What two executive functions are most called on when children engage in perspective taking?

From a social-cultural perspective (Vygotsky, 1986), children’s PS emerges spontaneously from social interaction with others where speech is used to guide children’s thinking and behavior during joint activities (Berk & Spuhl, 1995; Winsler et al., 1997). By definition, children (and adults for that matter: Alarcón-Rubio, Sánchez-Medina, & Winsler, 2013) use PS when it is needed, when the task at hand is sufficiently challenging that it cannot be done on (silent) automatic pilot (Winsler et al., 2009). Investigators consistently find that children use PS in moderately challenging tasks and situations, and not so much when tasks are particularly easy (Fernyhough & Fradley, 2005; Lidstone, Meins, & Fernyhough, 2011; Winsler & Diaz, 1995; Winsler, De León, Wallace, Carlton, & Willson-Quayle, 2003; Winsler, Naglieri, & Manfra, 2006). This makes understanding relations between PS, task performance, and EF particularly challenging for researchers, because amount of PS use on a given task depends on task difficulty, which itself depends on the task, the child’s current level of ability and EF, and (often, but not always) child age (Winsler et al., 2009).

Two types of developmental change are observed in PS – ontogenetic and microgenetic. On the ontogenetic level, researchers find a peak in the probability of observing spontaneous PS at around 4 years of age (Berk, 1986; Winsler et al., 2003). After this point, PS often becomes more quiet and partially internalized for children. Berk’s (1986) widely used coding system for PS breaks speech into three levels of internalization. Children first use Level-1 PS, which is wordplay, noises, comments about persons/things not present and statements irrelevant to the task. They then use more Level-2 as overt, task-relevant, self-guiding speech, followed later by Level-3 speech (partially internalized) which consists of whispered or silent, verbal lip movements (Berk, 1986). On the micro-genetic level, as children gain mastery over individual tasks across repeated trials, children change in how much PS they use depending on where they are in task mastery from moment to moment, regardless of age (Winsler & Naglieri, 2003; Winsler et al., 1997). Persons at any age will often use overt PS if the task is difficult enough and strains their current capacity (Duncan & Cheyne, 2001; Winsler et al., 2006).

The above discussion has been about children internalizing language in the form of PS for EF purposes. According to social-cultural theory (Vygotsky, 1986), children also internalize other social-cultural tools, norms, and symbol systems, such as music, to become part of a child’s internal mental world. Scholars have speculated that the internal mental world also includes internalized shared music, dance, and movement (e.g., drumming), and singing, and that evolutionarily, these came before full oral language (Casas-Mas et al., 2018; Hagen & Hammerstein, 2009; Zachariou & Whitebread, 2015). There are empirical links between music and cognitive/EF development. Exposure to music programs in school or private instrumental lessons causally relates to improved cognition and EF in children as revealed by randomized control trials (Moreno et al., 2011; Schellenberg, 2006; Schellenberg, Corrigall, Dys, & Malti, 2015). Dance and music/movement programs for preschoolers appear to have causal effects on enhanced social and behavior regulation skills (Lobo & Winsler, 2006). Part of these causal effects could be mediated by inner musicality or private song. Parents sing to children to regulate child behavior as a young infant (i.e., “Rock-a-bye-baby”) before they talk to them in full speech (Trevarthen, 2008). Parents and preschool teachers often use songs (i.e. “The Clean-up Song”) to regulate child behavior. It is certainly possible that children use music as a social-cultural tool in the form of private singing or private humming for fun, for emotion regulation, and as a tool for self-regulation during problem solving.

Researchers, however, have just started to examine the private singing and humming of children. Casas-Mas et al. (2018) investigated a few cases of private singing behaviors in novice and experienced musicians. The researchers qualitatively coded private singing (task relevant and irrelevant), sing-song noises, humming, and breathing while playing instruments during music instruction. The prevalence of private singing was low, but present in a young musician (age 7) engaging in task-related behaviors (e.g., finding notes on the instrument, positioning the instrument relative to the body). In the laboratory, Winsler et al. (2011) examined PS use and EF in a quasi-experimental study of young children who were and were not (naturally) enrolled in an early childhood music and movement program that involved many activities designed for behavioral regulation via music (i.e., stop-go, fast-slow, high-low). The researchers observed the PS that program attenders and non-attenders used on a switching/selective attention task (SAT) and a delay of gratification task. Children who received the music program used more Level-2 (task-relevant overt) PS on the SAT, and were able to wait longer for an experimenter to return in the gift-delay task. Notably, children attending the program were observed to sing and hum to themselves in strategic ways during the gift-delay task more than children not enrolled.

The PS literature has not yet systematically examined private singing and, thus, an exploratory study such as the current work is helpful. Indeed, commonly used PS categorization systems are not even clear on where private singing or humming should go. In Berk’s (1986) system, singing to the self is considered a subtype of Level-1 task-irrelevant overt PS assumed to be immature. This might make sense if children sung to themselves in a way that appears unrelated to the task at hand (e.g., “Puff the Magic Dragon”). However, it is possible that children’s private singing during EF tasks could be task-relevant (e.g., “Where is the green one? – in a sing-song manner to the tune of “Pop Goes the Weasel”). Indeed, Mead and Winsler (2019) showed that a music-exposed 45-month-old (Nora) used a significant portion of crib/PS to the self at naptime with much of it consisting of singing and humming. Some PS was song unrelated to her current activity, some of it was a known tune with words creatively task-relevant (e.g., “Where is the green one?”), and some song consisted of novel tunes yet the words were task-relevant (similar to Berk’s Level-2 PS) and related to the play she was engaged in with her stuffed animals (Mead & Winsler, 2019). Nora’s sung utterances were longer and more complex than her spoken utterances, suggesting the need to distinguish between PS song and speech duration. There may be important relations between bouts of private singing, PS utterances, and EF performance, but there has not been research relating these components together. We examine such issues here in the first paper on this topic to our knowledge.

In this exploratory study, we examine the PS and private song used by 4- to 8-year-olds and relate their speech/song use to EF skills measured with a battery of assessments as well as by parent report. Following prior research (Fatzer & Roebers, 2012; Lidstone et al., 2011; Winsler et al., 2000), we assess multiple facets in EF including simple inhibition, working memory, cognitive flexibility, and higher-order planning. We raise the following questions and hypotheses: 1) To what extent do children privately speak, sing, and hum to themselves during the task? It was expected that most children would use at least some PS to perform a selective attention/shifting task (2006, Winsler et al., 1997). PS was expected to be largely covert in nature, due to the age of our sample, but some would also be overt due to the difficulty of the task. We expected to see at least some singing and humming, but thought that this behavior would be rare. We were curious to see if the private singing observed would be task relevant (Level-2) or off task (Level-1), but did not have a priori hypotheses. 2) What is the relation of PS to private singing and humming? Based on Berk (1986), we hypothesized private singing and private humming would be observed more among children who also engaged in more overt speech. 3) Is PS and private song related to EF performance? We thought overt private, speech, singing, and humming would be related to other EF indicators (inhibitory control, cognitive flexibility, planning, parent-report). However, given the problems discussed above about inferring speech-performance relations from one-point-in-time global correlations, we remained agnostic about the direction of the links between Level-2 overt task relevant PS and song with EF performance.