Violations of Expectation in Infants

Author’s Note 

Violations of expectations, the topic of this paper, came from a course I took at UC Davis titled, PSC 154 The Psychology of Emotions. While reading the textbook for this course, Understanding Emotions (Keltner et al., 2018), I came across the concept of surprise as a means of facilitating learning behaviors in infants. As a Human Development major and someone who is interested in child development, I wanted to explore the literature of this phenomenon and for my UWP 102 Writing in Psychology and Human Development I was able to dedicate my time to compiling this literature review in hopes of disseminating information on infant learning as well as researching the importance of surprise as a behavioral trait and mechanism for learning for future studies in human development. 

Abstract

Human development researchers agree that surprises challenge core knowledge, preconceived notions about how the environment works, and lead infants to explore events and objects that do not fit within expectations. Violation of expectation (VOE) has been shown across studies to operate as a learning mechanism in infants. “Observing the Unexpected Enhances Infants’ Learning and Exploration” (Stahl & Feigenson, 2015) was the framework for VOE research and provided key research terms such as “violation of expectations” and “core knowledge” that I expanded on after consultation with a UC Davis librarian. VOE studies were organized and analyzed by the type of learning mechanism that was measured in the experiment. Three types of learning mechanisms have been delineated from VOE: kinesthetic learning behaviors, visual learning behaviors, and cognitive learning behaviors. Visual learning behaviors begin with selective attention which precedes exploration. Attention to environment cues promotes explanation seeking curiosity, leading an infant to display social referencing expressions to a nearby caretaker and providing a learning opportunity to explore VOE situations.  Kinesthetic learning is motor-based cognitive processing and was operationalized as approaching and playing behaviors such as banging, dropping, and grasping objects that violate expectations. Cognitive learning behaviors that have been observed in infants, including repeated exposure to a VOE event, increases familiarization and prediction patterns due to statistical learning. VOE does have a limited effect on learning mechanisms as there is not a significant effect of VOE situations on memory enhancement. Measures of cognitive learning behaviors and the validity of infant emotional learning studies have been criticized as having low validity of attention measurements. Three changes to increase validity of infant VOE studies are: measure looking away times as well as looking at time, provide a theoretical framework for parameters, and to include pilot test data used to determine the parameters of the study.

Infants are tasked with an exorbitant amount of learning at the beginning of their lives regarding their surroundings and caretakers. Infants interact with a variety of people, each have their own unique expressions of emotions.Violation of expectations can prompt learning behaviors. Stimulating learning behaviors in children helps them reach developmental psychosocial and biological milestones that will be built upon during later learning (Ramos, 2018). Violation of expectations is prompted by first-person cues in the environment which can foster cognitive development in infancy.  First-person cues grab the attention of infants as they prefer to explore stimuli that violate their expectations. Third-person cues come from others in the form of social referencing and guided questions that facilitate learning. Future-oriented cues involve the integration of newly-learned information into an infant’s understanding of the world. Liquin and Lombrozo (2020) determined that explanation-seeking curiosity (ESC) was the motivation for learning how or why an event occurs and is prompted by events that violate expectations. ESC is selective and based on attention to cues in the environment. These cues are important to know and study as they promote learning behaviors in infants, the developmental stage when the most learning occurs. 

These learning behaviors are measurements used to study VOE in infants because they are overtly observable and can be tested for understanding through experimentation. Directing infant attention to learning opportunities can be done by exposing them to surprising situations. Surprise can be provoked through a multitude of scenarios such as the violation of expectations as it defies preconceived notions about how the world operates. Infants prefer objects that violate expectation and engage with them more than novel objects which are new stimuli that align with previously established cognitive frameworks. Infants often learn through social referencing, the act of looking to others for cues on how to react. These learning mechanisms are experimental measurements of behavior such as visual learning behaviors, kinesthetic learning behaviors, and cognitive learning behaviors.

Methods

“Observing the Unexpected Enhances Infants’ Learning and Exploration” (Stahl & Feigenson, 2015) in a Psychology of Emotions class textbook was the primary article into further research. Observing the Unexpected was about violation of expectation (VOE) and the preference for VOE objects over novel objects. Since this article was published in Science, further research was continued through studies recommended by Science that would expand knowledge on violation of expectation and the emotion of surprise being used in affective learning that were cited in Observing the Unexpected recommended by Science. PsycInfo, Wiley Library, and Pubmed primarily provided access to the foundation articles. Articles that were included in the study were within the years 2015 - present to increase the reliability and accuracy of the content analyzed. Some search terms used were: surprise, infants, VOE, learning behaviors, selective attention. 

To further research on VOE, a librarian specialist in psychological research continued to collect articles from the UC Davis library databases. Six more articles were discovered but some were dissertations or reviews that were only to be used as reference points. Others added key information to the subtopics of learning mechanisms such as cognitive learning, which is incredibly difficult to reliably measure. The articles examined are relevant because they build on existing theories and add validity to previous research. The sources analyzed in the results section explore VOE and how to measure learning behaviors observed in infants.

All of the sources provided evidence that infants favor VOE experiences over novel ones. Sources were organized by their measurements as well as their operational definitions of learning mechanisms. Information was placed into a research matrix and measurements were sorted into three categories: visual learning behavior, kinesthetic learning behaviors, and cognitive learning behaviors. Visual learning behaviors were operationalized as studies contained measurements that focused on attention or social learning and were analyzed first because the infant must be focused on the VOE event before action and cognitive learning can occur. Kinesthetic learning behavior was defined as behaviors that were intentional in position and movement of sensory organs in an attempt to explain the VOE event. Articles that were organized into this category were playing and approaching behaviors. Cognitive learning behaviors were defined as studies that measured learning and integration of knowledge into schemas. Memory, familiarization, and statistical learning were cognitive learning behaviors because their study designs included retesting and demonstration of understanding. All selected sources strengthen the claim that VOE objects are preferred over novel ones and can facilitate learning behaviors. 

Results 

Studies examining infant learning indicate that violations of expectations facilitate learning behaviors in infants to encourage them to learn about the nuances in the ways the world operates. Selective attention is the beginning step in the process of exploration as infants stare longer at VOE situations as well as display surprise reactions. Infants’ attention can be directed to learning opportunities by their caretakers to further develop cognitive skills. Behaviors such as kinesthetic learning behaviors observed in infants are shown to build upon an infant’s core knowledge. Infants have shown a preference for VOE objects over novel objects because they do not align with core knowledge and therefore require greater attention. Attention can be measured through observing the dilation of pupils which has been noted as a display of learning. Cognitive learning behaviors are the measured analytical steps an infant takes. Repetition exposure to a learning opportunity such VOE events can lead to familiarization, a process in which an unexpected event becomes predictable. There has not been a significant increase in memory during the surprising event but VOE events create opportunities for cognitive learning. 

Visual Learning Mechanisms

Visual learning mechanisms are the preceding process in VOE learning opportunities because an infant must be engaged in order to learn. Looking time was operationally defined as attention to test if infants focused their attention on VOE events. Sim and Xu (2017) examined the looking time infants before they approached colorful balls that violated their expectations. Infants displayed surprised reactions after seeing four balls of different colors each enter a box and four balls of the same color leave it. The time the infants spent looking at the event that violated expectations was longer than the event where the balls remained different colors. When given the opportunity to explore, they preferred to touch and reach inside the box with the same colored balls. Obtaining the infants’ attention was a key component in facilitating learning behaviors and lends evidence to the idea that children can direct their attention towards learning opportunities. 

How can directing an infant’s attention promote learning behaviors in VOE situations? Social referencing, when infants look to their caretakers for how to emotionally respond, is one process of promoting learning behaviors. In situations where infants experience a VOE event, they have been recorded referencing caretakers in their environment in search of explanation or ways to react. Dunn and Bremner (2016) measured infant social looking, a look that is directed at the caregiver, after exposure to a situation that violated expectations. The social look is thought to obtain the caregiver’s attention and is a measure of behavioral learning. Infants were shown a toy before the task presenter came onstage in front of them and then were shown either the same toy or a different toy. Results showed that infants in the object-switch group looked longer and engaged in more social looking to their caregiver. The object switch group was a violation of expectations which facilitated learning through social looking behaviors. 

Learning through social referencing has also been observed through infant expression. False-belief, an emotional display of VOE, was examined in an experiment by Scott (2017) which looked at infants’ display of false-belief and their attribution of internal states to others. This phenomenon is described to be a form of affective learning in the form of displaying states that infants believe are expected of them through social referencing. The display of false belief was prompted by exposing infants two rattles, one that rattled and one that was silent. Caretakers displayed surprised or satisfied reactions when exposing infants to the rattle in either a congruent or incongruent manner. Infants were surprised by the VOE rattle that was silent and expected the caretaker to display a surprised reaction as well, which was measured by time spent looking. 

Kinesthetic Learning Mechanisms 

The longer an infant stared at events that defied expectations, the more likely they were to explore them. Violation of expectations is a first-person cue from the environment that facilitates explanation seeking curiosity. This is measured by kinesthetic learning mechanisms such as approaching and playing. Schulz (2015) found that play supported learning more so by selectively exploring objects that both violate previous expectations and have an equal probability between outcomes. Core knowledge, the idea that cognitive representations of objects are innate, support infant exploration and quick accumulation of knowledge by signaling opportunities to learn (i.e. situations that violate expectations). This indicates that infants seem to approach opportunities they believe they can learn about, such as events that defy expectations. 

Stahl and Feigenson (2017) studied the action of kinesthetic learning mechanisms and found that infants prefer the unexpected over novel objects. Infants watched an object that aligned with principles of object solidarity, the concept that solid objects can not pass through one another, compared to an object that violated those principles (a rolling toy car falling off a ledge vs a toy car continuing to drive and not falling off a ledge). After being exposed to the objects, infants preferred to explore the toy that did not align with expectations over the one that did. Two common exploratory behaviors, banging objects and dropping objects, were observed. Banging is associated with testing object solidarity and dropping tests object support. These behaviors were determined to be a greater effort to learn and explore why the object did not align with preconceived expectations of the world. 

Infants have been shown to seek out VOE situations and to explore and understand them, but at what age do infants display the curiosity to explore the out of ordinary? To determine the age at which this curiosity can be observed, Fargard et al. (2015) placed a rattle out of reach of infant participants to observe approach behavior towards objects of interest. 

The ball of the rattle was the object of interest and the handle was intended to measure grasping behavior as a means of obtaining an object of interest. Infants ages 6 months and ages 8-10 months were given a trick rattle in which the ball would fall off when moved. Infants 6 months old did not show surprised reactions to the trick rattle compared to 8-10 months old infants who displayed surprised reactions. The results add evidence to the idea that motivation to understand cohesiveness and events that defy previous expectation is present around 8-10 months.

Cognitive Learning Mechanisms

Cognitive learning mechanisms are different from visual and kinesthetic because they are internal cognitive processes rather than overt actions. Cognitive learning mechanisms can be studied through measures of pupil dilation and recall memory tests, both indicators of internal processes. Jackson and Sirois (2022) measured attention by how much larger the infants’ pupils dilated after witnessing an unusual event. They showed infants one of two video clips, a train going through a tunnel and reappearing the same color or a train reappearing through the tunnel a different color. After viewing the clips, the results showed that pupils dilated more during possible and familiar events leading to the conclusion that infants were familiarized with the violation of expectations and therefore became more responsive to the familiar events than unfamiliar. Repetition was associated with greater familiarization and therefore a greater interest in the familiar. 

Familiarization is the process when VOE becomes gained knowledge. This process of learning appears in the form of recognizing patterns, such as predicting number patterns. Mermier (2022) further examined the process of familiarization in infants through statistical learning. Infants in this study were first familiarized with separate parts of a turning sequence and then were shown either one of two clips of actors. This turning sequence involved actors facing each other and then turned to the camera repeatedly displaying a different facial expression during each turn. The actors’ turns had either a pattern of high predictability or a pattern of low predictability. While watching the clips, the participants were observed and coded for time spent looking at the clips and during which parts of the turning sequence they looked most frequently. The results indicated that infants were able to differentiate between high predictability and low predictability likely due to greater familiarization. Infants showed a slightly higher significant response to the looking towards parts than the looking away parts which indicates a level of sociality involved in the process of familiarization.

There are limitations to the ways in which VOE influences learning. Surprise at the time of encoding has not been found to lead to enhanced memory. In a study conducted by Csink et al. (2021) 17-month-old infants were assigned a memory encoding condition in which they were shown familiar stimuli labeled either incorrectly and correctly. This condition was followed by a recognition test that included new items as well. This test was designed to heighten infant attention to VOE and promote learning mechanisms. Surprise, which was operationalized as pupil dilation, was measured throughout the exposure to familiar/unfamiliar stimuli. Increased dilation was associated with incorrectly labeled stimuli as well as successful memory retrieval during the recognition test, but infants only displayed differentiation between familiar/unfamiliar stimuli during the memory recognition test and not during the encoding condition. The results show that surprise at the time of encoding does not lead to enhanced memory but agrees with previous findings that VOE increases attention and learning mechanisms in infants. 

Discussion 

VOE studies have limitations in reliability and accuracy. Reliability of cognitive and behavior infant research was examined by O’Grady and Dusing (2015) who concluded, after reviewing experimental designs, that cognitive and kinesthetic measures can be valid but measures can have poor interrater reliability or test-retest reliability. A way to mediate low reliability is to strengthen the assessment of psychometric properties, the experiment characteristics of human characteristic studies. Rubio (2019) reviewed the validity of VOE studies and false-belief reasoning. The proposed changes to increase validity after analyzing 15 VOE studies were to measure looking at and looking away times which were often not measured. Including multiple eye tracking measures, such as corneal reflection tracking, increases the validity and the accuracy of measurements as infants can look away without focusing their attention on something different. To increase the validity of the parameters characteristic of the variables studied, a theoretical justification for the definition of parameters should be included in study design. An example of this is explaining why looking times were defined as familiarization in a particular study. Finally, reporting pilot data from pretests increases validity and understanding of why certain criteria was used in the study. Increasing the accuracy of measures of infant attention creates more reliable measures.

Despite limitations, online testing has shown solid evidence as a valid study design for VOE experiments. Smith-Flores (2021) considers online testing for VOE experiments and the applicability of results outside of the laboratory. Infants were exposed to VOE experimental designs and were reported to stare longer at unexpected events, learn novel concepts better when they violated preconceived notions, and preferred to look the most at the source of the utterance in a familiarization test; all results concurred with previously mentioned VOE knowledge. 

References

Csink, Mareschal, D., & Gliga, T. (2021). Does surprise enhance infant memory? Assessing the impact of the encoding context on subsequent object recognition. Infancy, 26(2), 303–318. https://doi.org/10.1111/infa.12383

Dunn, K., & Bremner, J. G. (2016). Investigating looking and social looking measures as an index of infant violation of expectation. Developmental Science, 20(6), e12452. https://doi.org/10.1111/desc.12452

Fagard, J., Florean, C., Petkovic, M., Rat-Fischer, L., Fattori, P., & O'Regan, J. K. (2015). When do infants understand that they can obtain a desired part of a composite object by grasping another part?. Infant behavior & development, 41, 169–178. https://doi.org/10.1016/j.infbeh.2015.05.003

Jackson, & Sirois, S. (2022). But that’s possible! Infants, pupils, and impossible events. Infant Behavior & Development, 67. https://doi.org/10.1016/j.infbeh.2022.101710

Liquin, E. G., & Lombrozo, T. (2020). Explanation-seeking curiosity in childhood. Current Opinion in Behavioral Sciences, 35, 14–20. https://doi.org/10.1016/j.cobeha.2020.05.012

Mermier, Quadrelli, E., Turati, C., & Bulf, H. (2022). Sequential learning of emotional faces is statistical at 12 months of age. Infancy, 27(3), 479–491. https://doi.org/10.1111/infa.12463

O’Grady, & Dusing, S. C. (2015). Reliability and validity of play-based assessments of motor and cognitive skills for infants and young children: A systematic review. Physical Therapy, 95(1), 25–38. https://doi.org/10.2522/Ptj.20140111

Ramos, J. A. (2018). Important Child Development Stages - First 5 Years. Children's Bureau. Retrieved May 27, 2022, from https://www.all4kids.org/news/blog/why-the-first-5-years-of-child-development-are-so-important/

Rubio-Fernández P. (2019). Publication standards in infancy research: Three ways to make Violation-of-Expectation studies more reliable. Infant behavior & development, 54, 177–188. https://doi.org/10.1016/j.infbeh.2018.09.009

Schulz L. (2015). Psychology. Infants explore the unexpected. Science, 348(6230), 42–43. https://doi.org/10.1126/science.aab0582

Scott R. M. (2017). Surprise! 20-month-old infants understand the emotional consequences of false beliefs. Cognition, 159, 33–47. https://doi.org/10.1016/j.cognition.2016.11.005

Sim, Z. L., & Xu, F. (2017). Infants preferentially approach and explore the unexpected. The British journal of developmental psychology, 35(4), 596–608. https://doi.org/10.1111/bjdp.12198

Smith-Flores, A. S., Perez, J., Zhang, M. H., & Feigenson, L. (2022). Online measures of looking and learning in infancy. Infancy, 27(1), 4–24. https://doi.org/10.1111/infa.12435

Stahl, A. E., & Feigenson, L. (2015). Cognitive development. Observing the unexpected enhances infants' learning and exploration. Science, 348(6230), 91–94. https://doi.org/10.1126/science.aaa3799