Early linguistic reference to first and second person depends on social understanding as well as language skills: Evidence from Czech 30-month-olds

Abstract

The early use of first and second person pronouns has been viewed as a sign of emerging social understanding. However, it may also depend on general language development: pronouns do not appear among the first words children acquire. In addition, some languages conjugate verbs for person, and the inflections may thus show similar relations to social understanding and language as pronouns. Very few studies have examined the relations between early person reference on the one hand and both social understanding and language development on the other. The present study used data on spontaneous language transcripts and language tasks from 58 Czech-speaking children aged 30 months to examine the relations between person reference, social understanding, and general language development. Social understanding was estimated from children’s use of the mental state language (MSL). The results confirm that both MSL and general language development show unique relations to person reference with pronouns as well as verbs.

Keywords

Mental state language person personal pronouns social understanding verb conjugation

Children’s acquisition of personal pronouns depends on two distinct domains of their knowledge. On the one hand, it is related to the development of language, both grammar and semantics. Pronouns are closed-class elements that appear in children’s language only after many nouns, verbs, and social words (Brown & Fraser, 1964; Chiat, 1986; Morehead & Ingram, 1973; Tanz, 1980). They often have mostly grammatical functions, and they do not refer to a stable class of persons, but rather to the speaker’s or listener’s role in the discourse. On the other hand, knowledge of personal pronouns, especially the first and second person pronouns, depends on the development of social understanding (Levine & Munsch, 2013; Piaget, 1929). To distinguish and correctly use words such as I or you, children must have developed some concept of their own individuality as opposed to others. The early use of person-referring pronouns may thus be related both to language development and the development of social cognition. Similar relation to social understanding should be observed in other linguistic devices used for person reference, such as verb inflections for person. However, only a few studies of early pronouns have examined the effects of linguistic development and social understanding at the same time, and even fewer studies have examined how verb inflections used for person reference relate to social understanding and linguistic development. The present study addressed the emerging use of person reference in both pronouns and verbs, and its relation to language development and social understanding, as indicated by the use of mental state language in children. Additionally, the study uses a language – Czech – that has not been previously studied in this respect, adding to the cross-linguistic generalizability of this line of research.

Pronouns, verb inflections, and social understanding

One perspective on the emergence of pronouns in young children is that pronouns mark the emerging sense of individuality, or differentiation between self and others. Piaget (1929) pointed out that the emergence of the first person pronoun coincides with the increasing ability to understand spatial perspective, which indicates that there is a relation between children’s understanding of the differences between self and others and their use of the first person pronoun. Goodenough (1938) formulated a similar proposal. Some developmental psychology textbooks present the emergence of personal pronouns as one of the shifts in behavior that are related to the development of self-awareness and individuality, along with mirror self-recognition, visual perspective-taking, and possessiveness (Levine & Munsch, 2013). There is now solid evidence that pronouns are related to various tasks that require social understanding, such as self-recognition in a mirror (Courage et al., 2004; Lewis & Ramsay 2004), cooperation (Brownell et al., 2006), pretend play (Lewis & Ramsay, 2004), and visual perspective-taking (Loveland, 1984; Ricard et al., 1999). Another argument linking correct pronoun use and social understanding is based on findings from autistic children. Kanner (1943) was among the first to observe that verbal autistic children sometimes use I for the addressee and you for themselves. A number of subsequent studies suggested that autistic children have problems using person-referring pronouns, either avoiding them or using them incorrectly, and that also their mothers tend to use fewer second person pronouns (Bartak & Rutter, 1974; Charney, 1980; He et al., 2018; Jordan, 1989; Lee et al., 1994). This is the case even for autistic children who are native users of American Sign Language, in which pronouns are expressed as points to self or others and are thus transparent gestures (Shield et al., 2015). Based on these findings, pronoun problems have been viewed as one of the typical features of autism-spectrum disorders. It should be noted that the inverted use of personal pronouns – I for you and vice versa – is more frequent in autistic children compared to neurotypical children, but still rather rare (Naigles et al., 2016). At the same time, children with other conditions, as well as some typically developing, even precocious, children show frequent pronoun reversals, suggesting that this phenomenon may have variable causal underpinnings (Evans & Demuth, 2012; Oshima-Takane et al., 1993). Perhaps use of person-referring pronouns in general might be a better indicator of development in this domain for the majority of children.

Pronouns are not the only language structure used for person reference. In many languages, verb inflections for person play a similar role. The early use of person inflections should thus be related to emerging social understanding in a way similar to pronouns. Verb inflections for person may provide important validation for the relation between person reference and socio-cognitive development. If such a relation is present for both pronouns and verb inflections, it is clear that more abstract properties of person reference are involved and not only specific characteristics of pronominal reference. To the authors’ knowledge, there are only two studies that have addressed this possibility (Longobardi et al., 2019; Markova & Smolík, 2014), and both found similarities in the relations between pronouns, verb person inflections, and social cognition. However, they used parent report questionnaires as the source of their data, and further research is needed to establish the relations using measures obtained directly from children.

Person reference is just one of the components of language that are related to social understanding. Another such domain is the language describing mental and internal states and processes, such as perceptions, emotions, thoughts or beliefs; this is known as mental state language (MSL). Bretherton and Beeghly (1982) first pointed out that children’s references to mental states indicate their increasing understanding of other people’s minds. MSL first appears at about the same time as the expanding system of pronouns, i.e. around 30 months (Bretherton & Beeghly, 1982; Shatz et al., 1983), and there is a well-established relation between MSL and the emerging theory of mind and social understanding (Bartsch & Wellman, 1995; Lewis & Ramsay, 2004).

Since both MSL and personal pronouns are related to social understanding, the emergence of pronouns and MSL should also occur in parallel or in close sequence. This relation should be rather salient because both MSL and pronouns are a part of the same general cognitive domain, language. The two studies that examined verb inflections as a personal reference device (Longobardi et al., 2019; Markova & Smolík, 2014) also used MSL as an index of social understanding but the reporting parents’ views of children could overestimate the relations between MSL and other social vocabulary such as pronouns.

Person reference and grammatical development

The relation between social understanding and pronominal person reference is well-supported, but it is not the only challenge in acquiring person-referring pronouns. Pronouns are linguistically complex elements that fulfill important grammatical functions, and their emergence may depend on linguistic factors rather than social understanding. Most available studies of personal pronouns and social understanding did not control for the overall linguistic development, and thus could not exclude this possibility.

Personal pronouns are not among the first words children produce, despite their high frequency and ubiquity in adult conversational discourse. Data on the mastery of pronouns in children are relatively scarce, although a number of studies have addressed various questions relevant to pronouns, such as pronoun reversal errors (Dale & Crain-Thoreson, 1993; Loveland, 1984). Like other closed-class elements, pronouns are generally acquired later than content words. Brown and Fraser (1964) examined the ability to imitate different classes of words, and found that young children (aged 25–35 months) imitated pronouns less accurately than nouns and verbs but more accurately than purely grammatical words, such as articles or auxiliaries. Morehead and Ingram (1973) showed that the pronoun I emerges early but pronouns you, he, or she do not show up before mean length of utterance (MLU) values of 3 and the age of 2;9. Chiat’s (1986) review also suggested that children first acquire I and then you. In addition, various researchers have pointed out that the earliest references of children towards self are with their own names, and references to their conversational partners are with nouns or names (e.g. Bates, 1990; Brown, 1973; Budwig, 1985; Chiat, 1982, 1986; Tanz, 1980). Although this might be related to limited social understanding, it can also be a way for children to use familiar linguistic devices and avoid structures that they have not yet mastered, such as pronouns or verbs inflected for first or second person. A certain amount of social understanding is needed for correct person reference with nouns as well, so the reason to use nouns may be that children know the grammatical regularities governing nouns, but not pronouns or verb inflections.

Person reference via verb inflections is perhaps even more obviously related to general language development than person-referring pronouns. In languages that have person inflections, these are an inseparable part of the verb conjugation system. Children usually master the third person singular form first, even though other forms may be used from very early on (Smolík, 2002). The increased use of first and second person may relate to the expanding knowledge of verb inflection paradigms rather than social cognition.

Because the acquisition of personal pronouns and verb inflections is linguistically complex, it is possible that their relation with social understanding is spurious, and these forms actually emerge as a result of changes in the linguistic system. Such a spurious correlation would not be surprising, as many quantitative measures of children’s skills have a tendency to increase with age. On the other hand, there are theoretical reasons to expect a relation between person reference and social understanding, and empirical research supports this view. To have an explicit test, research on early pronoun use and person reference in general must take into account the level of language development and measures of social cognition at the same time. Such concurrent assessment of these two domains can confirm whether they both have specific relations to the early use of pronouns.

Overview of the Czech pronominal and verb system

Czech is a West Slavic language, along with Slovak and Polish, with complex morphology and morphosyntax and flexible word order. Nouns and nominal phrases are inflected for seven cases and two numbers. Verbs are conjugated for person, number, tense, and mood. Verbs also carry aspect, which is sometimes marked morphologically, but is sometimes just a component of the verb’s semantics.

Czech has a rich system of pronouns: the traditional Czech philology distinguishes personal, possessive, demonstrative, interrogative, indefinite, and reflexive pronouns. For the present work, pronouns expressing person are especially relevant, and these include personal and possessive pronouns. Personal pronouns are inflected according to their own inflectional patterns. Just like in nouns, inflected forms may be ambiguous with respect to case, with the same surface form expressing different cases. Personal pronouns in the noun phrase act like nouns, possessive pronouns like adjectives. Table 1 summarizes the forms of first and second person singular personal and possessive pronouns in standard Czech, which were the target forms in the present study. It should be noted that there are large differences between the forms and inflections in standard and colloquial Czech, and there is considerable dialectal variation in the colloquial forms.

Table 1.

All forms of singular first and second person possessive and personal pronouns.

	Personal		Possessive (masculine)
	First person	Second person	First person	Second person
Nominative	já	ty	můj	tvůj
Genitive	mě	tě, tebe	mého	tvého
Dative	mně, mi	ti, tobě	mému	tvému
Acccusative	mě	tě, tebe	můj (inanimate), mého (animate)	tvůj (inanimate) tvého (animate)
Vocative	—	ty	můj	tvůj
Local	mně	tobě	mém	tvém
Instrumental	mnou	tebou	mým	tvým

Czech is a pro-drop language, which means that nominative case pronouns in the subject position are not obligatory. Overt use of subject pronouns often means some level of emphasis or contrastive use. If used, subject pronouns are stressed elements, not clitics. Clitical forms of pronouns, however, exist in other cases, genitive, dative, and accusative. Most pronouns have a short clitical form (tě you, ho him, it), as well as a full stressed form (tebe, jeho). However, the clitical and nonclitical form of the first person pronoun are the same, differing only in stress. Clitics are placed in the second clause position, after the first constituent, in a fixed sequence.

The other aspect of Czech grammar relevant for the present study is the person inflection in verbs. Verbs are conjugated for person in all tenses. Generally, the first person is marked by the endings -m or -u, the second person by -š, both preceded by a theme vowel that is specific for the conjugational class (see Table 2). In the past tense, the verb forms consist of the auxiliary verb ‘to be’ in the present form that is inflected for person and number, and a past participle inflected for gender and number. For third person singular and plural, the past tense auxiliary is omitted. In the future tense, the auxiliary is the future form of the verb ‘to be’, and the main verb is in the infinitive.

Table 2.

Inflected verb forms in singular for three classes of verbs.

Infinitive	nést	trpět	dělat
First pers.	nesu	trpím	dělám
Second pers.	neseš	trpíš	děláš
Third pers.	nese	trpí	dělá

The relation between person reference using pronouns and verb inflections is shaped by the grammatical properties of Czech. Verbs agree with the subject, so the person reference in verbs is the reference to the subject role. Because subject–verb agreement is obligatory, the person inflection for subject is present for every finite verb. On the other hand, pronouns are not obligatory in the subject position because the information is carried by the verb inflection.

Present study: questions and hypotheses

The goal of the study was to test whether there is a relation between linguistic person reference on the one hand and the emerging social understanding and general language development on the other. To address this question, measures of person reference, grammatical development, and social understanding were obtained using spontaneous language transcripts and various language tasks. The age of children (30 months) was chosen because existing research indicated that children by this age already use some person-referring pronouns but the system is not highly developed (Brown, 1973; Chiat, 1986; Morehead & Ingram, 1973). The basic hypothesis was that both social understanding and grammatical development would show independent relations to person reference. Addressing this research question in our study extends the existing research in three areas. First, is the inclusion of verb person inflections as person reference devices, and the comparison of their use to person reference with pronouns. The second novel aspect was examining the social understanding, language development, and person reference concurrently, using direct measures of children’s language and pronoun use. While some of the afore-mentioned studies included both language and social cognition as predictors of person reference, they relied on parent reports or focused on children with developmental conditions such as autism. Finally, the use of Czech is a novel contribution by itself; most previous work on person reference has been done on English, and there is very limited amount of work on languages with rich morphology and case marking. The present work may thus add to the cross-linguistic generalizability of findings on person reference.

Method

Participants and procedure

The study used data from a larger two-wave longitudinal study that included language sample measures as well as experimental tasks examining language comprehension and children’s social understanding. The data examined here are limited to the measures from language transcripts and language tasks in the first wave of the study. A total of 63 children (29 girls) aged 30.4 to 31.7 months were recruited via online ads. The final sample consisted of 58 children for most analyses used here; in some analyses, some children were omitted because they were not able to complete the testing sessions or because of administration error. All children were native speakers of Czech, and had no major developmental or neurological disorders.

The standard procedure included two lab sessions. During the first visit, parents were asked to participate in free play with their children for approximately 20 minutes, using a selection of toys available in the lab. The play session was used to record the language sample that was later transcribed and analyzed. Additionally, a set of language measures as well as experimental tasks that examined the children’s linguistic skills and social understanding was administered. Some tasks used pictorial stimuli on paper or a computer screen, some used toy props. The whole session was audio- and video-recorded; the recordings were used for transcribing the language sample and coding of the behavioral tasks. Parents signed informed consent forms that confirmed explicit consent with recording and explained its purpose and confidential nature. Children provided age-appropriate informal assent with the procedure and recording.

Language sample measures

Measures of pronominal reference, verbal inflection, mental state language, as well as the mean length of utterance were obtained by analyzing transcripts of spontaneous conversations with mothers. The transcripts were transcribed using an adapted version of the CHAT conventions (MacWhinney, 2000). The transcribers used the audio recordings, as well as written notes made by the examiner who was present during the play sessions. The transcripts were analyzed using custom Perl routines. For pronouns and MSL, the analyses searched the transcripts for all words that appeared in a search list, and listed the utterances containing these words. The count was then performed manually from these utterance lists, checking if the form identified in the search was among the intended target forms and disambiguating homographic forms (e.g. ty may be the nominative singular form of you or a plural form of the feminine demonstrative pronoun, i.e. these).

Pronouns

The primary dependent variable was the number of pronoun tokens produced during the sample. All case forms of the first person and second person personal pronouns (first person: já, m(n)ě, mi, mnou; second person: ty, tebe, tě, tobě, ti, tebou), as well as the forms of the first and second person possessive pronoun (můj, moje, moji in the first person, and tvůj, tvoje, tvoji for the second person) were counted. We only counted uses of pronouns that were grammatical and appeared in complete and intelligible utterances. Two numbers were counted for each category: the number of usages of words belonging to that category, i.e. number of tokens, and the number of distinct words from that category, i.e. number of types.

Verb inflections

Verb inflections were counted manually. For each transcript, a concordance list was created that listed all distinct word forms used, and the number of their occurrences. Verbs in the first and second person singular forms were manually searched in this list, and the number of occurrences (tokens) and of different verbs in each form (types) tallied. Only grammatical occurrences in complete and intelligible utterances were counted.

Mental state language (MSL)

The use of MSL was taken as a measure of advanced social understanding, and a precursor of the theory of mind. A list of 27 words was compiled that included words referring to perception, emotions, cognitive processes, and desires (see Appendix). The list was adapted from Markova and Smolík (2014), who based it loosely on Bretherton and Beeghly (1982) and Wellman et al. (2000). The transcripts were then searched for all occurrences of words from this list. The listing of utterances containing these word forms was checked manually to exclude possible homographs or typing errors.

Mean length of utterance (MLU)

The main measure of language development level was MLU. This was calculated by dividing the total number of word tokens in each transcript by the number of utterances. Most words in Czech are morphologically complex but it is not clear if children at this early age fully represent the complex morphological structure of many Czech words. In order to be conservative, and given very high correlations between MLU in morphemes and words in a similar language, Russian (Tomas & Dorofeeva, 2019), MLU in words was used.

Standardized and experimental tasks

Lexical task

In order to examine the level of receptive vocabulary and verbal ability, a short vocabulary comprehension task was presented. The task used a format similar to the Peabody Picture Vocabulary Test (PPVT; Dunn & Dunn, 2007): children were shown a panel with four pictures and asked to point to the referent picture of a given word. Stimuli were taken from the lexical subtest of a bigger battery of language tasks (Seidlová Málková & Smolík, 2014). Raw scores were used in the analyses as no norms are available for the age range examined here.

Grammar comprehension

The grammar comprehension task was based on pictorial material from the Test for Reception of Grammar (Bishop, 2003), using the working Czech stimuli developed for the research project ELDEL (Caravolas et al., 2013; Kucharská, 2014). The whole stimulus set comprised five blocks of different sentence types, each with four items. Similar to the lexical test, participants were asked to point to the picture that corresponded to the sentence presented by the examiner. The task has no norms for Czech available, so raw scores were used in all analyses.

Sentence repetition

Sentence repetition task is a powerful tool to identify differences in language processing. Children were presented with 15 sentences of increasing length and complexity, and asked to repeat each item. Accurate responses were scored with 2 points, repetitions with minor modifications with 1 point. The 1-point response could differ from the target by one word or by one transposition only, and had to be grammatically correct. Total raw scores in points were used in the subsequent analyses.

Analysis

The basic approach required the use of regression analysis. However, ordinary least squares regression is highly susceptible to outliers, and some of the children in the present study achieved much higher scores than other children on most measures. While these numbers reflect the children’s advanced level of language and cognitive development, using them in ordinary regression would inflate the estimated relations between all measures. For this reason, the results were analyzed using robust regression as implemented in the R package Robust (Wang et al., 2014). For regression models with continuous predictors, this package uses the method of S-estimation, which is resistant to the effects of outliers in both the independent variable and the predictors.

The analyses used person reference as the dependent variable. Separate sets of analyses were performed with the number of pronoun tokens, number of inflected verb forms of interest (verb tokens), the numbers of different pronouns (pronoun types), and the number of different verbs used with an inflection (verb types). These numbers depend on the size of the language sample. In order to account for this effect, the counts were divided by the total number of word tokens. The dependent variables in the analyses were thus the proportions of pronouns or verb inflections in the token totals for each child. The independent variables were the number of mental state language tokens or types (also divided by the total number of tokens), and one of the general measures of language development. Besides MLU, we used composite measures based on the lexical and grammatical comprehension tasks and the sentence repetition task, i.e. measures of language that were not dependent on the language sample. The composites were calculated by summing the z-scores on each task. One composite included only lexical and grammatical comprehension. The other composite also included sentence repetition, which resulted in a more comprehensive measure, but at the expense of a lower number of participants as many children were not able to finish the task.

Results

The descriptive statistics in Table 3 show that children were just at the beginning of their mastery of personal pronouns and mental state language terms, with most children showing just a few instances of each category. The initial analyses examined the relation between the use of first and second person pronoun or verb inflections as the dependent variables, and two predictors, MLU and the use of MSL. For the use of pronominal reference, this analysis found a significant effect of both predictors; for verb inflections, only the effect of MSL was significant (see Table 4). Additional analyses used the different aggregate measures of language development. In the first analysis, the composite language measure included all tasks administered to children, i.e. lexical comprehension, grammatical comprehension, and sentence repetition. This analysis showed significant independent contributions of MSL and language measures to person reference, both by pronouns and by verb inflections, but it only included 43 children, since many children had problems completing the sentence repetition task. It is possible that these children would in fact score zero if they showed some compliance and would start the task. We thus repeated the analysis, assigning all the children with missing sentence repetition data the score of 0. For pronouns, the effects of MSL and the language composite were statistically significant. For verb conjugation, it was only the effect of MSL that was significant, but not the effect of language composite. Finally, we performed an analysis without the sentence repetition data, using a language composite that included only lexical and grammatical comprehension. There were significant independent effects of MSL and language on pronouns, even though the effect of language was near the threshold with p = 0.048. For verb inflections, the effect of MSL was significant, while the effect of language was only marginal (p = 0.058).

Table 3.

Descriptive statistics for all measures.

Measure	Mean	SD	Median	Min	Max
Total utterances	128.88	27.80	129	67	196
Total word tokens	255.71	92.41	254.5	85	602
Total word types	97.41	35.43	97	29	198
Total pronouns	2.71	3.64	1.5	0	19
1st pers. tokens	2.29	3.14	1	0	16
2nd pers. tokens	0.41	0.88	0	0	4
Person inflection tokens	11.43	9.30	9.5	0	46
1st person infl. tokens	7.93	6.67	6	0	31
2nd person infl. tokens	3.50	4.06	2	0	19
Total pronoun types	1.48	1.52	2	0	6
1st pers. types	0.74	0.76	1	0	3
2nd pers. types	0.29	0.56	0	0	2
Person inflections types	7.07	5.07	6	0	24
1st person infl. types	4.6	3.47	4	0	14
2nd person infl. types	2.47	2.45	2	0	12
MLU	1.97	0.53	1.96	1	3.07
Sentence repetition	16.50	10.00	20.5	0	28
Grammar comprehension	9.89	3.82	10	3	19

Table 4.

Regression of pronoun or verb conjugation use on mental state language and various language measures.

	Pronouns/tokens			Verb conjugation/tokens
	t	p	∆R²	t	p	∆R²
Models with MLU	N = 58, R² = 0.23			N = 58, R² = 0.38
MSL/tokens	2.45	0.017	0.04	4.35	< 0.001	0.18
MLU	3.98	< 0.001	0.14	1.00	0.321	—
Models with lexicon, grammar and SR	N = 43, R² = 0.23			N = 43, R² = 0.28
MSL/tokens	3.06	0.004	0.08	4.77	< 0.001	0.22
Lex + grammar + SR	3.38	0.002	0.11	2.71	0.010	0.09
Models with lexicon and grammar	N = 55, R² = 0.13			N = 55, R² = 0.27
MSL/tokens	3.49	0.001	0.09	5.10	< 0.001	0.24
Lex + grammar	2.03	0.048	0.04	1.94	0.058	0.03
Models with SR, missing children as 0	N = 55, R² = 0.17			N = 55, R² = 0.26
MSL/tokens	3.05	0.004	0.08	6.41	< 0.001	0.22
Lex + grammar + SR	2.98	0.004	0.08	1.67	0.100	—

Note: MSL mental state language; MLU mean length of utterance; SR sentence repetition.

The second set of analyses examined the relations between the numbers of pronoun, verb, and MSL word types, rather than tokens, paralleling the structure of models used above. It should be noted that the numbers of types were quite low and the variability was thus limited, as shown in Table 3, so these analyses were not expected to yield strong effects. The results are summarized in Table 5. For pronouns, the unique effect of language was significant in the model using MLU as the language measure, but not in any other model. In contrast, MSL had a significant effect in the model with grammatical and lexical comprehension aggregate, and a marginally significant effect in models with grammatical comprehension, lexical comprehension, and sentence repetition as language measures. No significant effect was found when MLU was used as the language variable. For verb conjugation models, there was no significant unique effect of MSL when MLU was the language measure, but the effect of MSL was significant in the remaining models. Language variables had no significant effect in any of the models.

Table 5.

Regression of pronoun or verb conjugation use on mental state language and various language measures.

	Pronoun types/tokens			Conjugated verbs/tokens
	t	p	∆R²	t	p	∆R²
Models with MLU	N = 58, R² = 0.19			N = 58, R² = 0.15
MSL/tokens	1.49	0.142	—	1.31	0.194	—
MLU	2.10	0.040	0.084	1.01	0.316	—
Models with lexicon, grammar and SR	N = 43, R² = 0.18			N = 43, R² = 0.15
MSL/tokens	1.78	0.083	—	2.38	0.022	0.102
Lex + grammar + SR	1.25	0.220	—	1.69	0.098	—
Models with lexicon and grammar	N = 55, R² = 0.13			N = 55, R² = 0.14
MSL/tokens	2.27	0.028	0.091	2.35	0.023	0.010
Lex + grammar	1.22	0.228	—	1.03	0.310	—
Models with SR, missing children as 0	N = 55, R² = 0.15			N = 55, R² = 0.15
MSL/ tokens	1.79	0.080	—	3.00	0.004	0.092
Lex + grammar + SR	1.38	0.175	—	1.56	0.125	—

Note: MSL mental state language; MLU mean length of utterance; SR sentence repetition.

Discussion

Our study tested whether social understanding and general language development have unique relations with person reference, each above and beyond the other. This was largely confirmed in analyses using word tokens, i.e. the frequencies of usages for pronouns, verb conjugations, and MSL items. In analyses using word types, the results were less clear. No analyses showed unique effects of both MSL and language, but unique effects of MSL were present in multiple models, and a unique effect of language was present in the model with MLU as the language measure for personal pronouns. Given the limited number and variability of word types, it is not surprising that the effects are weak, but the analysis of types still shows effects of MSL and some effects of language. The results thus broadly support the expected outcomes for the research questions.

The basic research question in the study was about the relation between person reference, social understanding, and language development, addressing three separate issues. First, we tested whether person reference via verb inflections is related to social cognition, and whether it is similar to person-referring pronouns in this respect. The second issue was whether both social understanding and language development have independent relations with person reference. The third issue was the use of Czech, and how it could complement the findings from other languages.

With respect to the first issue, the study confirmed that verb inflections for first and second person show clear relations to mental state language; in fact, verb inflections were related to mental state language more strongly than to language development using both token and type counts. This confirms previous findings by Markova and Smolík (2014) and Longobardi et al. (2019), but unlike these studies, our analyses are based on data collected directly from children, which provides important converging evidence to the previous results.

The second issue involved in addressing our research question is more generally whether language development and social understanding have independent unique relations to person reference, whether by pronouns or verbs. This was clearly confirmed by our analyses of tokens, and the analyses of types also suggested that social understanding and language have their effects, depending on the exact analysis. This has some precedents in previous studies but the evidence has been limited. Naigles et al. (2016) showed that the frequency of pronoun reversals, which may mark the lack of social understanding, depends both on language development level and social understanding. Studies by Markova and Smolík (2014) and Longobardi et al. (2019) suggested similar conclusions but were limited by the parent-report nature of the data. Our current findings support these previous results, but one should also note the differences. Studies using parent reports could not get reliable information on the frequency of use in pronouns, verb inflections, and mental state language terms, i.e. could not use the token counts. The present study could assess the number of occurrences (tokens) of pronouns, inflected verbs, or mental state terms, as well as the number of different words, i.e. types. However, the results for tokens are more reliable because the single-session transcripts were limited in the number of opportunities for using different words. The main conclusions from the present study are thus based on tokens, not on types. The token numbers reflect how often the child uses person reference and mental state language, emphasizing how familiar children are with these areas of language, and whether they seek opportunities to use them. We assume that this is a better measure of the level of development in these domains than the use of the number of different pronouns or mental state words. For one thing, the number of types may be limited by linguistic factors: some words or word forms are not very frequent in general, and their absence may thus be natural in children. In addition, some of the forms of interest, especially pronouns, have only a few distinct forms, limiting the possible number of types. Due to these methodological differences, the present study does not just replicate the findings by Markova and Smolík (2014) and Longobardi et al. (2019), but extends them using different types of measures.

A somewhat surprising aspect of the present study is the difference between pronouns and verb inflections in their relation to language development and mental state language. The acquisition of verb inflectional morphology is an important part of language development. One would thus expect that the relation between verb inflections and general language would be particularly strong. The present findings show much closer relation to social understanding, or more specifically, mental state language. Perhaps researchers should pay specific attention to verb inflections, rather than pronouns, when searching for linguistic markers of social understanding. It is also possible that the mental state verbs are among the first verbs to be used in the first and second person. This would strengthen the relation between mental state language (our selection included mostly verbs), and possibly mask the relation to overall language development. It is also worth noting that the non-significant effects of general language on person inflections were observed in token analyses that used MLU or sentence repetition in their language measure. While pronouns contribute to MLU in verbs, verb inflections do not, which could result in the relation between MLU and pronouns. In any case, our findings confirm that not only first and second person pronouns, but also similar verb inflections are related to the development of social understanding.

The third novel aspect of the present study was the use of Czech, a morphologically complex language with a rich system of case inflections and verb conjugations. Using Czech is valuable in that it is a less studied language structurally different from English, and the similarities in the findings between these languages thus support cross-linguistic generalizability of the findings. The complexity of pronominal inflection and verb conjugation might suggest that they will primarily depend on the general development of language and grammar. However, the findings also confirmed the specific relations between person reference and mental state language, indicating that social cognition is an important influence in the early use of person reference devices, whether pronominal or other. Together with findings from Italian by Longobardi et al. (2019), this highlights the importance of research on languages with rich morphology.

Czech and Italian, the two languages in which person reference in verbs was examined, are pro-drop languages. Verb reference to persons in such languages has an especially prominent role because subject pronouns are not obligatory. Therefore, in many sentences, the reference to the subject may only be established using the verb inflection. It would be interesting to compare the current results with results from languages that are not pro-drop but do inflect verbs for first and second person, such as German. It is possible that the relation between social understanding and pronouns will be stronger in such languages because of the pronouns’ more consistent presence. However, this question needs to be tested empirically.

The dual nature of person reference

The key finding of the present study has important methodological implications for any research using personal pronouns or other linguistic measures of person reference. Such research should take into account both the linguistic and socio-cognitive aspects of person reference. If personal pronouns are used as markers of emerging social understanding, the level of language development in children should be controlled to obtain more precise estimates of children’s socio-cognitive level. In studies of language acquisition that focus on early pronouns or verb inflections, including control measures of social understanding would result in cleaner estimates of how pronouns depend on language skills. The same considerations apply to the research on person reference using verb conjugation. Because the issues of inflection are traditionally viewed as a topic of language research, the present findings indicate that extralinguistic factors may be relevant here.

While the present study shows that both social cognition and language development relate to person reference, it does not preclude the possibility that one of these domains is more important for the development of the other than vice versa, or that one affects development prior to the other. The present design could only examine the concurrent relations between the variables. It could be the case that the very first usages of person reference are only related to one of the two key predictors, language or social understanding. In particular, it is easy to see how language could be a limiting factor for any use of pronouns or verb inflection. If children’s linguistic system is not developed enough, these elements could not appear even if their social understanding is at an advanced level. On the other hand, children without adequate social understanding could still use pronouns or verb inflections occasionally, even though with a higher incidence of errors or inappropriate expressions. If this were the case, the relations between language and person reference should be initially stronger than those with social understanding. This question will have to be addressed using detailed longitudinal data on the early use of pronoun reference.

The work presented here is related to a long-standing question about the role of language in the development in social understanding, even though the scope of our study is different. While we focused on identifying unique and specific contributions of social understanding and language to person reference, we believe our findings emphasize mutual developmental interrelations between social understanding and language. Person reference can be tied to previous social understanding, as demonstrated by its relation to previous mirror self-recognition (Lewis & Ramsay, 2004) and visual perspectives (Ricard et al., 1999), but we show that it is also related to language development. Language development, on the other hand, has been repeatedly shown to affect the development of more advanced social understanding, such as the theory of mind (see Milligan et al., 2007, for a meta-analysis). It is thus likely that social cognition and language have mutual supportive relations during development. It will be important to study the specific effects of person reference in this respect. Because person-referring pronouns provide children with more opportunities to talk about self and others, they could facilitate the development of social understanding, and they could contribute to the effects of language on the theory of mind. Such questions must be addressed in longitudinal research to study the direction of causality, but our current results provide evidence that these relations likely exist.

Conclusions

The relation between early use of personal pronouns and the development of social understanding has been supported by numerous previous findings. The present study confirms that the relation between pronouns and social understanding holds even after accounting for language development, and demonstrates independent effects of language and social understanding on pronoun use. In addition, we show that person reference via verb inflection is related to social understanding in a similar way that pronouns are.

Footnotes

Appendix

Acknowledgements

The data collection was based at the Laboratory of Behavioral and Linguistic Studies (LaBeLS), a joint lab of the Institute of Psychology of the Czech Academy of Sciences and the Faculty of Arts of the Charles University.

Author contributions

Filip Smolík: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Software; Supervision; Writing-original draft; Writing-review & editing.

Veronika Bláhová: Data curation; Investigation; Project administration; Resources; Writing-review & editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was supported by the Czech Science Foundation (GA ČR) Grant No. GA16-15123S awarded to the first author, as well as by the institutional support of the Czech Academy of Sciences RVO 68081740.

ORCID iD

Filip Smolík

References

Bartak

Rutter

(1974). The use of personal pronouns by autistic children. Journal of Autism and Childhood Schizophrenia, 4(3), 217–222.

Bartsch

Wellman

H. M.

(1995). Children talk about the mind. Oxford University Press.

Bates

(1990). Language about me and you: Pronominal reference and the emerging concept of self. In Cicchetti

Beeghly

(Eds.), The self in transition: Infancy to childhood (pp. 165–182). University of Chicago Press.

Bishop

D. V. M.

(2003). Test for reception of grammar, version 2 (TROG 2). Pearson Assessment.

Bretherton

Beeghly

(1982). Talking about internal states: The acquisition of an explicit theory of mind. Developmental Psychology, 18(6), 906–921.

Brown

(1973). A first language: The early stages. Harvard University Press.

Brown

Fraser

(1964). The acquisition of syntax. Monographs of the Society for Research in Child Development, 29, 64–100.

Brownell

Ramani

G. R.

Zerwas

(2006). Becoming a social partner with peers: Cooperation and social understanding in one- and two-year-olds. Child Development, 77, 803–821.

Budwig

(1985). I, me, my and ‘name’: Children’s early systematizations of forms, meanings and functions in talk about the self. Papers and Reports on Child Language Development, 24, 30–37.

10.

Caravolas

Lervåg

Defior

Málková

G. S.

Hulme

(2013). Different patterns, but equivalent predictors, of growth in reading in consistent and inconsistent orthographies. Psychological Science, 24(8), 1398–1407.

11.

Charney

(1980). Pronoun errors in autistic children: Support for a social explanation. British Journal of Disorders of Communication, 15(1), 39–43.

12.

Chiat

(1982). If I were you and you were me: The analysis of pronouns in a pronoun-reversing child. Journal of Child Language, 9, 359–379.

13.

Chiat

(1986). Personal pronouns. In Fletcher

Garman

(Eds.), Language acquisition. Studies in first language development (pp. 339–356). Cambridge University Press.

14.

Courage

M. L.

Edison

S. C.

Howe

M. L.

(2004). Variability in the early development of visual self-recognition. Infant Behavior and Development, 27, 509–532.

15.

Dale

P. S.

Crain-Thoreson

(1993). Pronoun reversals: Who, when, and why? Journal of Child Language, 20, 573–589.

16.

Dunn

L. M.

Dunn

D. M.

(2007). Peabody Picture Vocabulary Test IV. American Guidance Service.

17.

Evans

K. E.

Demuth

(2012). Individual differences in pronoun reversal: Evidence from two longitudinal case studies. Journal of Child Language, 39(1), 162–191.

18.

Goodenough

F. L.

(1938). The use of pronouns by young children: A note on the development of self-awareness. The Pedagogical Seminary and Journal of Genetic Psychology, 52(2), 333–346.

19.

A. X.

Luyster

Hong

S. J.

Arunachalam

(2018). Personal pronoun usage in maternal input to infants at high vs. low risk for autism spectrum disorder. First Language, 38(5), 520–537.

20.

Jordan

R. R.

(1989). An experimental comparison of the understanding and use of speaker-addressee personal pronouns in autistic children. British Journal of Disorders of Communication, 24(2), 169–179.

21.

Kanner

(1943). Autistic disturbances of affective contact. Nervous Child, 2, 217–250.

22.

Kucharská

(2014). Riziko dyslexie. Pregramotnostní schopnosti a dovednosti a rozvoj gramotnosti v rizikových skupinách [The risk of dyslexia: Pre-literacy abilities and skills and the development of literacy in risk groups]. PedF UK v Praze.

23.

Lee

Hobson

R. P.

Chiat

(1994). I, you, me, and autism: An experimental study. Journal of Autism and Developmental Disorders, 24(2), 155–176.

24.

Levine

L. E.

Munsch

(2013). Child development: An active learning approach. SAGE.

25.

Lewis

Ramsay

(2004). Development of self-recognition, personal pronoun use, and pretend play during the 2nd year. Child Development, 75, 1821–1831.

26.

Longobardi

Spataro

Pecora

Bellagamba

(2019). Indices of social knowledge in Italian-speaking children: Relations between personal pronouns, verb conjugations and mental state language. First Language, 39(4), 396–411.

27.

Loveland

(1984). Learning about points of view: Spatial perspective and the acquisition of ‘i/you’. Journal of Child Language, 11, 535–556.

28.

MacWhinney

(2000). The CHILDES project: Tools for analyzing talk. Lawrence Erlbaum.

29.

Markova

Smolík

(2014). What do you think? The relationship between person reference and communication about the mind in toddlers. Social Development, 23, 61–79.

30.

Milligan

Astington

J. W.

Dack

L. A.

(2007). Language and theory of mind: Meta-analysis of the relation between language ability and false-belief understanding. Child Development, 78(2), 622–646.

31.

Morehead

D. M.

Ingram

(1973). The development of base syntax in normal and linguistically deviant children. Journal of Speech and Hearing Research, 16(3), 330–352.

32.

Naigles

L. R.

Cheng

Rattanasone

N. X.

Tek

Khetrapal

Fein

Demuth

(2016). ‘You’re telling me!’ The prevalence and predictors of pronoun reversals in children with autism spectrum disorders and typical development. Research in Autism Spectrum Disorders, 27, 11–20.

33.

Oshima-Takane

Cole

Yaremko

R. L.

(1993). Pronominal semantic confusion in a hearing-impaired child: A case study. First Language, 13(38), 149–168.

34.

Piaget

(1929). The child’s conception of the world. Routledge & Kegan Paul.

35.

Ricard

Girouard

C. P.

Gouin-Decarie

(1999). Personal pronouns and perspective taking in toddlers. Journal of Child Language, 26, 681–697.

36.

Seidlová Málková

Smolík

(2014) Diagnostika jazykového vývoje: Diagnostická baterie pro posouzení vývoje jazykových znalostí a dovedností dětí předškolního věku [Assessment of language development: An assessement battery for evaluating the development of language skills in preschool children]. Grada.

37.

Shatz

Wellman

H. M.

Silber

(1983). The acquisition of mental verbs: A systematic investigation of the first reference to mental state. Cognition, 14(3), 301–321.

38.

Shield

Meier

R. P.

Tager-Flusberg

(2015). The use of sign language pronouns by native-signing children with autism. Journal of Autism and Developmental Disorders, 45(7), 2128–2145.

39.

Smolík

(2002). Osvojování českých slovesných tvarů v raném věku [Acquisition of Czech verb forms in early age]. Československá Psychologie, 46, 450–461.

40.

Tanz

(1980). Studies in the acquisition of deictic terms. Cambridge University Press.

41.

Tomas

Dorofeeva

(2019). Mean length of utterance and other quantitative measures of spontaneous speech in Russian-speaking children. Journal of Speech, Language, and Hearing Research, 62(12), 4483–4496.

42.

Wang

Zamar

Marazzi

Yohai

Salibian-Barrera

Maronna

, et al. (2014). Robust [Computer software]. http://cran.r-project.org

43.

Wellman

H. M.

Phillips

A. T.

Rodriguez

(2000). Young children’s understanding of perception, desire, and emotion. Child Development, 71(4), 895–912.