Mother and father language input to young children: Contributions to later language development

Abstract

There has been little research comparing the nature and contributions of language input of mothers and fathers to their young children. This study examined differences in mother and father talk to their 24 month-old children. This study also considered contributions of parent education, child care quality and mother and father language (output, vocabulary, complexity, questions, and pragmatics) to children's expressive language development at 36 months. It was found that fathers' language input was less than mothers' language input on the following: verbal output, turn length, different word roots, and wh-questions. Mothers and fathers did not differ on type-token ratio, mean length of utterance, or the proportion of questions. At age 36 months, parent level of education, the total quality of child care and paternal different words were significant predictors of child language. Mothers' language was not a significant predictor of child language.

Introduction

Broad changes over the last 30 years in family structure and functioning, including more women in the workforce, extended and less predictable work schedules, and the changing role of men in families, have highlighted the need to include both mothers and fathers in research concerning familial influences on child development. These changes in family structure and functioning have contributed to new ways of conceptualizing the role of fathers in understanding children's development. In previous studies fathers have been found to influence children's development through provisioning of resources and investment in the family (for reviews, see Amato, 1998, Marsiglio et al., 2000), but it is necessary for current models of father influence to consider the role father/child relationships in influencing children's development (Lamb, 1997, Marsiglio et al., 2000) and psychological well-being (Amato, 1998). Children, and especially young children, may benefit from interacting with two involved parents who may have different but complementary behavioral styles (Cabrera et al., 2000, Guzell and Vernon-Feagans, 2004).

The possible influence of both mothers and fathers may be especially important when children are very young and acquiring cognitive, language and social skills that undergird later development. Language develops in the context of social relations and has roots in the early interactions of children and caregivers (Bates, 1976, Bruner, 1981, Locke, 2001, Tomasello, 1992). Early language development occurs within, and is affected by, a variety of contexts (Bronfenbrenner, 1979), including familial and nonfamilial environments. Ecological systems theory argues that social interactions with caregivers are examples of proximal processes that are linked to early language development (Bronfenbrenner & Morris, 1998). To gain a more complete understanding of early language development, it is necessary to consider sources of influence from multiple caregivers. For many contemporary families, such influential caregivers include mothers, fathers and nonfamilial child care providers. This research aims to examine the differences between mother and father language input in a group of dual-earner families where fathers have major responsibility for their children and therefore may influence their children's language. This study is also aimed at a more complete consideration the contributions of multiple caregivers (including mothers, fathers, and child care providers) to children's early language development.

Several studies have found links between SES and early language development (Fish and Pinkerman, 2003, Hart and Risley, 1995, Hoff-Ginsburg, 1998). Hoff-Ginsburg (1998) found effects of SES even when looking at middle-class and upper-middle-class samples in that children from high-SES families used a richer vocabulary of object labels than did children from mid-SES families.

One way maternal education may influence early language development is through maternal language input (Hoff-Ginsberg, 1991). Previous research has found links between SES and maternal vocabulary and mean length of utterance during mother–child interactions (Hart and Risley, 1995, Hoff, 2003). Hammer and Weiss (1999) found that among 12 African American mothers from low- and mid-SES groups, mothers in the mid-SES group included more language goals in their play, and labeled, commented and imitated children's vocalizations more often than did mothers in the low-SES group. The authors hypothesized that observed group differences may be related to mothers' educational experiences. That is, mothers with higher levels of education might use more behaviors that are valued in school, such as play that incorporates more language goals.

While these findings suggest the possible importance of maternal level of education to both maternal language input and child language development, many studies that have controlled for maternal level of education or SES when considering the impact of parental language input on early language development have considered primarily low-income families (Pan et al., 2005, Weizman and Snow, 2001). Further, very few studies have controlled for the impact of paternal level of education on father language input to their young children. The present study simultaneously considers the impact of both mother and father language input on children's early language development, controlling for the possible impact of maternal and paternal level of education in a relatively homogenous sample of dual-earner middle-SES families.

While research on caregiver language input has traditionally focused on the influence of mothers' language use, the communicative environment of young children is much broader. Extra-familial factors, such as the quality of center-based child care, have also been found to be associated with children's language development, even as early as one year of age (Burchinal, Roberts, Nabors, & Bryant, 1996).

Several studies have indicated that young children in higher quality child care have stronger expressive and receptive language skills (Feagans et al., 1995, McCartney, 1984, NICHD Early Child Care Research Network, 2002). McCartney's (1984) study of child care in Bermuda found that overall quality of care was predictive of early language development. Specifically, children from centers with high levels of caregiver speech performed better on tests of language development than did children from centers with high levels of peer speech. The connection between quality of center-based care and language development during the first 3 years of life has been supported by more recent studies with diverse populations, which have found evidence that higher quality child care predicts higher measures of language development (Burchinal et al., 2000). Specifically, ratings of caregiver responsiveness, sensitivity, and frequency of language stimulation have been linked with early language development (NICHD Early Child Care Research Network, 2000).

The quality of child care may also be related to mother/child language in the home. Vernon-Feagans, Hurley, and Yont (2002) found that among 41 four year-old children enrolled in full time day care, those children from high-quality child care used more words and utterances in a picturebook task with their mothers. They also found that children in high-quality daycare were buffered against many of the negative effects of otitis media on their language skills.

The relationship between the quality of child care and child language has been found even after adjusting for selected child and family characteristics such as SES (Burchinal and Cryer, 2003, Burchinal et al., 2000). It is important to note that whereas the quality of child care accounts for a significant amount of the variance in children's language outcomes, family factors have been found to be more important predictors of these same outcomes (NICHD Early Child Care Research Network, 2000). Therefore, it appears important to consider the family context of language development, even for those children enrolled in full-day child care.

There is a growing but still modest literature on the relationship between mother and father behavior with their children and children's later development. A number of studies have suggested that mothers and fathers engage in different types of interactions with their children. Specifically, it has been reported that fathers spend a greater proportion of their time interacting with their children in play activities, and that their play is more physical than that of mothers (Parke, 2002, Yeung et al., 2001). Gottman (1998) found that in their interactions with children, mothers were more verbal and directing, whereas fathers were more physical and arousing. Although the description of these differences was important, further research is needed to understand whether mother and father styles of interaction are linked to specific outcomes for children. There have been a host of studies that have confirmed that mothers' behaviors are linked to concurrent and later child outcomes in language, cognition, and social behavior (Cassidy, 1988, Cummings and Davies, 1994, Hart and Risley, 1995, Huttenlocher et al., 1991, Olsen et al., 2002). Recent research has also found that fathers' supportive parenting behavior during parent–child interactions at 24 months was predictive of children's language development at 36 months (Tamis-LeMonda, Shannon, Cabrera, & Lamb, 2004). However, very few, if any, studies have considered the possible contributions of father language input to child outcomes.

To date, there has been limited research in the area of comparing mother and father language, and most existing research in this area was conducted nearly two decades ago. These studies generally were limited by a small sample size and the results of the studies were not always consistent with one another. Nonetheless, they lay the basis for the present study.

Overall, the studies comparing mother and father verbal input to their children have found that fathers, like mothers, engaged in language that was attuned to their children. Fathers adopted a simplified speech register, spoke with a higher pitch, and adjusted their language output in response to changes in their children's language (Fernald et al., 1989, Kavanaugh and Jirkovsky, 1982, McRoberts and Best, 1997, Rondal, 1980). In a meta-analysis of mother/father language use with their children from infancy to middle childhood, Leaper, Anderson, and Sanders (1998) created global categories of input. They found that during parent–child interactions, fathers used less total language, less supportive language, less negative language and more directive and informing language than did mothers. There were no differences in the use of questions or requesting information. Leaper et al. found that effect sizes associated with mother/father differences were larger with infants and toddlers than with older children.

In the present study we consider five broad areas of mothers' and fathers' language input to toddlers and only review studies where a language sample was coded. These five areas are not meant to be exhaustive but are representative of the important language variables that have been studied and shown to influence children's development. These five areas are: parental total verbal output, parental diversity of vocabulary, parental utterance complexity, parental use of questions, and parental pragmatic features of the discourse.

Many studies have found that fathers produced less total verbal output than did mothers in interaction with their children, although there was a context effect for some studies. Both Golinkoff and Ames (1979) and Hladik and Edwards (1984) examined mother and father language in a dyadic and triadic situation. Golinkoff and Ames studied 12 middle-class children in 10-minute free play sessions. They found that in the dyadic situation, mothers and fathers produced a similar number of utterances. On the other hand, in the triadic situation the fathers used fewer utterances than mothers. Hladik and Edwards included 10 middle-class children who participated in 30-minute sessions with their parents alone and together. They found exactly the opposite finding as Golinkoff and Ames: in the triadic situation, there were no differences between mothers and fathers on any variables, but in the dyadic situation, fathers spoke less than mothers did. Rondal (1980) studied five French speaking middle-class families in dyadic free play and story-telling situations and triadic meal situations when the children were 1;6 to 3 years of age. During the five recording sessions, fathers spoke fewer words than mothers.

In contrast, other studies have not found significant differences between mothers and fathers on measures of total output (McLaughlin et al., 1983, O'Brien and Nagle, 1987). A recent study by Rowe, Coker, and Pan (2004) compared fathers' and mothers' talk to toddlers in 33 low-income families. They reported that, in dyadic interactions with their toddlers, mothers and fathers did not differ in the amount of child-directed language. In one of the only previous studies to consider dual-earner middle-class families, Malone and Guy (1982) also found that in a sample of 10 families, fathers and mothers did not differ in total output during dyadic interactions with their 3-year-old sons.

Surprisingly few studies have compared mothers' and fathers' diversity of vocabulary in interactions with their young children. O'Brien and Nagle (1987) studied 10 middle-class European American children and their parents in a free play situation. The children ranged in age from 1;6 to 2 years of age. Rowe et al. (2004) studied 33 Early Head Start children and their parents when the children were 2 to 2½ years of age. These two studies compared mothers and fathers in terms of diversity of vocabulary and found no significant differences. Ratner (1988) studied eight mother–father–child triads, aged 1;6 to 2;0, in free play situations. Ratner found only minor differences between mothers' and fathers' type-token ratio, but found that fathers' speech was characterized by greater use of rare vocabulary and lower use of common vocabulary than was mothers' speech.

There have been a number of studies that examined the language complexity of mothers and fathers although there is little consensus among the results. McLaughlin et al. (1983) studied 24 middle-class European American children; eight at 18 months, eight at 2½ years of age, and eight at 3½ years of age. They found an overall effect for a shorter MLU for fathers. Rondal (1980) studied five boys who were between 1;6 and 3 years of age. This study also found a shorter MLU for fathers. Malone and Guy (1982) studied 10 first-born 3½-year-old boys from middle-class dual-earner families. They found that fathers had a shorter MLU than mothers and in addition that the five longest utterances of the fathers were also shorter than the mother's five longest utterances.

However, there were six studies that did not find utterance complexity differences between mothers and fathers (Golinkoff and Ames, 1979, Hladik and Edwards, 1984, Hummel, 1982, Kavanaugh and Jirkovsky, 1982, O'Brien and Nagle, 1987, Rowe et al., 2004). These studies were similar to the studies that did find differences with respect to sample characteristics. For instance, these studies also had small sample sizes from 4 to 33 children, similar ages of children from 8 months to 3½ years of age, and generally included middle-class European American families.

Some previous research has suggested that fathers may ask proportionally fewer total questions of toddlers, more wh-questions (open-ended) questions, and fewer yes/no questions than do mothers (Leaper et al., 1998, McLaughlin et al., 1983, O'Brien and Nagle, 1987, Rondal, 1980, Rowe et al., 2004). In contrast, Malone and Guy (1982) found that fathers had a lower proportion of total questions, a lower percentage of wh-questions, and a higher percentage of yes/no questions than did mothers. Again, other studies have not found a significant difference between fathers and mothers in terms of use of questions (Golinkoff and Ames, 1979, Hladik and Edwards, 1984, Hummel, 1982, Kavanaugh and Jirkovsky, 1982, Kruper and Uzgiris, 1987).

A common conversational variable that has been measured in some studies comparing mother and father language input has been conversational turns. While some studies have found that fathers took fewer conversational turns than did mothers (Golinkoff and Ames, 1979, Rondal, 1980), other studies have not found significant differences between mothers and fathers in terms of conversational turns (e.g., McLaughlin et al., 1983). Although there were contextual differences among these studies with respect to the way in which the language samples were obtained, there did not appear to be any consistent patterns that could explain these different findings, except the small sample size.

Most existing work on father language input has focused on the comparison of father and mother language input. However, due possibly to small samples sizes, previous research in this area has not yielded consistent results. The current study included a large sample of dual-earner families to compare mother and father language input during free play interactions when their children are 2 years of age. This study further extends the existing focus on father language input to consider the direct contributions of father language input on early language development.

Research has shown that parents modify their speech to their young children in ways that support their early language learning through, for example, simplified language that is less complex grammatically, more redundant, and with a higher pitch and exaggerated intonation pattern (Fernald et al., 1989, Kavanaugh and Jirkovsky, 1982, Kitamura and Burnham, 2003, McRoberts and Best, 1997, Rondal, 1980, Snow, 1977). Previous research linking parental language behavior to early child language development has focused primarily on mothers. Results of these studies have suggested that maternal output, vocabulary, complexity of speech and questions may contribute in important ways to children's early language development. However, many of these studies have not controlled for parent level of education. Additionally, most previous studies have considered families in which mothers stayed at home full-time and therefore, did not use center-based nonfamilial child care.

Previous research into maternal language has indicated that the amount of talk mothers direct to their children was associated with their children's gains in linguistic abilities (Barnes et al., 1983, Furrow et al., 1979, Huttenlocher et al., 1991). Barnes et al. (1983) considered a sample of 32 children from diverse socioeconomic backgrounds during the second year of life. They found that amount of adult speech was significantly correlated with gains in children's early language skills, in that children who made larger gains in language development also experienced greater amounts of adult speech. Huttenlocher et al. (1991) found similar results with a sample of 22 children, from 14 to 26 months of age. They found that the overall amount of maternal speech input at 16 months accounted for a substantial amount of variation of children's acceleration in vocabulary growth.

The diversity of maternal vocabulary has been a strong predictor of children's later language development and literacy (Bornstein et al., 1998, Hart and Risley, 1995, Hoff-Ginsberg, 1991). Hoff-Ginsberg (1991) found maternal vocabulary to be significantly associated with child vocabulary in both working-class and upper-middle-class mother–child dyads. Similarly, Hart and Risley (1995) found that the number of different words used by parents was strongly related to children's early vocabulary.

While still controversial, the complexity of maternal language input to young children has been associated with young children's language development (Furrow et al., 1979, Gleitman et al., 1984). This finding has not been supported by other studies (Barnes et al., 1983), which have not found significant links between the complexity of maternal speech and child language development.

In a study of 22 mother–child dyads, Hoff-Ginsberg (1985) examined the role of mothers' speech and children's syntactic growth during the second year of life. Hoff-Ginsberg found that frequency of mothers' open-ended, or wh-questions was a significant predictor of children's auxiliary growth. Similarly, Rowland, Pine, Lieven, and Theakston (2003) found in a sample of 12 English children ages 1;8–3;0, that the frequency of mothers' use of particular wh-words and verbs were predictive of children's order of acquisition of wh-questions. In addition, mothers' use of yes/no questions has been related to greater language achievement in children during the 2nd year of life (Furrow et al., 1979).

Previous research suggests that multiple aspects of children's familial and nonfamilial environments influence early language development. This body of research suggests that the interactions between children and adults within and outside of the family shape early language development. While there is considerable research on maternal language input, to date very few, if any, studies have considered the contributions of father language input to early language development. The present study aims to more fully include fathers in the discourse surrounding parental language input. Previous studies on father versus mother input to young children have found some differences, but many of these findings are not consistent from study to study. These discrepancies may be due to the limited sample size in most of the studies. In addition, most of the studies were done almost 20 years ago with families in which the mother generally did not work outside the home. None of the studies examined the possible influence of father language input on the child's later development. A goal of this study is to develop a model predicting children's expressive language development at 36 months of age that controls for the effects of parent level of education when considering the influence of quality of care and mother and father language input.

The present study builds upon the limited previous research on mothers' and fathers' language input to consider triadic language interactions with a large sample of 92 dual-earner families. It extends the previous body of work in this area by examining multiple sources of influence in the familial and nonfamilial caregiving environments on children's early language development. This study addresses the following research questions:

• How does the language input of fathers differ from that of mothers during triadic free play sessions in the home when their children were 24 months old?

• What are the unique contributions of parent education, quality of child care, and fathers' and mothers' language input when their children were 24 months old in the prediction of children's expressive language development at 36 months of age?