately explicated in lines 66-70.

Secondly, when presenting the two main hypotheses, preparedness and anthropomorphism, there is some ambiguity about whether the idea of God can be considered intuitive in each approach. It is characteristic of the preparation hypothesis that the idea of God is intuitive. Precisely, it is postulated that it is an intuitive concept because children seem to understand it early (before understanding ordinary minds) by virtue of their cognitive equipment that allows them to attribute by default traits that are characteristic of the monotheistic conception of God, such as omniscience and almightiness (e.g., Barrett & Richert 2003). Accordingly, it is argued "Children show ‘preparedness’ to represent extraordinary minds and it is essential that these intuitive ideas are preserved until adulthood (Barrett and Richert 2003; Richert and Barrett 2005)." (Lines 60,61)

When presenting the anthropomorphic view, the authors claim that the idea of God is also intuitive (line 50). This affirmation seems to be related to the Piagetian claim about the influence of religious education in having an early concept of God (lines 47-49). However, in the anthropomorphic most recent version hypothesis, the idea of an extraordinary mind seems not to be 'intuitive' or attributed by default but just the opposite. The concept of an extraordinary mind (God) can be attributed once the characteristics of ordinary minds are understood. For this reason, it is confusing to characterize the idea of God as intuitive within the framework of the anthropomorphic hypothesis.

Additionally, the actual use of the word 'intuitive' is also confusing. In association with the anthropomorphic hypothesis, the concept is intuitive to the extent that it is promoted by the environment (e.g., religious education). In association with the preparedness hypothesis, due to the cognitive equipment. One way to solve this is to avoid using the word intuitive. However, if you want to use it, select a criterion to establish that something is intuitive. In my opinion, it is more appropriate the association with innate equipment, than with the influence of the environment.

Thirdly, it results notably the claiming that there has been no attempt to connect TOM with conceptual reasoning (Line 71-72). TOM development is considered the development of a conceptual capacity. Once we understand the mind as representational, we can understand and correctly use the mental state concepts such as beliefs, desires, and intentions. Acquiring a representational theory of the mind (or TOM) implies understanding that people behave based on the representation of the world but not on the state of affairs (Welllman 1990). In sum, acquiring a representational theory of the mind is considered a conceptual achievement.

In particular, the scientific child approach postulates that the development of TOM depends on general theorizing abilities, which are used in childhood to form and change the theory underlying mentalistic ability, and in adulthood to form and change scientific theories (Gopnik & Meltzoff 1997). The development of TOM involves the internalization of intentional psychology according to which behavior is caused by beliefs and desires. A theory analogous to scientific theories underlies mentalist understanding insofar as ordinary mental concepts are theoretical constructs, legally related to one another in ways that allow for causal explanations and facilitate predictions (Wellman 1990, Gopnik & Wellman 1992, Gopnik & Meltzoff 1997). Now, to arrive at this fully developed theory, changes occur in children's mental understanding that can be considered similar to those that take place in scientific knowledge (Gopnik & Wellman 1992, Gopnik & Meltzoff 1997). Children are considered as little scientists who build and review their thinking in relation to various domains of knowledge (physics, biology, psychology).

In principle, the development of TOM and the development of conceptual reasoning are continuous, and this would imply moderating some statements, for example:

The development of conceptual reasoning skills has also been linked to the development of the attribution of false beliefs. 261-262

(This link was made before)

Another novelty in the field of research on the capacity of conceptual reasoning was its connection with the attribution of false beliefs. 277- 278

It was found that along with an increase of conceptual knowledge, the knowledge about minds also grows. 280- 281

(TOM is conceptual knowledge and TOM sophistication is conceptual sophistication or conceptual growth)

Observation of simultaneous growth 283 may suggest that these abilities develop at the same time. 284

(To the extent that TOM is a conceptual ability, TOM will have the same timetable in development as conceptual abilities in general)

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Some of the revisions from the previous version helpfully address weaknesses in the manuscript. It appears, however, that in the attempt to make revisions, various new errors have been made. Throughout there are missing words, punctuation, spaces, and occasionally some construction weaknesses that distract from the main message of the manuscript.

More substantively, there are a number of problems that are still present from the original manuscript that really need to be addressed.

Previously I suggested a stronger framing that connects the work to lived religious thought in children. I fear the author(s) misunderstood my comment. What I meant to suggest that it would strengthen the manuscript to help readers -- readers who may not be psychologists, let alone developmental psychologists -- to understand better why it is that this study matters to them. What does it add to our understanding of religious cognition? Introducing the psychological approach with some kind of more contextual, religion-relevant framing could really help the study demonstrate its usefulness.

Second, concerning the place of conceptual reasoning, it would be helpful further clarify what has and has not been suggested by the results. For example, it is claimed that conceptual reasoning is correlated with scores on "understanding of religious concepts" but exactly what that understanding amounts to isn't clear in the manuscript. It sounds as if it just means that children know who God is. If that is right, it appears that children who are stronger conceptually are also better at understanding who God is; and may, in turn, perform better on the false-belief task. All of that is reasonable, but do we have reason to think that such relationships are of particular interest? One interpretation would be that children who are poorer in conceptual skills are less clear on who this character "God" is, are more easily confused by the drawn picture of God into confusing God with an ordinary person, and so answer more as if God is a human being. That isn't uninteresting, but it isn't the same as claiming that conceptual understanding is playing a large role in the underlying developmental process. Indeed, the reported regression analysis suggests that the measure of conceptual reasoning does not significantly contribute to predicting strong performance on the false-belief task for God when child's age and their performance concerning Mum are included in the model. This suggests that successful performance on the false belief task (with Mum) is a much better predictor. One way to understand that is that task-related competence (e.g., Theory of Mind competence? Understanding the task demands?) leads to better performance for God on the task, quite aside from conceptual reasoning. (This picture might be clearer if another regression were run predicting Mum from God, child's age, etc.)

Third, given the new text on lines 71-76, it is surprising that the relative uniformity of responses across agents (below 50%) for the three year-old children isn't given more attention in the discussion. See also the claim in the abstract that the data "confirm" the anthropomorphism hypothesis. This conclusion is far from obvious and needs considerably more support.

Fourth, and relatedly, the description of the preparedness hypothesis in the introduction has not been improved and it is still in error. For instance, preparedness does not require any claims about innateness, omniscience, or thought about other minds as being completely sectioned off from cultural influences. The claim, as I understand it, is that young children find it easier to think of others' minds (whomever's mind it is), as being super-knowing and/or super-perceiving; they default to the assumption that if something is knowable (from their perspective), the other knows it. (This overlaps with an egocentric bias, but seems to be slightly different as studies by Barrett, Richert, and others have shown that children in states of ignorance will sometimes assume God will know the contents of a container.) And so, under preparedness, children have to learn that humans and other agents have perceptual and conceptual limitations. For this reason, it is easier for children to reason with gods or super beings who have relatively few limitations -- it is closer to the developmental primitive state (or default assumptions). The data presented here seem to align reasonably well with such an account. And so, even if the author(s) doesn't set out to adjudicate between preparedness and anthropomorphism, the author(s) owes it to readers to present the accounts accurately and to explain how the data do and don't align with the predictions of the two accounts.

Other important matters that require attention:

- On lines 389-390, the word "not" seems to be missing.

-  On Table 1, the p-value of .000 appears twice. Technically, this should be <.001 or <.0001, because a p-value cannot (strictly) = 0. I recognize that stats packages give such values as outputs but they shouldn't be reported as .000. This would mean that the probability of the results given the null hypothesis is true in the population is zero or utterly impossible. Strictly speaking, it couldn't be entirely impossible.

- Perhaps something is lost in translation, but I assure you (and have double-checked multiple authoritative sources) that ages 3-6-years old are standardly considered early childhood. Before that children are infants or toddlers. The early school years are considered 'middle childhood'. 

I hope these suggestions are helpful. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 3

Reviewer 3 Report

The manuscript contains a number of typographical errors, some leading to ambiguities of meaning. (For instance, in the abstract the term 'anthropomorphic hypothesis' is used instead of 'anthropomorphism hypothesis'.) 

This passage contains one unsupported claim and two ambiguous ones:

"The performed logistic regression analysis demonstrates that only two variables have an impact on assigning knowledge to the figure of God. These are the age of the child and the knowledge of the ordinary mind (assigning mental abilities to Mum). The ability of children to attribute omniscience to God is significantly linked to their age and competence in the attribution of the ordinary mind. Considering our research group (which is primarily Christian children), it can be said that our research is consistent with previous findings that in the first developmental phase, children think of God as having human-like mental states and limitations (Richert et al. 2017; Kiessling and Perner 2014; Lane et al. 2010, 2012; Makris and Pnevmatikos 2007). The research has shown that conceptual knowledge is important to ordinary and extraordinary minds, but children do not need any more conceptual knowledge to be able to understand an extraordinary mind." 

First, the analysis mentioned here does not show that "in the first developmental phase, children think of God as having human-like mental states and limitations." Rather, it shows that three-year-old children do not think of any agent as having ordinary, human-like mental states and limitations. Four-year-olds seem uncertain with the exception of Superman who does not have human-like mental states and limitations.

This claim is ambiguous: "The research has shown that conceptual knowledge is important to ordinary and extraordinary minds, but children do not need any more conceptual knowledge to be able to understand an extraordinary mind." More precisely, the study has suggested that there is some relationship between passing the false-belief test and scoring higher on the 'conceptual reasoning' test, but the study has not ruled out that this relationship is anything other than children with greater conceptual reasoning skills better understanding the task demands. Furthermore, it could be that the causal relationship goes the other way: children with better social intelligence (including ability to pass the false-belief task) have acquired stronger conceptual reasoning. And so, perhaps it isn't that 'conceptual knowledge' is particularly important.

An additional ambiguity appears here: "The ability of children to attribute omniscience to God is significantly linked to their age and competence in the attribution of the ordinary mind." First, strictly speaking, the children haven't been asked whether God is omniscient, only whether God knows the contents of a box. Second, as you haven't presented a regression analysis that treats attribution of limitations to the mother as the dependent variable and 'passing' the task concerning God as a predictor variable, it very well could be that you have found that distinguishing between God and mother mutually inform each other, or that understanding God drives better understanding of mother. 

The analysis still neglects to give clear descriptive data concerning whether children are more accurate in reasoning about the various super agents versus the human agents. What are the means and counts for each age group? Are any significantly below or above 50%-chance answering? It looks like the mental states of Superman, for instance, are reasoned about more accurately than the girl or mother. If so, that would count against the following claim: "The lack of differences between age groups in terms of responses to these figures indicates a different treatment of ordinary and extraordinary minds. The knowledge of extraordinary minds is more difficult to master." It appears that Superman was easier to master. That is, the data presented seem to contradict -- not support -- this conclusion.

The results of this study make a valuable contribution but some elements of the interpretation do not align well with the results and should be revised so as not to misrepresent the findings.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf