Introduction

Although there is a large body of literature focused on children’s understanding of emotions and another body focused on parent-child conversation about emotions, little research has examined parent-child talk about complex emotions. The majority of research is dedicated to simple emotions and their functions in children’s life. However, the studies which are dedicated to simple emotions admit the sufficient importance of parent-child interaction for better emotional understanding. The emerging of complex emotions become a disputable process, moreover these emotions are functionally more complicated than simple. The following research has argued the importance of parent-child conversation in terms of understanding complex emotions by children. To account for this gap in the literature, the present study focuses on parent – child talk about complex emotions.

To understand this topic, a review of the literature on complex emotions is needed. Much past research has looked at ambivalent emotions. However, researchers use different names such as ambivalent emotions (Donaldson & Westerman, 1986), mixed emotions (Harris, 1989a, 1989b; Harter & Buddin, 1987; Kestenbaum & Gelman, 1995), conflicting emotions (Bennet & Hiscock, 1993) and multiple emotions (Meerum Terworgt, Koops, Oosterhoff, & Olthof, 1986; Wintre & Vallance, 1994). There are many definitions that these authors use, but in spite of this fact, these authors express the similar idea about this kind of emotion: when an individual has two contradictory feelings at the same time. For instance, we may be happy to receive a present, but disappointed at the same time, as a present is not what we have expected to receive.

Children’s understanding of emotions is a gradual process that develops from simple types of emotions to more complex ones (Tenenbaum, Visscher, Pons, & Harris, 2004). According to Pons, Harris, de Rosnay. (2004) and Tenenbaum et al. (2004) the understanding of emotions has three broad categories which develop gradually: external, mentalistic, and reflective. The external component consists of recognizing faces, the ability to realise emotions when they are affected by different external circumstances and understanding the fact that past emotions might affect present ones. Recognizing faces means that children start identifying emotions according to a facial expression. Then, they understand how external cases can cause different feelings and the last component is when they realize that some subjects from the past might cause emotions in the present. This type of emotion understanding usually develops among children between the ages three to five. The second component, mentalistic emotions, contains of realizing beliefs and desires, also identifying the difference between real and apparent emotions. Children start assuming the idea that people may have different emotional reactions due to the fact that they have dissimilar desires. The same situation occurs with beliefs, as children start understanding that people’s beliefs will influence their emotions to a situation. The last element is when children can understand the link between memory and emotions, for example, the strength of emotions might reduce with time. The authors claim that this process usually occurs between the ages five to seven. Finally, the third component consists of reflective emotions which include moral, regulated and ambivalent emotions. Children start experiencing several moral emotions such as guilt, shame or embarrassment when their actions did not correlate with generally accepted rules. They also start realizing the way they can control their emotions, using either behavioural or psychological strategies. The last stage is an understanding that people might have two different emotions at the same time. Overall, the third component usually emerges at ages seven to nine. That classification is widely used in terms of defining emotions. Consequently, ambivalent emotions, which are one of the primary aims of the following work, are situated at the third level and presented to be one of the most complicated emotional categories to understand.

Some authors (Donaldson & Westerman, 1986; Harter & Buddin, 1987) created a model of ambivalent emotions which consists of four different stages. This model was invented after clinical observations of play therapy. Children used to experience problems in understanding two emotions at the same time. Consequently, a question about developing of understanding of mixed emotions arouses. The following model consists of four stages including such factors as valence (positive or negative emotions) and a target (one or several targets). The stages are presented gradually from the least to the most advanced. The first level, which usually arises at seven years old, is presented the same valence and target, where two positive or two negative emotions are expressed towards one object (e.g., “A boy was happy and excited because of his birthday”). The second level, which might emerge around eight years old, is represented as same valence, but different targets (e.g., “A girl was happy about her birthday and excited that so many guests had came”). The third combination, usually develops at ten years old, is shown as different valence and targets (e.g., “A boy was sad he could not go for a walk, but happy because he could watch TV”). Finally, the fourth combination, arises approximately at 11 years old and consists of different valence and same target, a child has different feelings towards one object or situation (e.g., “A girl was happy to go for holiday with parents, but sad to leave her friends”). Though, authors give an explanation that understanding and experience may not be contemporary, as it is possible that a child might experience two emotions at the same time, but not be able to realize it yet.

Though, there is a contradictory view about the levels of emotional development. Wintre and Vallance (1994) present their theory where alongside with multiple emotions and valence, they also depicted the intensity of emotions. Based on several previous studies, they offered the theory of development of understanding mixed emotions, which include such components as intensity and valence and multiplicity. For example, at level A that emerges usually at 4 years, there might be observed only intensity directed to one situation, though valence and multiplicity have not arouse yet. At level B which develop around 5 years old, children may recognize several mixed emotion of the same valence, but they do not use different intensity. Next level C, usually arise at 6 years old, where intensity is combined with mixed emotions, but there is only the same valence involved yet. At the last level D, which might emerge at 8, where children show the ability to use all three components together. Besides they differentiate not only stages and the main components of emotional understanding of ambivalent emotions, but the age of emerging as well. Thus, for Harter and Budden (1987) the last stage might be reached children approximately by ages 10 and 11, whereas for Wintre and Vallence (1994) the period for the last component is age 8.

However, there is another arguable point is the age period at which complex emotions arise. For example, several studies (Donaldson & Westerman, 1986; Meerum Terwogt et al., 1986; Harter & Buddin, 1987; Harris, 1989a, 1989b; Pons et al., 2004; Tenenbaum, 2008) pointed to the approximate period of developing complex emotions from seven to eight years old until eleven to twelve. To illustrate this point, there are several examples based on the studies. According to Harter & Buddin (as cited in Donaldson & Westerman, 1986) understanding of ambivalent emotions is quite a late cognitive process. She found that children were not able to distinguish emotions that occur simultaneously toward the same person until ten and a half years old. Taking into consideration her study, Donaldson and Westerman (1986) replicated the findings, showing that though few characteristics of understanding ambivalence appeared at seven to eight years old, only older children of ten to eleven showed this ability entirely. A study conducted by Harter and Buddin (1987) supported the fact of gradual development of understanding ambivalent emotions. Thus, children four to five ages cannot believe that two feelings could be together, six to eight years old placed feelings in “temporal” order as when a child did not see the object which had caused his emotions, he simply forgot about it. However, children from eight to twelve usually can describe the appearance of two emotions at the same time. Meerum Terwogt (1986) argued that ten year old child understood the fact of having both emotions equally strong, but younger children had always chosen one to be stronger than another. They present an example of two groups of children: six and ten years old. Undoubtedly, ten years old children understand ambivalent emotions more often than six years old children, especially if the situation consists of opposite valence emotions. To sum up these studies, the most widespread period of time when mixed emotions develop is from seven to twelve years old.

Though, there are some authors (Gnepp and Klayman, 1992; Wintre and Vallance, 1994) whose studies admit the arising of understanding two simultaneous emotions at the age of 8. They gave some explanations for this, where not only cognitive factor plays a crucial role, but social experience also becomes an important factor. Consequently, these authors did not share either late or early emerging of mixed emotions.

However in spite of all these findings, there are other studies, which contradict the idea of such a late arising of ambivalent emotions. For instance, Kestenbaum and Gelman (1995) pointed that the development of mixed emotions appeared between ages four and six. Children of four and five years old can recognize ambivalent emotions when they are presented in a facial expression, even if they were on the same face (sad eyes and happy mouth), as well as on a picture with two-headed alien who expressed different emotions. Moreover, the results confirm that five year old children might understand mixed emotions within a simple story. Furthermore, it was offered two levels of understanding emotions at the preschool ages. The first one is recognition of ambivalent emotions showing on the face. The second level is an ability to match a face that expressed mixed emotions with a situation. All in all authors claim that this is a gradual process which starts early and increases with age. Their findings contradict the other researches which admit that a development of mixed emotions is a process starting much later. Likewise, Brown & Dunn (1996) mentioned that even six years old children might show an understanding of mixed emotions when they are given a slight prompt about feelings of characters in stories. Similarly, Harris, P. L., Johnson, C. N., Hutton, D., Andrews, G., & Cooke, T. (1989 a) consider the fact that even young children may predict about different emotions in a more advanced way. Even children between the ages of three to seven are very sensitive to other people’s desires, emotions and beliefs. Moreover, they do not merely take into account a situation by itself, but modify their reaction due to different situational changes. That in case might show children’s advanced ability to understand second mentalistic component: beliefs and desires. Considering this fact, they suggest that understanding of the second level arouse at three, but set up only between the ages five to six years. Consequently, the next level to develop will be the third, which contains moral, ambivalent and regulated emotions. The possible explanation of this phenomena could be found in Harris’s (1989 b) study, he pointed that even in a behaviour of one-year-olds might be seen the ambivalence, whereas conscious realization does not arise before seven or eight years old. He stressed that children start expressing ambivalence quite earlier, but understanding it emerges only from eight to ten, young children can merely express ambivalence without understanding of this fact.

To sum up all the previous research made according to this topic, it can be clearly seen that these studies implied a large range of age when complex emotions appear in the childhood. Though, it might be criticised the period of appearance of mixed emotions. As if to consider all researchers, described earlier, there are some who consider the emergence of ambivalent emotions as a very early process (between ages 4 to 6 years old). Some consider it to be the late process (between 9 to 12 years old). Besides, there are some who support the idea of appearance of multiple emotions at age 8. However, the researchers who offered the late period of appearance ambivalent emotions might be criticised for a big age gap between groups. For example, in the research of Meerum Terwogt et al. (1986), they conducted their study with two groups of children by ages 6 and 10. What may be considered not exactly relevant, as during these 4 years undoubtedly children will have a great step forward in terms of understanding ambivalent emotions.

In present study the age group of seven, eight and nine were chosen in order to observe the difference between the age group according to rather contradictory fact of arising the understanding of mixed emotions. However, there are some findings that admit the fact of emerging mixed emotions quite early (Harris et al., 1989a; Kestenbaum & Gelman, 1995; Brown & Dunn, 1996), the excessive amount of research claim the period from seven to twelve years old. Taking into consideration both contradicting findings together, it was decided to concentrate the research on the following groups of children due to some reasons. For instance, such ages as seven, eight and nine present an ideal age group where ambivalent emotions might just have started and developed at the late stage. Besides, it appeared to be not enough amounts of studies to support the idea of early development of mixed emotions, hence, children of younger ages were not considered. Moreover, according to Harris’s (1989 b) study, where his notable remark cannot be ignored, children might notice the ambivalence, but not understand it. As the major aim of this work is to find out when mixed emotions are understood by children, these period of ages were chosen.

Parent-child conversation

One way in which children understand emotions may be through the family via interaction with parents (Dunn et al., 1991; Harris, 2004; Racine et al., 2007). Different researchers have examined different types of conversations that may support children’s understanding. For example, Engel (as cited in Fivush & Fromhoff, 1988) described two maternal styles, which are called: conversation-eliciting and directive. The first one is characterized by asking many questions, rare interruption of children, and a regular tendency to integrate a child’s answer what in turn leads to more equal and productive conversation. The second type is characterized by different types of commands, frequent interaction in child’s actions, as well as the high parental control during the conversation.

Another classification was offered by Reese & Fivush (1993) and Reese, E., Haden, C.A. & Fivush, R (1993) and it claimed that children’s understanding emotions correlates with the parental conversational style. They offered two parental styles: high – elaborative and low- elaborative. For instance parents with “high-elaborative” style, which consists of a lot of details, explanations, always try to arouse children’s interest. On the contrary there is other kind of style called “low-elaborative” where parents ask simple questions during their interaction with children and give them little new information and quite often change topics of conversation. Similar types of maternal styles were offered by Fivush and Fromhoff (1988) where the styles are called: elaborative and repetitive. Elaborative mothers are likely to provide a lot of details during the discussion and tend to ask many open-ended and complex questions during the interaction. On the contrary, mothers who have repetitive conversational style tend to ask “yes/no” questions; do not provide so much detailed information during the discussion. Besides, the authors depict that elaborative style is significantly better for children to remember past events during the conversation. Overall, all previous research concerning maternal style of conversation was summarized (Reese et al., 1993) The mothers who are more engaged in a conversation with their children – “elaborative, high-elaborative, reminiscing or topic-extending” and mothers who are less involved in a communication process – “repetitive, low elaborative, practical remembers or topic-switching” (p. 404).

There are many studies that support the link between parent-child conversation and children’s emotional understanding (Dunn, Bretherton & Munn, 1987; Fivush & Fromhoff, 1988; Dunn, J., Brown, J., Slomkowski, C., Tesla, C., & Youngblade, L., 1991; Dunn & Brown, 1994; Brown & Dunn, 1996; Steele et al., 1999; Pons et al., 2003; Laible & Song, 2006). Some of the authors connect emotional understanding with maternal attachment (Brown & Dunn, 1996; Steele et al., 1999; Pons et al., 2003), the other with the quality of maternal speech (Dunn et. al., 1987; Dunn & Brown, 1994).

To begin with, there is as well a link between quality of parent-child attachment and the quality of family talk about emotions with children’s emotional understanding (Pons et al., 2003). Brown and Dunn (1996) also consider a possibility that children’s development of ambivalent emotions can depend upon the family where a child grows. Moreover the findings, presented by Steele et al. (1999) suggested that the development of secure mother-child attachment at one year helps to shape a good understanding of mixed emotions later at six years.

Much past work has argues that when mothers talk a lot about emotions, children consequently start using more emotional words in conversation (Dunn et. al., 1987). However, , children in families in which the use of negative emotions is greater than positive emotions may have some difficulties expressing and recognizing emotions (Dunn & Brown, 1994).

Studies conducted by Dunn et al. (1991) showed a significant link between the parent-child conversation and children’s emotional understanding. Thus children whose parents spoke with them a lot about emotions at the age of 3 demonstrated better understanding about the emotions of unfamiliar adults at the age of 6 in comparison with the children who did not experience feeling- state talk so frequently. Laible (2004) and Thompson et al. (as cited in Laible and Song, 2006) argue that style is more influential than content. Parent- child conversation helps children realise different situations of everyday life, especially if these situations are not so obvious to see. That usually includes emotions, motives and intentions. Besides, when mothers involved with children into conversation about past experience, they not only give their children an emotional understanding of the past, but also bring extra knowledge for future experience. The other findings of Laible & Song (2006) indicated that parent-child discourse was a crucial predictor of the child’s socio-emotional development. For example, during the given task if mother used more elaborative style rather than repetitive, children had higher scores on emotional understanding. The similar results were presented by Steele et al. (1999) where children and parent’s language abilities did not affect children’s understanding of ambivalent emotions, what in turn depicted the fact about the major influence of socio-emotional rather than cognitive elements.

Considering all previous research conducted on this topic, it is without doubt the effect that parent-child conversation plays in terms of children’s emotional understanding. However, some studies use different core point to correlate this connection, the influence is still might be observed. Particularly, the more elaborative and positive discourse parents and children have, better emotional understanding children express in comparison with children whose parents use “low-elaborative” style, and consequently these results show the great impact of parental conversation with children on children’s emotional understanding. The same situation is presented with an attachment, the more secure attachment is observed between mother and child, and the better emotional understanding will be expressed by a child later. Hence, the parent-child conversation was chosen in order to correlate it with children’s understanding of mixed emotions.

Parent – child discourse may however vary with child gender. That is the importance of gender difference in terms of parent-child interaction, which cannot be denied, although the studies are rather controversial, thus some researches refute it (Racine et al., 2007; Dunn et al., 1991) and some support (Dunn et al., 1987; Fivush, 1991;Kuebli & Fivush, 1992; Reese & Fivush, 1993; Cervantes& Callanan, 1998). To begin with, there are examples of several studies which support the gender difference in parent-child interaction. The findings of Dunn et al. (1987) claimed that mothers had more communication with girls rather than boys and later girls referred to emotional states more often than boys. Moreover, mothers expressed more emotional words to daughters rather than to sons. Consequently, girls used more feelings’ words during the discourse in comparison with boys. Though authors pointed one of the limitations to be a small sample of children who took part in the research, the studies made afterwards can support these results. Namely, Reese & Fivush (1993) argued that parents of daughters expressed more elaborative style than parents of boys and the authors believed that due to more prolonged conversation with daughters, they can grow up to produce more elaborative narrative style in comparison with sons. The authors admit that in the majority of cases parent-daughter pairs involved in more detailed conversation and as well as this style required more involvement of children’s memory, daughters may develop more elaborative style in comparison with sons. Similar results are presented by Kuebli & Fivush (1992) where they admit parents of daughters use quite more emotional words in comparison with parents of sons. Their study shows that parents speak more about sadness with girls rather than with boys. Besides, the study conducted by Cervantes & Callanan (1998) also showed the gender difference for children at the age of 2 but not at the age 3 and 4 in terms of children’s talk, there were also gender-related differences in mother’s talk for using more explanations for boys and more labels for girls. In the light of this evidence, it is obvious that if there is a gender difference during parent-child conversation, parents usually pay more attention to interaction with girls rather than with boys, that is why the aspect of gender difference seemed to be quite significant in the present study. One good point might summarize the gender difference idea. Though, mothers tend to socialize with a child in different ways according to the gender. Besides, it may be that boys and girls are experiencing and discussing emotions in different ways and consequently, their mothers simply respond to this difference (Fivush, 1991).

All things considered, in the following study, there is an aim to find out whether parent-child discourse about mixed emotions plays a significant role in children’s understanding of these emotions. Based on a substantial literature (ref), the first hypothesis predicted that the more parents explain complex emotions, the better children will understand emotions. A second related hypothesis is that the more emotion words parents use, the better the child’s emotional understanding will be. The third hypothesis is that parents will use more emotion words with girls than boys and consequently, girls will use more emotion words in comparison to boys.

Method

Participants.

The participants were 16 parent–child pairs. Children were dived for three groups according to their age: the younger group was 7 years (M = 7.5, ranging from 7 years, 4 months to 7 years, 10 months), 8 years (M = 8.48, ranging from 8 years, 1 month to 8 years, 9 months) and 9 years (M = 9.42, ranging from 9 years, 1 month to 9 years, 8 months). There were both 5 children at the group of 7 and 9 years old and 6 children at the age group 8 years old. Overall there were nine boys and seven girls. Participants and their parents were recruited from one primary school in Dunstable. All participants were white British. It is notable that all parents that took part in this study were mothers; no fathers participated in the research. The majority of mothers (12 people) were employed, 1 mother was employed, but worked at home and 3 mothers were homemakers. All mothers have educational background ranging from primary school to university.

Materials

The session with each pair will be held only once. To begin with, it is notable to mention that the books were chosen for this research as a discussion in general seems to be a quite crucial for children in order to reach better emotional understanding. Therefore, in the joint activity, such as reconstruction of events, adults give an idea for children what emotions are appropriate and inappropriate for different situations (Fivush, 1991). A created story book which contains 6 vignettes about 3 complex, 2 regulated and 1 moral emotions. The book was created by the researcher and was particularly organised for boys and girls separately. The story was about two children: a boy – Peter and a girl – Anna, who took part in several activities, such as: passing an exam, having a birthday, going to amusement park, home interaction, his/her friend leaving. There was one question concerning emotions for every situation to which a child was involved in. The second book is called “Frog goes to dinner” by Mercer Mayer (1974). This book contains several vignettes, presenting a story about a boy who has a frog. Accidentally, a frog goes to dinner with a boy and his parents without being noticed by anybody. A frog spoiled the dinner and the parents were very angry with a boy. The important fact is that this book is wordless; obviously the task was to create a story together. This particular method was chosen because Kestenbaum and Gelman (1995) pointed the importance of presenting information to children during the research not only verbally, but giving photographs or drawings as stimuli. To justify their point, there were two reasons presented in their study. The first reason is that a facial expression helps children understand emotions significantly better. The second reason is that the absence of pictures might cause some difficulties as children might not know how two different emotions can be expressed at the same time. After reading two books with the parents, children were tested by the Test of Emotional Comprehension which was created by Pons, Harris and de Rosnay (2004). The TEC presents vignettes in which a gender-matched protagonist encounters simple to complex situations that elicit different emotional responses. After each vignette, the child is asked how the protagonist is feeling, by choosing from four illustrations of faces representing different emotional states (Tenenbaum et al., 2004).

The TEC consists of nine sections which included several pictures and a question about emotional condition of protagonist. Though, the first section requires naming emotions that the faces show. The second includes the situation where a book character expresses the influence from external factors. The third one consists of desires that a protagonist expresses. The fourth factor involves understanding of false beliefs and the way they influence on emotions. The fifth section expresses the situation with reminders that might influence on children’s present condition. The sixth one asks children to control their emotions. The seventh section includes hidden emotions that a child should define. The eighth factor presents the situations with mixed emotions, particularly on what this study focuses. The last component involves understanding of moral emotions.

Parents were offered to fill in a “Parent Questionnaire” which contains several questions about background information and social economic status.

Procedure.

Parents with children read two books: one is created by the researcher and another is “Frog goes to dinner” by Mercer Mayer. The first book was created by the researcher and concluded six situations including such emotions as mixed, moral and regulated. A task required to read the stories and a child should have answered an offered question. It was not specially defined who suppose to read a book. Consequently, mothers could read it as well as children, or they could do it in an order. The vignettes, which were presented in a book, concerned several situations at school, at home, at an amusement park, on holidays. For example, a situation for girls that involved mixed emotions: “Anna’s birthday is during summer holiday, so she can do everything she’d like to do- ride a bicycle, play with friends and eat ice-cream! Her parents decided to arrange a birthday party for Anna. She likes parties very much and a lot of guests will bring gifts to her. But her best friend, Lola, is unwell and now she can’t go to Anna’s birthday party. Anna wants Lola to come so much. How does Anna feel now?” After reading a created book together, they were offered to make up a story together. A book “Frog goes to dinner” by Mercer Mayer (1974) was used as a good example of wordless story. These two tasks were chosen in order to give parents and children enough time to interact with each other. Moreover, they both contain different emotional situations describing which should promote an excessive usage of emotions.

After making up a story a child was tested with the Test of Emotional Comprehension, which was created by Pons, Harris and de Rosnay (2004) in order to measure the period of time when children start understanding different emotions. The model of Pons, Harris and de Rosnay (2004) was accepted as the main one in a theoretical description of emotional understanding of complex emotions. Besides, the test consists of pictures, and the questions are always asked by a researcher. It is also very convenient for children to conduct this test due to the absence of necessity to read. Consequently, their test was assumed to be relevant to use during the research. Besides, it is worth mentioning that the researcher was trained before to conduct this test, as the requirements were to conduct it with neutral voice, do not give any prompts to children which emotions a protagonist might express.

During conducting the TEC with children, parents were offered to fill in the “Parent Questionnaire” which contains of such sections as child and parent’s names, child’s age, date of birth and birth order, also social economical situation of parents (ethnicity, occupation, level of education, marital status). Besides, parents signed a consent form which stated that they and their children agreed to take part in the present research and they did not object to be video taped. Moreover, this form included all information about research, it was also mention that parents were free to withdraw at any time they want from the study and contact details of researchers were provided. Children were asked oral permission if they agreed to help a researcher to conduct the study and neither parents nor children refused. Parents were given a choice to stay or leave during a conducting the TEC with children, but it is worth mentioning that all parents were present during children’s testing.

Coding

A created book that was basically consisted of 6 vignettes was scoring in the following way. It is notable that in terms of this book the amount of using em