Emotion-specific vocabulary and its relation to emotion understanding in children and adolescents

ABSTRACT

Among children and adolescents, emotion understanding relates to academic achievement and higher well-being. This study investigates the role of general and emotion-specific language skills in children’s and adolescents’ emotion understanding, building on previous research highlighting the significance of domain-specific language skills in conceptual development. We employ a novel inventory (CEVVT) to assess emotion-specific vocabulary. The study involved 10–11-year-old children (N = 29) and 16–17-year-old adolescents (N = 28), examining their emotion recognition and knowledge of emotion regulation strategies. Results highlight the ongoing development of emotion-specific vocabulary across these age groups. Emotion recognition correlated with general vocabulary in the younger group. In the older age group, emotion recognition was related to emotion-specific vocabulary size, but this effect only became apparent when controlling for the depth of emotion-specific vocabulary. Against expectation, there were no significant contributions of general or emotion-specific vocabulary to knowledge of emotion regulation strategies in either age group. These findings enhance our comprehension of the nuanced interplay between language and emotion across developmental stages.

KEYWORDS:

• Emotion understanding
• language development
• emotion-specific vocabulary
• children
• adolescence

Introduction

The transition from childhood to adolescence is characterised by profound biological and social changes, which present a considerable challenge to socio-emotional development. Longitudinal evidence suggests that negative emotions increase while positive emotions decrease from childhood to late adolescence (Bailen et al., 2019; Larson et al., 2002; Weinstein et al., 2007). To adapt to these changes, adolescents must re-evaluate and adjust their understanding of emotions and the strategies they use to regulate them (Young et al., 2019). Adolescents who have a better understanding of emotions tend to have higher levels of academic achievement and well-being (Rivers et al., 2012). However, those who struggle to identify their emotions often report experiencing more negative affect and show deficits in socio-emotional functioning (Ciarrochi et al., 2008). Language plays a crucial role in the development of emotion understanding and emotion regulation (Cole et al., 2010; Holodynski et al., 2013; Lindquist et al., 2015). Therefore, investigating the intersection of language skills and emotion understanding in middle childhood and adolescence is crucial to advancing the understanding of socio-emotional development in this transition period.

Emotion understanding encompasses recognition and conceptual knowledge of emotions and is linked to social and academic competencies (Denham et al., 2003; Trentacosta & Izard, 2007). Differential associations between general and emotion-specific language skills and emotion understanding are well documented in young children (Beck et al., 2012; Ornaghi & Grazzani, 2013; Streubel et al., 2020). Concurrently, previous research and research from other domains highlight the importance of domain-specific language skills for conceptual development (Bartsch & Wellmann, 1995; Dehaene et al., 1999; Grazzani & Ornaghi, 2012; Streubel et al., 2020). However, little is known about the development of emotion-specific language skills as well as the associations of these skills with domain-specific concept development in later childhood and adolescents.

Building on previous research highlighting the significance of domain-specific language skills for conceptual development, the present study investigates the role of general and emotion-specific language skills in children’s and adolescents’ emotion understanding.

Early on, children begin to recognise and name facial expressions of basic emotions (e.g. sadness, anger, happiness) and understand that emotions are linked to situational causes. As they acquire theory of mind, they begin to understand that mental states can trigger emotions (Denham et al., 2003). In middle childhood (from around age 6 up to grade school), more advanced cognitive skills and self-awareness enable children to gain a more reflective understanding of emotions including knowledge of emotion regulation strategies (Waters & Thompson, 2014) and understanding of mixed and more complex emotions, such as embarrassment, jealousy or pride (Lagattuta & Thompson, 2007; Larsen et al., 2007). Children become more fluent in reasoning about emotions and the connection between situations, thoughts and feelings (Davis et al., 2010; Lagattuta et al., 2016). By adolescence, emotion understanding continues to deepen due to advanced cognitive skills like metacognition, self-monitoring, cognitive flexibility, and more abstract reasoning about emotion-relevant circumstances, such as principles like fairness and justice (Steinberg, 2005).

Language as a (general) means of cognitive representation plays a key role in acquiring and developing emotion understanding as it helps children to mentally store and access knowledge about emotion concepts, including related experiences, causes, and consequences, leading to better understanding of emotional processes (Cole et al., 2010; Holodynski, 2013; Pons et al., 2003). In line with these theoretical assumptions, numerous studies document strong relations between general language skills and emotion understanding throughout early and middle childhood (Beck et al., 2012; Griffiths et al., 2020; Pons et al., 2003). More recent approaches stress the role of emotion-specific (rather than general) language skills for the development of emotion understanding: They state that particularly emotion terms might help children to connect diverse aspects of emotion concepts, facilitating their representation as entities (Lindquist, 2017). Concurrently, empirical studies show that a differentiated emotion-specific vocabulary positively relates to various aspects of emotion understanding (Bazhydai et al., 2018; Beck et al., 2012; Ornaghi & Grazzani, 2013; Streubel et al., 2020). Streubel and colleagues (2020), for example, examined the relative contributions of emotion-specific and general vocabulary to several components of emotion understanding in children aged four to nine years. In their study, they differentiated between emotion vocabulary size (i.e. number of words) and emotion vocabulary depth (i.e. adult-like usage of words). As vocabulary acquisition in a specific content domain entails both the number of words recognised or used as well as the accuracy of the knowledge about the meaning of each word, the term “vocabulary depth” has been introduced to refer to the convergence of children’s usage of words with an adult-like usage (see Saji et al., 2011). Streubel and colleagues found a significant positive contribution of emotion-specific vocabulary size to 4–5-year-old children’s knowledge of emotion regulation strategies beyond general vocabulary and in all age groups a positive contribution of emotion-vocabulary size on emotion recognition. However, in one measure of emotion understanding (i.e. knowledge of mental emotion triggers) this relation changed in 8-to-9-year-old children, and the size of their emotion-specific vocabulary had a negative impact. Post-hoc analysis revealed that a larger repertoire of emotion words (size) seemed to be detrimental only in those children who had acquired comparatively little depth. This highlighted the importance of considering both size and depth of emotion-specific vocabulary to understand the relation between emotion specific vocabulary and conceptual development.

Research on the development of emotion-specific language skills suggests that children’s initial concepts of emotion are broad and valence-based (e.g. good or bad; Widen & Russell, 2008). Between the ages of two and five, children commence naming basic emotions. In the subsequent years, their emotion-specific vocabulary seems to double every two years, until age 12, when it seems to plateau (Baron-Cohen et al., 2010; Grosse et al., 2021; Nook et al., 2020). Some emotion concepts are well-developed by the end of preschool years (e.g. happiness, sadness, anger), but others (e.g. fear, surprise, disgust, guilt, pride) are not fully acquired until middle childhood or adolescence (Widen & Russell, 2008). Studies including older participants (up to age 16–25) suggest that emotion-specific vocabulary development continues to progress (Nook et al., 2020; O’Kearney & Dadds, 2004). With increasing age, adolescents produce more and more differentiated and abstract emotion words (Bazhydai et al., 2018; Nook et al., 2020). But research on adolescents’ emotion-specific vocabulary has thus far been limited to a small number of emotion categories (e.g. O’Kearney & Dadds, 2004: “anger” and “fear”, Bazhydai et al., 2018: five emotion categories: “happy, relaxed, angry, sad, and nervous”). Except for the study by Nook and colleagues (2020) in which participants were asked to define 27 emotion terms, and their responses were scored for comprehension in the sense of conceptual knowledge and abstractness of this response. The study found that emotion comprehension scores plateaued at age 11 and abstractness scores plateaued only later at 18 years of age. No relationship between emotion language and emotion understanding was investigated in their study.

Taken together, the importance of emotion-specific language skills in promoting related conceptual knowledge is well established in theory and empirical evidence in younger children. However, research on these relations in older children and adolescents is scarce, while we know that both emotion understanding and emotion-specific vocabulary are subject to ongoing development.

The purpose of the present study was to investigate (1) age-related differences in emotion-specific vocabulary size and depth concerning a broad range of emotion concepts and (2) the relative contributions of emotion-specific and general vocabulary to individual differences in emotion understanding (i.e. emotion recognition and knowledge of emotion regulation strategies) in 10-11-year-old children and 16-17-year-old adolescents, while taking into account potential diverging effects of emotion-specific vocabulary size and depth.

Based on previous research, we hypothesised that: (1) emotion-specific vocabulary (size and depth) would increase with age. Furthermore, we hypothesised that in both age groups, (2) general vocabulary would be positively related to measures of emotion understanding, and (3) emotion-specific vocabulary would contribute to emotion understanding beyond general vocabulary.

Materials and methods

Participants

The study included 57 participants in two age groups: 10–11-years (n = 29; M = 11.0; SD = 0.4; 13 girls and 16 boys), and 16–17-years (n = 28; M = 17.0; SD = 0.6; 19 girls, 8 boys, and 1 diverse). The younger age group was recruited via a pre-existing database of families. The adolescent age group was recruited by distributing flyers, via social media, mailing lists and by direct contact outside schools. The population from which participants were drawn is 87.1% native Germans and predominantly middle class (Statistisches Landesamt des Freistaates Sachsen, 2023). The study was conducted in accordance with the ethical principles of the Declaration of Helsinki, approved by an ethics review board.

Sample size – An a-priori power analysis was conducted using G*power Version 3.1.9.7 (Faul et al., 2007) for sample size estimation based on prior studies that predicted children’s emotion understanding by (at least two) different language or emotional abilities (Bazhydai et al., 2018; Ornaghi & Grazzani, 2013; Streubel et al., 2020). Effect sizes f² derived from the reported R² ranged from .39 to .69, which is considered large according to Cohen’s (1988) criteria. With a significance criterion of α = .05 and a power of .95, a minimum sample size of N = 30 is required for linear multiple regression to achieve an effect size of f² = .39. The obtained sample size of N = 57 and n = 28/29 within subsamples is sufficient to test the study hypothesis.

Procedure

The 10–11-year-olds underwent one-on-one testing with female investigators in a quiet environment in the university’s laboratory. The 16−17-year-olds underwent testing with a female investigator via online video conference. Test sessions lasted up to 50 min and included assessments of general and emotion-specific vocabulary as well as emotion understanding tasks in a predetermined order.

Instruments

General expressive vocabulary – was assessed by a German adaptation of the Expressive One Word Picture Vocabulary Test (EOWPVT-4; Martin & Brownell, 2011). For 10–11-year-olds, we used a 42-items short version (Saalbach et al., 2016). For 16–17-year-olds, we used a translation of the 190-items original long version with item 95 being the age-based starting point. Translation was performed by means of a dictionary and verified by back-translation.

Each item presents a to-be-named illustration of objects, actions or concepts with increasing difficulty (e.g. dog, binoculars, directions). For each correctly named picture, participants receive one point – summing up to a maximum total score of 42 in the short version and 190 in the long version.

Emotion-specific vocabulary – was assessed using the Children’s Emotion-specific Vocabulary Vignettes Test (CEVVT; Streubel et al., 2020). The test comprises 20 illustrated emotion vignettes, each with a drawing and a recorded narrative of a child protagonist in a typical emotional situation with emotion-specific facial and bodily expressions, and thoughts (see Figure s1 in supplementary material for example). The test covers 6 basic emotions (joy, fear, sadness, anger, disgust, and surprise) and 14 semantically subordinated complex emotions (e.g. pride and contentment as subordinated to joy). After the presentation of each vignette, participants were asked how the vignette’s protagonist child feels. The transcribed responses were spell-checked and then coded by EmotionTool Version 1.0 (ET1.0; Bräuer & Streubel, 2023), a python-based computer program designed for automated CEVVT coding. ET indicated whether participants’ responses contained emotion words, using a pre-established word list. For more information on the vignettes, the procedure and the coding process please refer to the supplementary material.

Size was indicated by the number of different emotion words used across all 20 vignettes. Depth was indicated by an individual score representing the extent to which each participant’s use of emotion words matched the naming pattern produced by the adult sample in Streubel et al. (2020; for details see supplementary material).

Emotion understanding – was assessed by two distinct components – emotion recognition and knowledge of emotion regulation strategies – using different measures for each age group, as instruments covering both age groups were not available at time of data collection.

Emotion recognition. We used a subscale of the Intelligence and Development Scales (IDS; Grob et al., 2009) for 10–11-year-olds – comprising 10 photographs of prototypical facial expressions of the emotions joy, anger, fear, sadness, and surprise. Children were asked to name the expressed emotions, receiving one point for each correct answer and a maximum sum score of 10. For 16–17-year-olds, we used the short version of the Geneva Emotion Recognition Test (GERT-S; Schlegel & Scherer, 2016). This computer-based test comprises 42 video clips with sound (duration 1–3 s), in which actors express 14 different emotions (7 positive, 7 negative). The clips present mimic, gestural and prosodic cues simultaneously. After each clip, participants are asked to choose which of the 14 given emotion terms best describes the expressed emotion, receiving one point for each correct answer with a maximum sum score of 42.

Knowledge of emotion regulation strategies. We used the IDS subscale emotion regulation (Grob et al., 2009) for 10-11-year-olds which is embedded in the emotion recognition task. For three emotions (anger, fear, and sadness), children were asked (a) what the pictured child could do to not be angry (or afraid/sad) anymore, (b) what they would do if they were angry/afraid/sad and did not want to be so anymore, and (c) what else they could do. If children initially misidentified the pictured emotion, the investigator corrected them before asking about emotion regulation. Children scored two points for naming adaptive emotion regulation strategies (e.g. problem-focused coping, distraction, reappraisal), zero points for maladaptive strategies (e.g. aggressive behaviour, self-deprecation, perseveration), and one point for other strategies (e.g. seeking social support, expressing feelings; see Grob et al., 2009 for details) – yielding a maximum sum score of 18. Since the IDS subscale emotion regulation is not applicable for 16–17-year-olds, we used the emotion regulation subscale for 11-20-year-olds from the IDS-2 (Grob & Hagmann-von Arx, 2018). The participants were read three different scenarios about an adolescent experiencing anger, fear or sadness (e.g. “Imagine that you are angry because your teacher made you take a difficult test”), while simultaneously being shown a photo of an adolescent with a facial expression indicating the respective emotion. After each scenario, participants were asked what they could do to not feel angry/afraid/sad anymore. As with the IDS, adolescents scored two points for naming adaptive emotion regulation strategies, zero points for maladaptive strategies, and one point for other strategies – again, summing up to a maximum score of 18 (see Grob & Hagmann-von Arx, 2018 for more details).

Analytic strategy

Age-related differences in emotion-specific vocabulary (research question 1) were analyzed via multivariate analyses of variance. To test the relative contributions of emotion-specific and general vocabulary to individual differences in emotion understanding (research question 2), we conducted multiple regression analyses with participant’s scores of emotion recognition and knowledge of emotion regulation strategies as the dependent measures. There was some missing data. Specifically, in the adolescent age group, we had 12 incomplete datasets of the GERT-S. Little’s MCAR test did not reach significance, indicating that missing data was independent of other main variables or gender in the whole sample (χ²(4, N = 57) = 1.29, p = .864) as well as in the adolescent age group (χ²(4, n = 28) = 1.67, p = .797). Independent t-tests and Levene’s Tests showed no significant differences between the adolescents who had GERT-S data available and those who did not, in terms of the mean and variance of all other main variables. Incomplete data sets were listwise excluded in the central analyses. All analyses were performed using SPSS.

Results

Descriptive statistics for all main variables and the two age groups are shown in Table 1. Since Shapiro-Wilks tests showed that neither emotion recognition in the group of 10–11-year-olds nor emotion regulation knowledge in the group of 16–17-year-olds were normally distributed, we used a non-parametric test for bivariate correlations between all measures of language skills and emotion understanding as well as gender. We calculated Spearman correlations across age groups (using within-age-group z-standardised measures) and for each age group separately (see Table s1 in supplementary material). Size and depth of emotion-specific vocabulary were significantly and positively associated with each other both across (r = .565, p < .001) and within age groups (r_10–11y = .496, p = .006; r_16–17y = .650, p < .001). We found a significant positive association between general vocabulary and emotion recognition for 10–11-year-olds (r = .553, p = .002). For 16–17-year-olds, the size of emotion-specific vocabulary was positively related to emotion recognition (r = .463, p = .055). No other significant correlations were found.

Table 1. Descriptive statistics and Shapiro-Wilk statistic for all main variables.

	Maximum score	10–11-year-olds (n = 29)				16–17-year-olds (n = 28)
		M (SD)	Min	Max	W	M (SD)	Min	Max	W
General expressive vocabulary	42 / 190	36.1 (2.5)	31	40	.93	145.8 (19.4)	105	182	.97
Emotion-specific vocabulary
Size	20	10.4 (2.1)	7	16	.96	14.2 (2.4)	9	18	.96
Depth	5.0	2.9 (0.6)	1.6	4.3	.99	3.3 (0.5)	2.3	4.1	.97
Emotion understanding
Emotion recognition (IDS)	10	8.8 (1.1)	7	10	.85***	–	–	–	–
Emotion recognition (GERT-S)	42	–				29.4 (5.1)	21	37	.95
Emotion regulation knowledge (IDS)	18	12.5 (3.0)	6	17	.94	–
Emotion regulation knowledge (IDS-2)	18	–				16.2 (1.5)	13	18	.92*

Note: M = Mean; SD = Standard deviation. W = Shapiro-Wilk Statistics for Normal Distribution. IDS, Intelligence and Development Scales. GERT-S, Geneva Emotion Recognition Test (short version).

*p ≤ .05. ***p ≤ .001.

To test for age differences in emotion-specific vocabulary (research question 1), we conducted a one-way-MANOVA with age group as between-subject factor and size and depth of emotion-specific vocabulary as dependent measures. Age group resulted a significant factor for the variance in emotion-specific vocabulary (F(2, 54) = 20.1, p < .001, η² = .43). The 16-17-year-olds had a significantly larger emotion-specific vocabulary size (F(1) = 40.3, p < .001, η² = .42) and depth (F(1) = 7.9, p = .007, η² = .13) than 10-11-year-old children.

To investigate the contribution of general and emotion-specific language skills to emotion understanding (research question 2), we conducted two blockwise multiple linear regression analyses for emotion recognition and emotion regulation knowledge as respective dependent measures. To test for possible interactions between age group and the effects of general or emotion-specific vocabulary, all measures were z-standardised within age groups. Note that z-standardisation does not allow us to compare means of dependent variables across age groups, but does allow us to examine relations between inter-individual differences in variables and their interaction with age. In a first step, we examined contributions of general vocabulary to both components of emotion understanding (Model 1 in Table 2). We also entered an interaction term (Age group × General vocabulary) to investigate whether contributions of general vocabulary differed among age groups. To investigate the contribution of emotion-specific vocabulary beyond general vocabulary, we entered emotion-specific vocabulary size (Model 2a) or emotion-specific vocabulary depth (Model 2b), or both size and depth (Model 3) as additional predictors in our regression analyses. In all models, we also added interaction terms (Age group × Size/Depth of emotion-specific vocabulary) to investigate whether contributions of emotion-specific vocabulary differed between age groups.

Table 2. Blockwise multiple regression analyses predicting emotion recognition and knowledge of emotion regulation strategies from general and emotion-specific vocabulary (β-values).

	Emotion recognition	Emotion regulation knowledge
Model 1
Constant	.017	.000
General expressive vocabulary	1.33**	.444
Age Group × General expressive vocabulary	−1.10**	−.421
R^2	.22	.02
F	5.89**	.55
Model 2a
Constant	.001	.000
General vocabulary	1.20**	.396
Age Group × General expressive vocabulary	−.914*	−.381
Emotion-specific vocabulary size	−.420	−.382
Age Group × Emotion-specific vocabulary size	.609	.330
ΔR^2 / R^2	.06 /.28**	.02 /.04
ΔF / F	1.61 / 3.83**	.42 /.48
Model 2b
Constant	.023	.000
General vocabulary	1.40**	.487
Age Group × General expressive vocabulary	−1.20*	−.475
Emotion-specific vocabulary depth	.032	−.483
Age Group × Emotion-specific vocabulary depth	−.118	.642
ΔR^2 / R^2	.01 /.23*	.06 /.08
ΔF / F	.17 / 2.29*	1.61 / 1.09
Model 3
Constant	.017	.000
General vocabulary	1.40**	.451
Age Group × General expressive vocabulary	−1.17**	−.473
Emotion-specific vocabulary size	−.689	−.057
Age Group × Emotion-specific vocabulary size	1.04*	−.197
Emotion-specific vocabulary depth	.456	−.495
Age Group × Emotion-specific vocabulary depth	−.755	.808
ΔR^2 / R^2	13 /.35**	10 /.12
ΔF / F	1.86 / 3.36**	1.36 / 2.28

Note: Coding of age group: 1 = 10–11-year-olds, 2 = 16–17-year-olds. ΔR²: Additional explained variance compared to Model 1.

* p ≤ .05. ** p ≤ .01.

Findings revealed no significant contributions of general or emotion-specific vocabulary to knowledge of emotion regulation across age groups. As expected, we found a significant positive contribution of general vocabulary to emotion recognition. There was a significant interaction effect with age group, indicating differences among age groups in the influence of general vocabulary on emotion recognition (Model 1). The inclusion of emotion-specific vocabulary size (Model 2a) or depth (Model 2b) did not improve the initial model: there were no significant contributions of emotion-specific vocabulary size or depth to emotion recognition beyond general vocabulary, and there were no age differences in these influences. Notably, after adding emotion-specific size and depth to the model at the same time (Model 3), the interaction effect between age group and emotions-specific vocabulary size became significant. That is, after controlling for emotions-specific vocabulary depth, emotion-specific vocabulary size significantly contributed to emotion recognition beyond general vocabulary and depending on the age group.

To explore these age group effects, we computed separate regression analyses for each age group, regressing general vocabulary (Model 1 in Table s2 in supplementary material) and children’s emotion-specific vocabulary size (Model 2a) or depth (Model 2b), or both size and depth (Model 3) on emotion recognition. For 10–11-year-olds, we found a significant positive relation of general vocabulary to emotion recognition (Model 1: β = .556, p = .002; F(1,27) = 12.10, p = .002; R²= .31). The inclusion of emotion-specific vocabulary size (Model 2a) or/and depth (Model 2b/3) did not lead to an improvement of the initial model, indicating no significant contributions of children’s emotion-specific vocabulary size and depth to their emotion recognition beyond general vocabulary.

For 16–17-year-olds, general vocabulary did not relate to emotion recognition. Instead, findings revealed a significant contribution of emotion-specific vocabulary size to emotion recognition after adding both size and depth of emotion-specific vocabulary to the model (Model 3: β = .754, p = .023; ΔF(1,12) = 3.83, p = .074; ΔR²= .36) . As the improvement of the regression model after adding emotion-specific vocabulary size and depth does only reach marginal significance (p = .074), the coefficient should be interpreted cautiously. Figure 1 illustrates the relation between general vocabulary (A) and emotion-specific vocabulary (B) with emotion recognition as a function of age group.

Figure 1. A: Relation between general expressive vocabulary and emotion recognition for 10–11-year-old children and 16–17-year-old adolescents. B: Relation between emotion-specific vocabulary size and emotion recognition for 10–11-year-old children and 16–17-years-old adolescents.

Discussion

The current study investigated age differences in emotion-specific vocabulary as well as the contribution of emotion-specific and general vocabulary to emotion understanding in two age groups: 10–11-year-old children and 16–17-year-old adolescents.

As expected, and in line with previous literature (Bazhydai et al., 2018; Nook et al., 2020; O’Kearney & Dadds, 2004), we found an age-related increase in emotion specific vocabulary size and depth. The adolescents used more emotion words than the 10-11-year-olds, and they used them in a way that was more similar to that of adults, supporting the notion of a prolonged development of emotion-specific language skills.

Our findings indicate that there are significant age-related differences in the importance of general and emotion-specific vocabulary regarding the contribution of language skills to emotion understanding. For 10–11-year-olds, we found a significant positive contribution of general vocabulary to emotion recognition. On the one hand, this finding is consistent with previous research documenting strong relations between general language skills and emotion understanding in younger children (Beck et al., 2012; Pons et al., 2003), particularly emotion recognition (Griffiths et al., 2020). On the other hand, this finding contradicts prior studies, as it shows that emotion-specific vocabulary did not contribute to explaining either indicator of emotion understanding in this age group (Bazhydai et al., 2018; Beck et al., 2012; Ornaghi & Grazzani, 2013; Streubel et al., 2020).

In 16–17-year-olds, however, emotion-specific vocabulary size rather than general vocabulary was positively related to emotion recognition. This relation only became apparent when controlling for the depth of emotion-vocabulary. This pattern is consistent with the interpretation of emotion-specific vocabulary depth as suppressor variable. There is a high correlation between emotion-specific vocabulary size and depth, but their joint variance does not seem to be relevant for predicting emotion recognition. Instead, the depth of emotion-specific vocabulary suppresses some of the relevant variance of emotion-specific vocabulary size. Without considering depth, the predictive power of emotion-specific vocabulary size would be underestimated. Firstly, this finding is consistent with prior studies showing relations between emotion-specific vocabulary size and emotion recognition in younger age groups. Secondly, this finding supports the idea that it is beneficial to examine different aspects of emotion-specific language skills to gain a more complete understanding of the role of domain-specific language in conceptual development.

One possible explanation for the age-related differences in the contribution of general and emotion-specific vocabulary to emotion recognition is that younger children’s ability to describe emotional displays may be limited by their general language skills. In contrast, in adolescents, general language skills may be sufficiently developed, so their ability to use specific emotion words becomes the distinguishing factor. However, caution is warranted in interpreting our findings due to the number of missing values in the emotion recognition measure in our adolescent sample. Furthermore, our data indicate a potential ceiling effect for the emotion recognition task among 10–11-year-old children. The reduced variance within this age group may have hampered our ability to determine the influence of emotion-specific vocabulary beyond that of general vocabulary in our sample. A ceiling effect was not expected as the IDS is a validated and normed instrument for this age group. It is valuable that the GERT-S has been developed as an alternative. However, it has not yet been validated with younger age groups so far, only from 18 years upwards (Schlegel & Scherer, 2016). Moreover, we cannot rule out the possibility that the age-related differences in the contribution of general and emotion-specific vocabulary to emotion recognition are (at least partly) due to the use of two different instruments for measuring emotion recognition. The GERT-S task requires recognition of more specific emotions and emotion labels than the IDS task, making it more dependent on mastery of emotion-specific language. Further studies are necessary to investigate the robustness of our results.

Unexpectedly, in both age groups, we found no significant contribution of either general or emotion-specific vocabulary to knowledge of emotion regulation. In line with this, Streubel et al. (2020) reported that emotion-specific and general language skills had an impact on knowledge of emotion regulation only for the two youngest age groups they studied (i.e. 4–7 years). They argued that the IDS captures knowledge of emotion regulation strategies for only three basic emotion concepts: anger, sadness, and fear. This limited number may explain the observed result considering that even children (and adolescents) with a relatively small (emotion-specific) vocabulary might know these emotion words and related strategies. Our data suggest a potential ceiling effect in the task of emotion regulation knowledge among 16–17-year-old adolescents. However, it is the only validated and normed instrument available to measure this construct, that we are aware of. Future research may benefit from adapting other instruments developed for adults, such as the Situational Test of Emotion Management (STEM; MacCann & Roberts, 2008), for use in the study of adolescents.

Some additional limitations of the present study should be noted. Importantly, due to the language-based nature of the instruments used for both emotion-specific vocabulary and the different aspects of emotion understanding, some basic level of correlation was expected. However, our results did not support this claim as we found no correlation between emotion vocabulary and either emotion understanding measure in the younger age group. This suggests that the different instruments are not measuring the same underlying construct. Furthermore, we assessed general language competence to control for its impact on all related measures. Another important limitation is that due to the onset of COVID-19 and corresponding restrictions during data acquisition, we used different test settings (i.e. face-to-face and online testing) for the two age groups. This may have affected some of the results. Finally, this study was conducted in a typical Western culture with a mid-to-high educational background. Verification of the findings requires further research with more diverse samples and in different cultural settings.

In conclusion, this study is, to the best of our knowledge, the first to investigate the contribution of different (emotion-specific) language skills to various aspects of emotion understanding in children and adolescents. Our findings suggest that the relations between general language, emotion-specific vocabulary size and depth, and different components of emotion understanding throughout development are more complex than has been captured in this and previous research. Future research should aim to continuously track the development of emotion understanding and language skills across a wider age range, potentially in longitudinal studies. It is important to use emotion understanding measures that are applicable to a wider age range and cover different aspects of emotion understanding.

Acknowledgments

We want to thank the participating children, adolescents and their families! We also want to thank Lisa Grosse for help with data collection.

Disclosure statement

No potential conflict of interest was reported by the author(s).