CDS: a historical perspective

Linguists and developmental psychologists have reported on the phenomenon of Child-Directed-Speech (CDS) since the 1960s and 1970s. It is also referred to as motherese, parentese, or caregiver speech (Catell, Reference Catell2000, p. 104), and in the case of younger children it is known as Infant-Directed-Speech (IDS) (Ferguson, Reference Ferguson1964). There is now unequivocal evidence that CDS is a different register from Adult-Directed-Speech (ADS) and that it differs at all levels of the grammar and phonology (Cruttenden, Reference Cruttenden, Gallaway and Richards1994; Golinkoff et al., Reference Golinkoff, Can, Soderstrom and Hirsh-Pasek2015; Snow & Ferguson, Reference Snow and Ferguson1977; Soderstrom, Reference Soderstrom2007) and across different cultures (Broesch & Bryant, Reference Broesch and Bryant2015; Hilton et al., Reference Hilton, Moser, Bertolo, Lee-Rubin, Amir, Bainbridge, Simson, Knox, Glowacki, Alemu, Galbarczyk, Jasienska, Ross, Neff, Martin, Cirelli, Trehub, Song, Kim, Schachner and Mehr2022). Features of CDS include the use of slower speech or longer sound durations; higher and more varied pitch; more extreme/peripheral articulations; shorter utterances; simpler syntax; fewer false starts, hesitations and incomplete utterances; and baby words with simple syllable structure and reduplication (Aslin, Reference Aslin1993; Bernstein Ratner, Reference Ratner1984; Broen, Reference Broen1972; Drach, Reference Drach1969; Fernald & Mazzie, Reference Fernald and Mazzie1991; Fernald et al., Reference Fernald, Taeschner, Dunn, Papousek, de Boysson-Bardies and Fukui1989). From a theoretical perspective, these differences can be interpreted differently in generativist and constructivist approaches to language acquisition. From a nativist approach to language (Chomsky, Reference Chomsky1981; Chomsky & Halle, Reference Chomsky and Halle1965) the differences are inconsequential as input is not considered to provide children with the deep underlying structure of the complex grammar that they need to acquire, but rather just contains enough material to set the parameters that are innate (e.g., Hyams, Reference Hyams1996). Input-based theories (e.g., Tomasello, Reference Tomasello2000), on the other hand, emphasize the importance of input and argue that caregivers simplify the language addressed to their children in order to aid acquisition. This entails the maximization of linguistic contrasts which are important for the language in question. Proponents of this approach point to the correlation between the frequency of certain structures in the adult input and early child forms, as well as the way children use structures in the early stages (Pine & Lieven, Reference Pine and Lieven1997; Theakston et al., Reference Theakston, Lieven, Pine and Rowland2001). Studies have also demonstrated how CDS changes over time to adjust for the child’s growing competence (Kitamura & Burnham, Reference Kitamura and Burnham2003; Niwano & Sugai, Reference Niwano and Sugai2002; Soderstrom, Reference Soderstrom2007).

A pre-requisite question to this debate relates to whether CDS is actually used by caregivers to facilitate language development, or whether its primary function is socially-driven. Some studies suggest that the main function of CDS is to convey affection and foster social interaction (Singh et al., Reference Singh, Morgan and Best2002; Tamis-LeMonda et al., Reference Tamis-LeMonda, Kuchirko and Song2014). In other words, it is the caregivers’ way to express positive emotion to their child, encourage social interaction and foster joint attention to objects, places and events. Since this involves verbal as well as non-verbal means of communication, the fact that this in turn focuses the child’s attention on language is a side-effect rather than the main aim under this view. The attention to the modified speech signal has been shown to aid the child in speech segmentation (Jusczyk et al., Reference Jusczyk, Luce and Charles-Luce1994; Seidl, Reference Seidl2007; Thiessen et al., Reference Thiessen, Hill and Saffran2005), a pre-requisite for learning words, as well as to play a role in the development of various receptive and productive oral skills (Broen, Reference Broen1972; Hart & Risley, Reference Hart and Risley1995; Soderstrom et al., Reference Soderstrom, Blossom, Foygel and Morgan2008; Song et al., Reference Song, Demuth and Morgan2010).

While the majority of early work on CDS was carried out with WEIRD (Western Educated Industrialized Rich and Democratic) communities (Henrich et al., Reference Henrich, Heine and Norenzayan2010), Cameron-Faulkner et al. (Reference Cameron-Faulkner, Lieven and Tomasello2003) point to the fact that the register these families used cannot be necessarily generalized to other populations and cultures. For instance, caregivers from some non-Western cultures have been found to speak less frequently to their children (e.g., Lieven, Reference Lieven, Gallaway and Richards1994; Schieffelin & Ochs, Reference Schieffelin and Ochs1986; Shneidman & Goldin-Meadow, Reference Shneidman and Goldin-Meadow2012) and to use various speech acts in different quantities (Harkness, Reference Harkness, Snow and Ferguson1977; LeVine et al., Reference LeVine, Miller, Richman, LeVine, harkness and Super1996; Rabain-Jamin, Reference Rabain-Jamin2001; Vogt & Mastin, Reference Vogt and Mastin2013). Caregivers from different cultures stimulate their offspring through different activities to foster development in particular domains (Bornstein & Putnick, Reference Bornstein and Putnick2012; Keller, Reference Keller and Valsiner2012; LeVine et al., Reference LeVine, Miller, Richman, LeVine, harkness and Super1996), e.g., book reading, story-telling, counting and object labeling to foster cognitive skills (Bornstein & Putnick, Reference Bornstein and Putnick2012); physical activity such as walking to stimulate motor skills (Keller, Reference Keller and Valsiner2012); and singing, playing with other children and going outdoors to stimulate socioeconomic skills (Bornstein & Putnick, Reference Bornstein and Putnick2012). These different activities lead to different conversational settings, which in turn influence language development. For instance, the amount of directives or imperatives has a negative relation with children’s grammar and vocabulary development (Newport et al., Reference Newport, Gleitman, Gleitman, Snow and Ferguson1977) while book reading contains more object labeling and questions (Hoff, Reference Hoff2006). Within the same (non-Western) community, a change in geographic location and lifestyle can have consequences for interactions with children and what is expected of them. For instance, Vogt and Mastin (Reference Vogt and Mastin2013) found CDS use in urban Mozambique to more often convey socio-emotional than cognitive intention, and the opposite was true for CDS by parents living in rural Mozambique. This emphasizes the importance of scrutinizing the role of parental styles within and across cultures. The following section will delve into the nuances of Arabic CDS, building upon the current understanding of this subject.

CDS in Arabic

Research on Arabic CDS is rare. Studies in this domain typically adopt a cross-linguistic framework in order to identify universal linguistic features of CDS (e.g., Ferguson, Reference Ferguson1964), to address methodological issues in its investigation (Haggan, Reference Haggan2002), or to exemplify use of child directed speech forms in political discourse (e.g., Płonka, Reference Płonka2012). While several universal simplifications at the phonological, lexical, grammatical as well as stylistic features have been identified in CDS and reviewed in the preceding section, one less universally identified feature of CDS is Gender-shift. It was first reported in Arabic and Marathi by Ferguson (Reference Ferguson1964), as illustrated in example 1 below, which is based on real-life observation made by the first author witnessing a colleague interact with her 2-year-old boy.

As opposed to

In Gender-shift “a feminine noun, pronoun, adjective, or verb form is used in reference to a boy or vice versa” (Ferguson, Reference Ferguson1964, p. 106). It is assumed that these shifts are used for expressing endearment (Tobin, Reference Tobin, Hellinger and Bubmann2001), but little is known about any role they may play in language learning. A potentially unintended effect of Gender-shift in CDS is that of increasing or decreasing linguistic complexity, depending on the sex of the child: applying a Gender-shift when interacting with a female child frequently leads to a decrease in the linguistic complexity of the shifted form since the male forms are more frequently used, considered to compound the ‘default’ uninflected form, and exhibit higher levels of matching form and function ‘designations’. This assumption is made based on robust empirical evidence from computational linguistics demonstrating that masculine forms are more frequent and less complex than feminine forms in Arabic (e.g., Alhafni et al., Reference Alhafni, Habash and Bouamor2022; Alkuhlani & Habash, Reference Alkuhlani and Habash2011). Alkuhlani and Habash (Reference Alkuhlani and Habash2011) conducted the first quantitative analysis of gender and number agreement in the Arabic nominal system including nouns, proper nouns, and adjectives with a corpus containing 16.6K sentences and over 400K tokens with diacritics. In their analysis, Alkuhlani and Habash (Reference Alkuhlani and Habash2011) highlight the distinction between form and function gender designation and the occurrence of discrepancy between formal and functional designation in the Arabic nominal system. For example, they present the word حامل /ħamǝl/ ‘pregnant’ to have a masculine form designation and a feminine function designation (MS/FS) while the word حمرا /ħamɹa/ (red-fem.) is a masculine templetic form with a feminine function (MS/FS). The study results revealed that masculine forms were more frequent both in form and function and that matching form and function designations was more evident for the masculine forms (M/M = 62.5) than feminine forms (F/F = 28.9) which supports the general premise that masculine forms are generally simpler than feminine forms.

Similar findings showing that masculine forms are more frequent than feminine forms were reported in Alhafni et al.’s (Reference Alhafni, Habash and Bouamor2022) analysis of a 58035 first and second person sentences and 424,254 word corpus. The sentence level statistics analysis of the original corpus had 4% of sentences with masculine first person markers and 2.2% with feminine first person markers, whereas second person masculine was annotated for 21.7% of the sentences and second person feminine for 9.3% of the corpus. A word level statistical analysis showed that feminine was annotated for 2% and masculine was annotated for 4.5% of the corpus. These results provide further empirical evidence that masculine forms are the unmarked, more frequent, and less complex forms in Arabic. Although there are no theoretical accounts of this phenomenon yet, this work provides detailed data on the frequency of masculine and feminine forms in Arabic. The skew towards more complex forms in Arabic feminine morphosyntactic agreement based on the available computational data allows us to test potential linguistic simplification associated with CDS. Specifically, the higher complexity and lower frequency of feminine forms, combined with a semantic mismatch, should discourage their use with male children. For female children, however, semantic concurrence could offset the increased complexity of grammatically female forms.

On the other hand, studies on Gender-shift in adults propose that it is used to endorse affection (Tobin, Reference Tobin, Hellinger and Bubmann2001); if this is relevant in child-caretaker interactions too, we would expect to find its presence in CDS irrespective of the child’s sex and to be correlated to parenting styles that exhibit higher levels of endearment. Parenting styles are little-studied in CDS use and function when it influences the nature of caregiver-child interaction and potentially the resulting messages conveyed in CDS (e.g., Tarabeih, Reference Tarabeh2013). We turn to this next and incorporate it into our examination of the patterns of CDS exhibited in Arabic-speaking parents.

Parenting style and CDS interactions

A central and widely used typology in understanding parenting styles in Western cultures is the one developed by Baumrind (Reference Baumrind1967, Reference Baumrind1971, Reference Baumrind1991; see also Berg-Cross, Reference Berg-Cross2000). Baumrind’s typology characterizes typical parenting styles based on their demandingness and responsiveness into three different categories. (1) Authoritative parenting style consists of the three positive dimensions autonomy granting, warmth/support, and regulation. Autonomy granting refers to the ability of parents to allow their children a high level of psychological freedom and to have democratic participatory interactions with them. “Warmth/support” refers to the parents’ ability to accept their children with emotional warmth and support them and their choices in different situations. Regulation refers to the parents’ behavioral control that sets clear and consistent limits on the child’s behavior through inductive thinking about rules and determining consequences for inappropriate behavior (Matejevic et al., Reference Matejevic, Todorovic and Jovanovic2014; Sharma & Sandhu, Reference Sharma and Sandhu2006). (2) Authoritarian parenting style consists of the three negative dimensions physical coercion, verbal hostility, and non-reasoning punitiveness. Physical coercion describes the parents’ use of physical punishment such as slapping in order to control and discipline the child. “Verbal hostility” describes the parents’ use of abusive hostility to control, discipline, or frighten the child. “Non-reasoning punitiveness” describes the tendency of parents to punish their children without justification or reasonable reasoning (Matejevic et al., Reference Matejevic, Todorovic and Jovanovic2014; Sharma & Sandhu, Reference Sharma and Sandhu2006). (3) Permissive parenting style consists of two dimensions, autonomy granting and indulgence. “Indulgence” describes parents’ tendency to give children too many resources, do things for children that they should be doing for themselves, or set few rules or consequences for children’s behaviors (Clarke et al., Reference Clarke, Dawson and Bredehoft2004). A fourth parenting style has been added, the uninvolved parenting style. This consists of two negative dimensions, non-reasoning punitiveness and indulgence, which were defined earlier (e.g., Maccoby & Martin, Reference Maccoby and Martin1983). It is currently referred to as one of the four main parenting styles (Sorkhabi, Reference Sorkhabi2012).

The Parenting Styles and Dimensions Questionnaire (PSDQ) was designed by Robinson et al. (Reference Robinson, Mandleco, Olsen, Hart, Perlmutter, Touliatos and Holden2001) and found to be a valid measure to examine the four parenting styles. The PSDQ has been widely used in large samples and various cultures (e.g., Kern & Jonyniene, Reference Kern and Jonyniene2012; Topham et al., Reference Topham, Hubbs-Tait, Rutledge, Page, Kennedy, Shriver and Harrist2011). The Parenting Styles and Dimensions Questionnaire (PSDQ) was adapted and validated for Arabic, specifically focusing on Palestinian families, the target population of this study with high levels of internal consistency (Yaffe, Reference Yaffe2018). Studies using PSDQ (e.g., Brosnan et al., Reference Brosnan, Kolubinski and Spada2020; Davis et al., Reference Davis, Coleman and Kramer2021; Neel et al., Reference Neel, Slaughter, Stark and Maitre2019) treat the four PSDQ subscales as separate, continuous measures, with higher scores indicating more behaviors consistent with each parenting style respectively (rather than a single category). Hence, we focus on the seven dimensions of parenting behaviors that yield responsiveness and control rather than requiring parents fit into narrow classifications of “parenting styles”. These seven dimensions, which have been defined earlier, are autonomy granting, warmth and support, regulation, non-reasoning punitiveness, indulgence, verbal hostility, and physical coercion.

While research on CDS has pointed to cultural differences in caregiver practice (e.g., Sorkhabi, Reference Sorkhabi2012), we are not aware of studies looking at how parental styles may affect Gender-shift use in a particular culture. If we assume that caregiver input does not only bear linguistic based meanings, then examining parenting style dimensions may add depth and breadth to our understanding of CDS. Hence, we argue that incorporating parenting style and its interaction with caregivers’ linguistic productions is needed for comprehensive understanding of CDS in the child’s language, cognitive, and emotional development. Pioneering work has focused on the mental and emotional input infants are exposed to in relation to maternal mind-mindedness and its potential effect on infant attachment in Arabic (Tarabeh et al., Reference Tarabeh, Zreik, Oppenheim, Sagi-Schwartz and Koren-Karie2019). Maternal Mind-Mindedness refers to mothers’ attunement to their infant’s mental states, including their thoughts, feelings, motives, and goals, and is evidenced in verbal references to the mental states that might be governing the child’s behavior (Meins, Reference Meins1997). Such references are termed mind-related comments (Meins, Reference Meins2013) and can be assessed from mothers’ child-directed speech during mother–infant interactions (Meins et al., Reference Meins, Fernyhough, Fradley and Tuckey2001). Mind-related comments are judged as appropriate when they match the child’s behavior or as non-attuned when they are not matched to the child’s behavior (Meins & Fernyhough, Reference Meins and Fernyhough2015). Tarabeh et al. (Reference Tarabeh, Zreik, Oppenheim, Sagi-Schwartz and Koren-Karie2019) revealed that mothers of secure infants used more appropriate and less non-attuned mind-minded comments than mothers of insecure-disorganized infants.

No study in any language has so far incorporated examination of Gender-shift within the linguistic input directed at infants by their parents, while controlling for parenting styles. This might be due to the difficulty in collecting such data or the fact that most studies are conducted in languages that do not present the use of Gender-shift. Hence, examining the use of Gender-shift in CDS in Arabic may contribute to our understanding of the nature of these productions in CDS and the underlying function and role of CDS in language acquisition in general. We hypothesize that the use of Gender-shift as an index of endearment within child-caretaker interactions would be more evident in the input of parents who exhibit increased linguistic responsiveness in their parenting style (i.e., authoritative and permissive styles) and less in those parents with decreased responsiveness (i.e., authoritarian and uninvolved styles).

Even though Gender-shift in CDS was reported by Ferguson in 1964, there have been no recent studies of its use. Most of the available investigations of Gender-shift have focused on its presence in adult interactions (e.g., Wilmsen, Reference Wilmsen and Suleiman1999; and for Hebrew, Tobin, Reference Tobin, Hellinger and Bubmann2001). These studies support the notion that Gender-shift is used in adult interactions for endearment. Hence, it is logical to believe that Gender-shift would be used for the same function with infants/children too; however, due to the fact that it incorporates presentation of more complex linguistic structures when addressed to male infants/children, observable variation may be exhibited in its use based on the child’s sex too.

The aim of this study is to explore whether the use of Gender-shift in child-directed-speech in Palestinian Arabic has potential linguistic motivation and/or is used as a tool to express affection and endearment. If Gender-shift in CDS plays a linguistic simplification role, it would be present more in interactions with female infant/children and less with male infant/children, given the presumed different complexity levels of the feminine and masculine forms in Arabic. Moreover, Gender-shift in CDS (and CDS in general) may have an affective function for expressing endearment. This should correlate with greater use of Gender-shift in parenting styles characterized by increased responsiveness levels, namely authoritative and permissive parenting styles. This is the first study to investigate the phenomenon of Gender-shift in CDS in Arabic using an interprofessional framework within which parenting styles and demographic factors are accounted for. We therefore asked the following research questions:

One challenge in studying Gender-shift is the difficulty in eliciting it in naturalistic interactions and very large corpora. In fact, preliminary analysis of pilot child-adult play sessions showed no presence of Gender-shift within these samples (Tarabeh, Reference Tarabeh2013). It is hard to establish whether this was due to the presence of a fieldworker or the low frequency of occurrence of Gender-shift constructions in general. We therefore decided to survey parents on the reported use of Gender-shift in their CDS using a large participant pool and controlling for potentially relevant factors (gender, age, and parenting style).Future studies would then build on our results and consult dense corpora for a better chance at finding robust patterns in naturally occurring interactions.