Received 31 March 2016; accepted 19 April 2016; published 22 April 2016
1. Introduction
Over the last few decades there has been growing interest in research on parents and their involvement in children’s lives. Research has shown that parents’ engagement with their children is related to children’s linguistic and cognitive skills (e.g. Tamis-LeMonda, Shannon, Cabrera, & Lamb, 2004 ). Research on language input to children and parent-child interactions have mostly focused on motherese, that is, maternal language use. Studies on paternal language have been very limited. However, recent studies on paternal language have shown that fathers also contribute to their children’s language development (e.g. Tamis-LeMonda, Shannon, Cabrera, & Lamb, 2004; Duursma, Pan, & Raikes, 2008; Pancsofar & Vernon-Feagans, 2006, 2010; Pancsofar et al., 2013; Duursma, 2014 ).
Research also documents that family socioeconomic status (SES) is a powerful predictor of many aspects of child development (Hoff, 2003) , especially linguistic development. It is well established that children from lower SES build their vocabularies at slower rates than children from higher SES. More educated and advantaged parents have children with greater vocabulary skills and faster vocabulary growth during early childhood than less educated and advantaged parents. Further research has shown that this relation between SES and child vocabulary skill is due, in part, to the speech that parents offer children during day-to-day interactions (Hoff, 2003; Rowe, 2008) . Studies focusing on SES and language development in children (Hart & Risley; 1995; Hoff, 2003, 2006; Cabrera, Shannon, & Tamis-LeMonda, 2007; Duursma, Pan, & Raikes, 2008; Rowe, 2008) found differences related to socioeconomic backgrounds of the families. Hoff-Ginsberg (1998) observed that child-directed speech in high SES families contained richer vocabulary of object labels. Especially high SES mothers prefered longer utterances and a variety of words. Thus, since high SES children had access to different words, they possesed larger vocabularies. Low SES mothers, on the other hand, talked less and used less varied vocabulary during child-directed speech. Heath (1983) and Ninio (1980) also found that parents from low SES talk much less to their children than high SES parents. According to Hoff, Laursen & Tardif (2002) , maternal language input to children changes in accordance with family SES. High SES mothers support child language development. McCabe & Peterson (1991) and Peterson & McCabe (1992) point out that maternal and paternal language input to young children have an impact on children’s later narrative styles. Studies on shared book reading found that parents who have a better education demonstrate better literacy skills. Thus, these studies showed that parents’ socioeconomic backgrounds have an impact on the emergent literacy skills of their children (Raikes et al., 2006; Duursma, Pan, & Raikes, 2008; Duursma, 2014) .
Studies focusing on maternal and paternal language (Fernald, Taeschner, Dunn, & Papousek, 1989; Kava- naugh & Jirkovsky, 1982; Rondal, 1980; McRoberts & Best, 1997; Leaper, Anderson, & Sanders, 1998; Panc- sofar & Vernon-feagans, 2006, 2010) have found similarities as well as differences between parents’ child- directed speech. Golinkoff & Ames (1979) compared maternal and paternal language in a dyadic and triadic situation and found that in the dyadic situation, mothers and fathers produced similar number of utterances but in the triadic situation, the fathers produced fewer utterances than mothers. Fernald, Taeschner, Dunn & Papousek, (1989) compared mother and father verbal input and they found that fathers adopted a simplified speech register and spoke with a higher pitch. Studies examining language input in the context of book reading suggest that maternal as well as paternal book reading can help children’s language development (Raikes et al., 2006; Duursma, Pan, & Raikes, 2008) . There are some other studies in which differences between parents on measures of total output have not been observed (McLaughlin, White, McDevitt, & Raskin, 1983; Hladik & Edwards, 1984; O’Brien & Nagle, 1987) .
This cross-sectional study, focusing on the link between families’ SES and their language use, aims to investigate whether parents’ SES, i.e. education and income, influence their interaction styles with their children in terms of dialogicality. Several studies on parent-child interaction investigated language input of either mothers or fathers in contexts such as book-reading, picture task, meal time and dressing. This study investigates the properties of both maternal and paternal language input by examining the interaction between parents and their preschoolers in the context of toy play with regard to dialogicality.
The idea that knowledge is constructed through dialogue goes back at least to the time of Socrates and has been reiterated by many others since then. Dialogicality refers to “the ways in which one speaker’s concrete utterances come into contact with the utterances of another” (Wertsch, 1991: p. 54) . Utterance refers to Bakhtin’s notion that meaning does not reside in words; rather, it resides in the ways in which words are used in particular contexts (Ball, 1999: p. 232) . As Bakhtin (1986: p. 92) points out, “After all, our thought itself―philosophical, scientific and artistic―is born in the process of interaction and struggle with others’ thought”. According to Bakhtin, dialogicality refers to “all texts―written as well as spoken―that set up, in one way or other, relations between different voices” (Fairclough, 2003) . For Bakhtin (1986: p. 69) , an utterance is always and necessarily part of an ongoing dialogue in some sphere of activity. No one ever has the last word and equally, nobody breaks the silence of the universe. Thus, whenever we speak, we necessarily enter into an ongoing dialogue (Wells, 2007) . While monologic discourse is usually associated with fixed transmission of unchanging ideas and status inequalities; dialogic discourse connotes social relationships of equal status, intellectual openness, and possibilities for critique and creative thought (O’Connor & Michaels, 2007: p. 277) .
Research suggests that when adults give children opportunities to become active participants during conversations, i.e. when the interactions are more dialogic, children show greater language gains (Whitehurst & Lonigan, 1998) . There is an increasing body of research that supports the view that talk is the key to learning (Lyle, 2008) . This talk, i.e. dialogicality, introduces children to higher order thinking.
As can be seen in Table 1, Wells (2007) deploys three distinct binary oppositions in order to illustrate the differences between monologic and dialogic discourse, and claims that education requires both monologic and dialogic interaction. However, Wells essentially focuses on how dialogic discourse offers opportunities for learning of a kind not generally supported by monologic discourse.
In terms of classroom context, monologic and dialogic talk can be conceptualised as binary opposites and while monologic talk focuses power on the teacher, dialogic talk creates a space for multiple voices and discourses that challenge the asymmetrical power relations based on monologic practices (Lyle, 2008) .
In Turkey, there are many studies examining Turkish maternal language (Küntay & Slobin, 1995, 1996, 2001, 2002; Küntay & Ahtam, 2004; Türkay, 2007; Cengiz, 2010, 2013; Cengiz & Çakır, 2012a, 2015) ; however very little research examined the properties of paternal language (e.g. Cengiz & Çakır, 2012b, 2016 ). Studies on fathers in Turkey usually focused on fathers’ attitude towards child care, perceptions of their fathering roles and their participation levels in child care (e.g. Kuzucu, 2011; Taşkın, 2011; Zeybekoğlu, 2013 ). There is almost no research comparing maternal and paternal language use compared to the number of studies in other countries. The purpose of the present study is therefore to compare mother and father language input with their preschoolers and focus on the link between parents’ SES and their dialogicality in the context of toy play.
To our knowledge, the relation between socioeconomic status, dialogicality and Turkish parents’ language use has not yet been investigated. Although there are some limitations, this is the first study in Turkey comparing maternal and paternal child-directed speech and the first attempt to analyze Turkish parents’ language input in relation to dialogicality. Thus, within the framework outlined so far, this study aims to provide an overview of linguistic development in the preschool years with a particular focus on parents’ language input and their dialogicality. It tries to build upon the previous research on parents’ language input and also extend the examination of the relation between SES and parents’ dialogicality in the context of toy play. The following research question guides this study:
Do the parents’ interaction styles differentiate in terms of dialogicality depending on their SES; thus, which children are encouraged more (by their parents) to become active speakers before they start their formal education?
2. Method
2.1. Participants
Ten fathers, ten mothers and their five-year old preschoolers participated in this study. Families were married, with both parents living in the home. They all lived in İzmir and were native Turkish speakers. Table 2 and Table 3 present and overview of the characteristics of the participants. Table 2 provides demographic information
![]()
Table 1. Two functions of discourse: monologic and dialogic
O’Connor & Michaels (2007: p. 276) .
![]()
Table 2. Demographic characteristics of mothers
![]()
Table 3. Demographic characteristics of fathers
on the mothers and Table 3 on the fathers contributing to this study. Low SES families had only a primary or secondary school education, and lived in suburban areas, whereas high SES families had completed four years of college or received a master’s degree/PhD, and lived in more affluent districts of İzmir. All children attended nursery schools in their neighborhoods on weekdays.
As illustrated in Table 2, mother participants ranged in age from 30 to 40 years, with a mean of 34.9 years. The average age of high SES mothers was 34.4 and that of low SES mothers was 35.4.
As shown in Table 3, the father participants ranged in age from 29 to 54 years, with a mean of 37.4 years. The average age of high SES fathers was 34.6 and that of low SES fathers was 40.2. While low SES fathers had only a primary or secondary school education, high SES fathers had completed four years of college (n = 3) or received a master’s (n = 1) or doctoral degree (n = 1). The fathers were employed outside of the home.
2.2. Data Collection Procedure
Participants were selected by means of purposive and snowball sampling methods. The reason for the small number of participants is that fathers had little or no incentive to attend, or were not available due to their working hours. Difficulties in recruitment of fathers were also caused by suspicion of a project on testing intelligence. Therefore, mothers were enlisted to encourage their partners to participate. Mothers’ encouragement and information on the study helped to ensure father participation. The participation in the study was voluntary and all 20 participants provided an informed consent prior to the study. After obtaining the participants’ consent to contribute to the research, the families were visited at home by the researcher.
Before the observational session, each mother and father was interviewed regarding her/his education, age and employment. The parents’ gender, age, education and monthly income were recorded. After the interview process, each mother-child dyad was invited into a separate room. Some toys were given and a picture where there was a bridge from building blocks was shown to the mother-child dyads. The toys used in this context were building blocks with a set of wooden blocks. They had different colors and shapes and contained a little bell, ramps and glass marbles. They were told to use the toy blocks either to build what they saw in the picture or that they were free to construct whatever they wanted.
Mothers were instructed to play with their children for about 15 - 20 minutes. The mother-child dyads sat on the floor during the play session. The researcher placed two voice recorders on the floor and was not present in the room during the recording in order to make the mothers and children feel comfortable. Recording started when they had settled in the room, which was one minute after they came into the room.
After the mothers played with their children, the fathers repeated the same procedure. Although they were playing with the same toys for the second time, the children were very interested in the toys and wanted to play again with their fathers.
2.3. Data Analysis
All mothers, fathers and their five-year old children showed an interest in the interactive toy play session. Each interaction lasted between 06:00 and 36:00 minutes (M = 15:00 min.). These interactions were taped and then transcribed. The unit of transcription was an utterance, so the speech of parents and children was coded as utterances. Two or more independent clauses occurring within the same conversational turn were considered as separate utterances. An utterance is defined as a conversational turn that contains one or more syntactic units and it is usually preceded and followed by a pause (Huttenlocher et al., 2010; Rowe, 2012) . The utterances were then coded as exchanges. An exchange is the minimal interactive unit and involves the negotiation of a single piece of information (Stenström, 1994) . Exchanges are independently observable entities. Each exchange was ascribed to the corresponding type of pattern; that is, each exchange was coded as an II, an IR or an IRF pattern. Sinclair & Coulthard (1975) developed a model of classroom discourse. They found a structure of three-part exchanges in the traditional school classroom: Initiation, Response, and Feedback, known as IRF. According to them, the following is the typical classroom discourse sequences:
T: What’s the capital of France? (Initiation)
S: Paris. (Response)
T: Yes, Paris. That’s right. (Feedback)
These three moves, the teacher’s initiation, the student’s response and the teacher’s feedback, consist of an exchange. IRF pattern, also called the triadic dialogue, is the dominant form of interaction in a classroom setting. The model proposed to show how interaction in the classroom takes place by taking a linguistic and functional look at discourse. The model was originally developed through the application of transcripts taken from classroom settings in the 1970’s. However, these settings were primarily teacher-centered which has led to some criticisms of the model. Nevertheless, IRF pattern has been widely accepted by the researchers as a useful category to analyze classroom discourse, and it remains a relatively powerful model that allows us to objectively evaluate communication that occurs in the classroom (Liu & Le, 2012; Nicholson, 2014) .
Focusing on the interaction between parents and their children, this study examined the exchange patterns of parents and their preschoolers, and tried to reveal the dominant interactional patterns among families coming from different socioeconomic backgrounds. Thus, the pattern was coded as an IRF where the mother or father made the initiation and feedback/follow-up moves while the child responded; the pattern was coded as an IR where the mother or father made the initiation, the child responded and there was no feedback/follow-up move, and lastly, the pattern was coded as an II where the parents made the initiation and when there was no response from the child, they continued with another initiation, so there was no speaker shift, thus no dialogicality. The use of IR pattern shows that the discourse was dialogic, at least there was a speaker shift, thus a space for children’s voices. The use of IRF, that is when there occurs an F move, the discourse was again dialogic and the F move has a great importance; research has shown that children always need some kind of feedback and this is the F move that supports and encourages children.
Since the sample size of the study was small, a nonparametric test, two-independent-sample test, Mann Whitney U was applied using IBM SPSS statistical packages (version 21.0). The critical alpha value was set at 0.05 for this investigation. Significant p-values (p < .05) are highlighted in grey in tables displaying scores.
3. Results
This study aimed to examine the link between mothers’ and father’s socioeconomic status and their language use with regard to their dialogicality. Figure 1 presents the raw numbers on total utterances, IR, II and IRF moves of mothers and fathers.
As seen in Figure 1, the analyses on the number of utterances revealed that fathers produced more utterances than mothers while playing with their children. In other words, fathers spoke to their children much more than mothers did. The IRF move is used extensively by high SES mothers and fathers. The results also revealed that the II structure was preferred by low SES mothers and fathers. The scores on the IR structure demonstrated similar results.
Table 4 presents the descriptive statistics of total utterances produced by mothers and fathers during toy play.
As shown in Table 4, high SES mothers’ total utterance score was Mdn = 221 while low SES mothers’ score was Mdn = 128. The difference between these two groups was not significant (p-value = .175). However, the difference between high SES fathers (Mdn = 300) and low SES fathers (Mdn = 212) was statistically significant (p-value = .047).
As to whether there was a significant difference between high SES mothers (Mdn = 221) and high SES fathers (Mdn = 300) in terms of their scores on the total utterances, Mann Whitney U results (p = .175 > .05) indicated that there was no significant difference between these two groups. There was also no difference between low SES mothers (Mdn = 128) and low SES fathers (Mdn = 212) in total utterances (p = .753 > .05).
Table 5 illustrates the descriptive statistics of the IR (Initiation-Response) moves of mothers and fathers during free play.
As presented in Table 5, high SES mothers’ score on the IR pattern was Mdn = 9 and low SES mothers’ score was Mdn = 3. The difference between mothers was not significant (p-value = .175). The difference between
![]()
Figure 1. Raw numbers of mothers’ and fathers’ total utterances, IR, II and IRF moves.
![]()
Table 4. Descriptive statistics of mothers’ and fathers’ total utterances
![]()
Table 5. Descriptive statistics on mothers’ and fathers’ IR (Initiation-Response) moves
high SES fathers (Mdn = 6) and low SES fathers (Mdn = 5) was not statistically significant either (p-value = .525).
There was no significant differences between high SES mothers (Mdn = 9) and high SES fathers (Mdn = 6), and (p = .207 > .05), also between low SES mothers (Mdn = 3) and low SES fathers (Mdn = 5) in terms of IR structure use (p = .344 > .05).
Table 6 shows the descriptive statistics of the II (Initiation-Initiation) moves of mothers and fathers during play time.
Table 6 illustrates that high SES mothers’ use of the II pattern was Mdn = 0 while low SES mothers’ score was Mdn = 11. The difference between high SES and low SES mothers was statistically significant (p = .007 < .05). Moreover, the difference between high SES fathers (Mdn = 2) and low SES fathers (Mdn = 17) was also statistically significant (p = .009 < .05). A significant difference was also found between high SES mothers (Mdn = 0) and high SES fathers (Mdn = 2) (p = .043 < .05). However, the findings on low SES mothers (Mdn = 11) and low SES fathers (Mdn = 17) did not reveal a significant score (p = .530 > .05).
Table 7 presents the descriptive statistics of the IRF (Initiation-Response-Follow Up) moves of mothers and fathers.
As for IRF pattern, a preliminary inspection of the median values showed a significant difference between the scores of high SES mothers (Mdn = 14) and low SES mothers (Mdn = 1; p = .008 < .05). There was also a significant difference between high SES fathers (Mdn = 36) and low SES fathers (Mdn = 0; p = .008 < .05). Additionally, a significant difference was found between high SES mothers (Mdn = 14) and high SES fathers (Mdn = 36; p = .047 < .05). No significant difference (p = .100 > .05) was observed between IRF moves of low SES mothers and fathers.
4. Discussion
This study aimed to compare maternal and paternal language use with a specific emphasis on their dialogicality. There was no significant difference between high SES and low SES mothers in terms of total utterance and IR moves, while significant difference was found on II and IRF moves. As for the high SES and low SES fathers, no significant difference was observed on IR moves; however the scores on total utterances, II and IRF moves revealed significant differences. In terms of gender, the comparison of low SES and high SES parents also demonstrated differences. No significant difference was found between high SES mothers and fathers on total utterances and IR moves, but the scores on II and IRF moves revealed a significant difference. The differences between low SES mothers and low SES fathers demonstrated no significant results on total utterances, IR, II and IRF moves.
Although the sample size was small, according to the findings, socioeconomic status of parents showed a difference in dialogicality, i.e. in the IR and IRF patterns, and also in the total amount of utterances. Since IRF consists of more multiple voices, it might be related with dialogic discourse, rather than monologic. Thus, families coming from high education and income produced more utterances and they tried to create more dialogic interactions with their preschoolers. They tended to give more chance to their children to talk during toy play interactions and encouraged their children to speak more before starting formal education. On the other hand, the power was more on the parents’ in families coming from low SES, so there was more II patterns where only parents talked. As for the importance of the F move, children need guidance in building knowledge. Particular talk moves, such as feedback/follow-up, support children; it is for this reason that the importance of increasing the use of dialogic practices is emphasized. Classroom settings where there were many IRF patterns are accepted
![]()
Table 6. Descriptive statistics on mothers’and fathers’ II (Initiation-Initiation) moves
![]()
Table 7. Descriptive statistics on mothers’and fathers’ IRF (Initiation-Response-Follow Up) moves
as teacher-centered, and they have come under criticisms; even sometimes IRF patterns have been charged with being monologic. However in our context, we accepted the IRF patterns dialogic. When compared to II patterns, where there are no speaker shift, IRF patterns can be accepted more dialogic. High SES families provide speaking opportunities for their children with the R moves, and they provide guidance with the F moves.
The findings of this study are consistent with previous findings. Heath (1983) and Ninio (1980) also found that parents from low SES talk much less to their children than high SES parents. According to Hoff, Laursen & Tardif (2002) , maternal language input to children changes depending on family SES; high SES mothers contributes more to child language development. Fish & Pinkerman (2003) and Hoff-Ginsberg (1998) also observed that child-directed speech in high SES families contained richer vocabulary of object labels. Low SES mothers, on the other hand, talked less and used less varied vocabulary.
Several limitations of this study should be noted. First, this study is limited in that it examined parents’ language use only in the context of toy play. Future research could include other activities such as book reading, picture reading, meal time etc. There is also a need for studies with an intervention program to raise parents’ awareness on their contribution to child language development. Another limitation of this study was the non-homogeneity of mothers’ and fathers’ personal characteristics and the fact that the sample size (n = 20) was small. Future multidisciplinary and longitudinal studies are needed to investigate maternal and paternal language input and their effects in different contexts with a greater number of participants. Since this research was a cross-sectional study and the sample was not representative especially with regard to parents’ personal characteristics, the parents cannot be generalized to the general population of parents. Thus, the results of this study should be interpreted with caution, and the limitations of the study should be borne in mind.
5. Conclusions
Most of our learning about the world we inhabit is through dialogue with others. The more parents give their children the opportunity to speak at home, the more children will be accustomed to speak at school. The development of ability to use school-based registers is crucially dependent on learners’ experience. Thus, particularly “learning through conversation with a responsive adult” should occur more often at homes. Through dialogic discourse, children can have the opportunities of becoming active speakers and thinking critically.
The findings of this study may shed light on the fact that the characteristics of the child-directed speech is influenced both by the gender of caregivers and by the socioeconomic backgrounds of parents. The present study can be regarded as having extended the existing Turkish literature on parents’ language use and their dialogicality. However, a larger group of participants is necessary to identify quantity and quality of language differences between parents.
Acknowledgements
This study is the result of great effort on the part of many people. I thank all the mothers, fathers and their children who took part in this study. Without their cooperation, this study would not have been possible.