A longitudinal project on early language development was carried out by recruiting 186 families. The inclusion criteria for the present study were that the child was born on time (gestational age > 37 and birthweight >2,500) and was taken to the baby lab at least once. All infants included in the present investigation were of low social risk and were the first-born children of their mothers. As is common in developmental studies with voluntary participants, mothers with higher education were overrepresented, with 75% having college or university degrees. All participants came from metropolitan (Budapest) or agglomeration areas, all children were monolingual. No hearing problems were reported. The sample was ethnically homogeneous Caucasians of Hungarian origin. Families were recruited at the infant’s birth, 4, 9, and 18 months of age (see Table 2). All families received detailed information on the study, and informed consent was obtained. The first wave included 74 middle-class mothers recruited in the HONVED PMC hospital’s maternity ward. Data were collected up to 18 months in this phase. An additional recruitment at 18 months was planned to increase the sample size continuing to the second phase, however, as the dropout rate was higher than previously expected due to the COVID-19 epidemic, additional recruitment of 4- and 9-month-old infants was carried out (see Table 2). The sex ratio and infant characteristics in the participating and the dropout families did not differ significantly. The present data set includes varying numbers of infants at different ages due to the disruption caused by the pandemic breaking out during data collection and preventing families from visiting the child laboratory. The exact numbers of available data at each age are presented in Table 2.

According to the original protocol, mothers were to fill in the questionnaires in the baby lab while lab assistants played with the children and administered tests in the passive presence of the mother. However, the Covid-19 pandemic resulted in some mothers completing the questionnaires from their homes online. There were no significant differences in temperament or language between the questionnaires administered before and during the pandemic.

Ethical approval for the study was granted by ETT-TUKEB (1942-12/2016) and EPKEB (77/2015).

Means and standard deviations for the whole sample, and for boys and girls separately are presented in Table 4. The age-related growth of expressive vocabulary between 9 and 30 months is depicted in Figure 1.

One-way ANOVA or Kruskal-Wallis tests were applied to assess sex differences. No gender differences were observed in gestational age, but boys were significantly heavier at birth [H(1, n = 179] = 20.394, p < 0.001). Likewise, no significant differences between boys and girls appeared on temperament scales at any age, apart from a statistical trend towards boys scoring higher on Surgency at 9 months [F(1,119) = 2.891, p = 0.092]. However, significant sex differences were found in CDI language scores (Table 4). Girls scored higher on all CDI sub-scales at most time points, except for receptive vocabulary at 18 months [H(1, n = 147] = 2.939, p = 0.086), yet with a tendency in favor of girls.

Bivariate relationships between language development and temperament, demographic and perinatal variables were explored by correlation analyses (Table 5) to select variables for regression analyses predicting language outcomes.

Concurrent correlations were investigated at 18, 24 and 30 months. 18-month Surgency showed a consistent relationship with all measures of language development: higher Surgency was related to better language skills. Better Effortful Control was significantly related to more developed use of gestures and there was a tendency toward better receptive vocabulary. At 24 months, higher Surgency was related to better expressive vocabulary. At 30 months, the association between these two measures only showed a trend-level correlation, however, in the same direction as at earlier ages.

Longitudinal correlations were weak and sparse, however, Surgency at various ages tended to be related to measures of language and communicative development at later ages. Higher Surgency at 9 months was related to higher receptive vocabulary and gesture use at 18 months. Similarly, positive associations appeared between 18-month Surgency and expressive vocabulary at 24 months, and 24-month Surgency and expressive vocabulary at 30 months, with higher Surgency being related to a more extensive expressive vocabulary. In addition, 18-month receptive vocabulary and gesture use were associated with 9-month Effortful Control. Higher Negative Affectivity at 6 months was significantly correlated with more developed expressive vocabulary at 24 months. However, later Negative Affectivity (at 18 months) had the opposite relationship with expressive vocabulary at 30 months: more negative affect was associated with lower expressive vocabulary a year later.

Of the temperament factors in infancy and early childhood, Surgency seems to be indicated in language acquisition both concurrently and longitudinally, spanning from receptive to expressive language.

To better understand how temperament and perinatal factors affect each language and communication skill measured by the H-CDI, we conducted linear regression analyses with stepwise selection separately for each CDI variable at 18, 24 and 30 months. Temperament variables of preceding ages, perinatal variables, and infant sex showing significant correlations with the predicted variable were included in the regression.

First, we examined receptive vocabulary at 18 months (Table 6). Birthweight, Surgency and Effortful Control at 9 months were entered in the model. In the final model [R²⁼0.089, F(1,104) = 10.161, p = 0.002], the single significant predictor was Surgency measured at 9 months (β = 0.298, p = 0.002).¹

In the model predicting gestures at 18 months, sex, Surgency, and Effortful Control at 9 months were entered (see Table 7). In the final model [R²⁼0.111, F(2,103) = 7.559, p = 0.001], predictive variables were Surgency measured at 9 months (β = 0.272, p = 0.004) and sex (β = −0.275, p = 0.004).²

To predict expressive vocabulary at 18 months, only sex and gestational age were entered, as no temperament variable showed a significant correlation with this language outcome (see Table 8). Here, only one model was generated [R²⁼0.076, F(1,143) = 11.778, p < 0.001], with gestational age reaching significance (β = 0.276, p < 0.001).³

Table 9 presents the regression model predicting expressive vocabulary at 24 months, in which sex, gestational age, Surgency and Negative Affectivity at 18 months were entered. In the final model [R²⁼0.112, F(2,129) = 8.163, p < 0.001], gestational age (β = 0.248, p = 0.004) and Negative Affectivity measured at 18 months (β = −0.199, p = 0.019) were the significant predictors.⁴

Infant sex, birthweight, Negative Affectivity at 6 months, and Surgency at 24 months were entered into the regression to predict expressive vocabulary at 30 months (see Table 10). In the final model [R²⁼0.088, F(1,64) = 6.152, p = 0.016], the only significant contributor was Surgency at 24 months⁵ (β = 0.296, p = 0.016).

Several studies have linked positive affectivity with language development in infancy and early childhood (Laake and Bridgett, 2014; Pérez-Pereira et al., 2016; Davison et al., 2019). Positive affectivity contributes to the Surgency factor in Rothbart’s temperament model (Gartstein and Rothbart, 2003). Laake and Bridgett found that 10-month-old infants with higher Positive Affectivity/Surgency, as measured by IBQ-R, showed improved expressive language at 14 months. Davidson’s study also supported these findings, as infant Positive Affectivity/Surgency measured at 8 and 12 months predicted expressive language skills at 24 months.

Consistent with the literature, we also found Surgency to be related to early language skills. Surgency at 9 months predicted receptive vocabulary and gesture use at 18 months, while Surgency measured at 24 months was a significant contributor to expressive vocabulary at 30 months. Of concurrent associations between Surgency and language measures, only correlations with 18-month receptive vocabulary and gesture use remained significant after Bonferroni correction. However, at least a trend-level association with concurrent Surgency pointing in the same direction can be observed for expressive vocabulary at all ages. Thus, infants with higher Surgency scores demonstrated better language abilities, both in terms of receptive and expressive language. These results suggest that Surgency may be related to language development over an extended period. Since there is some stability in Surgency over time (correlations among Surgency values measured between 9–30 months ranged between 0.358–0.694), temperament can be expected to show a weak longitudinal correlation with expressive communication.

As children with high Positive Affectivity/Surgency are more likely to engage in and elicit social interactions, they have more opportunities to practice and improve their expressive language skills (Laake and Bridgett, 2014). This assumption could also apply to gesture use and receptive language, as both are related to expressive language use. Extensive social interactions provide more opportunities not only for the use of expressive vocabulary but also for gestural communication. More social interactions may result in varied, and increased amounts of language stimuli, fostering the development of language skills.

In our study, Effortful Control was not a significant predictor of language development in the regression models. Only weak correlations were observed between Effortful Control at 9 months and gesture use and receptive vocabulary at 18 months. Medium concurrent correlation with gesture use was also observed at 18 months.

The link between effortful control and language development remains unclear, despite some studies (Salley and Dixon, 2007; Keller et al., 2016) suggesting a positive relationship that could potentially be attributed to varying attentional capacities, which are thought to support language acquisition (Snijders et al., 2020). Effortful Control, as measured by Rothbart’s temperament questionnaires, is related to the functioning of the executive network (Posner et al., 2016). In turn, a link was demonstrated between the executive network and language development, production, and comprehension (Ye and Zhou, 2009; Shokrkon and Nicoladis, 2022). Furthermore, language development may also contribute to executive function development and self-regulation (Roben et al., 2013; Bruce et al., 2023).

However, Bruce et al. (2022) found that Effortful Control was only related to concurrent language, and 10-month Orienting/Effortful Control did not predict 24-month expressive language. Similarly, Ishikawa-Omori et al. (2022) did not find a predictive link between Effortful Control at 18 months and language development at 40 months. Keller et al. (2016) only demonstrated a significant relationship in the second language competence of dual language learners in childhood. The lack of predictive power of Effortful Control preceding the age of language assessment in the regression models and the separately observed concurrent correlation are in line with these results.

Negative Affectivity was entered in regressions at 24 and 30 months, however, only 18-month Negative Affectivity proved to be a significant predictor for lower expressive vocabulary at 24 months. This result supports earlier findings that Negative Affectivity may be associated with worse language skills (Dixon and Smith, 2000; Garello et al., 2012; Ishikawa-Omori et al., 2022). For instance, Garello et al. found that at the ages of 24–30 months, increased Negative Emotionality and Motor Activity correlated with poorer language production and comprehension. Similarly, Ishikawa-Omori et al. (2022) reported that Motor Activity, a scale of the Negative Affectivity factor, measured at 18 months, predicted lower expressive and receptive language skills at 40 months. They suggested that fidgeting behavior may reduce the availability of attentional resources, and as a result, it could hinder language learning. Excessive negative emotions could limit the resources children can allocate for information processing and language learning. They may also influence the way the children and their social partners interact. Children displaying more negative affect indeed performed worse on a joint attention task at 21 months (Salley and Dixon, 2007).

Gestational age proved to be a significant predictor for expressive vocabulary at 18 and 24 months. It’s been well-documented that both preterm birth and low birthweight can negatively impact language development into school age and beyond (Husby et al., 2023). Emerging findings, however, suggest variation in the development of term-born children, indicating differing developmental trajectories for early-term, full-term, late-term and post-term children (MacKay et al., 2010; Espel et al., 2014; Bentley et al., 2016; Snijders et al., 2020; Dhamrait et al., 2021). Our results suggest that longer in-utero development may support the development of expressive language. The final weeks of intrauterine development are characterized by rapid brain development. Children born early-term will not benefit from the effect of uterine neurosteroids (Hüppi et al., 1998; Limperopoulos et al., 2005; Shaw et al., 2019) as long as children born at later gestational ages. Increasing evidence shows long-lasting brain structure differences in preterm infants (Inder et al., 2005; Rogers et al., 2018). For instance, variations in functional connectivity were present even in adolescence after preterm birth, suggesting distinctive neurodevelopment potentially underlying behavioral differences (Lubsen et al., 2011). Term-born infants’ brain development also seems to benefit from longer gestation, within the time window of 37–41st weeks. Gestational age was related to differences in brain development in school-age children (Davis et al., 2011; Nivins et al., 2023). Such a variation may contribute to the observed differences in cognitive functioning and language skills (Ma et al., 2022).

Other than temperament and gestational age, infant sex also seems to contribute to variations in language skills. Previous studies have shown sex differences in language acquisition (Eriksson et al., 2012; Law et al., 2019), which aligns with our findings. Except for receptive vocabulary at 18 months, girls performed significantly better on all language measures and infant sex predicted the use of gestures at 18 months. Although we have only assessed gestures at 18 months, previous studies found girls using more gestures and starting earlier than boys (Özçalişkan and Goldin-Meadow, 2010; Germain et al., 2022).

Our aim was to investigate the role of temperament and some perinatal and maternal characteristics on early language development in a sample of low-social-risk, first-born term infants. Our sample was rather homogeneous as all participants were Caucasian of Hungarian origin, and the maternal education level was generally high across the sample. Results indicate the contribution of Surgency both concurrently and longitudinally on various measures of language development and the influence of gestational age on expressive vocabulary at 18 and 24 months. Negative Affectivity only predicted expressive vocabulary at 24 months. Despite Effortful Control being correlated with 18-month language, it was not a significant predictor in the regression models.

However, a major limitation of the study was the sample size and the missing data due to the pandemic. The COVID-19 pandemic hit during data collection and caused unexpected loss of data (data collection could not be conducted due to closures) and thus a higher dropout rate. Although the pandemic had no direct influence on the data presented (no differences were observed on any measures between data collected pre-pandemic and pandemic, post-pandemic periods), there might be hidden underlying effects of the quarantine period. Only 3 families reported contracting COVID-19 during the data collection. Thus, we may assume that results were not influenced by the neurological effects of the viral infection. Another limitation was the relatively low reliability of some temperament factors (Surgency at 9 months: 0.61, Negative Affectivity at 24 months: 0.63). Since correlations of the Surgency factor were consistent with those of other ages (albeit weak across the board), we have decided to include it in the regression analyses.

Our results extend previous findings as we have demonstrated associations with Surgency at the early stages of language acquisition for both receptive and expressive vocabulary, and showed the additional significant contribution of gestational age and Negative Affectivity. Gestational age was identified as a predictor for language in preterm infants previously. Our results extend this association to the narrower time window of gestational age of full-term infants. The latter finding may have relevance for medical practice and child educational support agencies. In line with other studies highlighting difficulties in later academic performance, this calls for increased attention to the early development of early-term and term infants.

Our results highlight the importance of longitudinal studies using tools to measure temperament based on the same theoretical concept over time. Also, investigating the small differences in gestational age in term infants in a larger sample may reveal important effects on language acquisition. With more evidence on how early-term status may influence later cognitive and language development, research on how certain environmental factors, such as socioeconomic status, maternal education, and quality of mother–child interaction might interact with gestational age can yield important results that can be translated into practices supporting early childhood development.

Extending the study beyond 30 months is crucial to identify early characteristics of developmental pathways leading to language impairment. The role of Surgency and the relative lack of power for Effortful Control in this sample calls for experimental investigation of the development of very early executive functions and attentional functioning.