Children benefit from morphosyntactic cues such as grammatical number marking and derivational rules when learning cardinals (one, two, three) and ordinals (first, second, third). However, children with developmental language disorder (DLD) often have difficulties with morphosyntax. We therefore hypothesized that they cannot use morphosyntactic cues to the same extent in their numerical development. We investigated this by comparing the performance of 27 kindergartners with (suspected) DLD (4;04-6;09) and 42 typically developing (TD) peers (4;03-6;04) on two counting measures (Rote Counting & Next Number and two measures of numerical knowledge (Give Me, and Tell Me). The study also explored how verbal and non-verbal background measures predicted children's counting ability and ordinal comprehension. In line with previous research, we found that children with DLD have counting deficits. In addition, our results revealed substantial delays in ordinal acquisition and qualitative differences between the groups. Unlike TD peers, children with DLD frequently interpreted ordinals as cardinals, and they also failed at different points in the counting routine. We argue that these deviating error-patterns reflect difficulties with rule-based learning. The explanatory role of language deficits was further substantiated by the fact that differences between the tasks were mediated by language abilities. These findings shed new light on numeracy deficits associated with DLD and emphasize the importance of not only investigating quantitative, but also qualitative differences between groups on numeracy tasks.
Although extensively studied in monolinguals, research on the development of functional categories in bilingual children remains limited. This paper contributes novel data through two corpus studies on the development of the left periphery in Germanic-Romance bilinguals acquiring German-Italian, German-Spanish, and Dutch-Italian. In the first study, I show that (some) complementizer phrase (CP) structures are produced early across all languages and children, while the development of topicalization strategies (e.g. topicalization vs. clitic left dislocation [CLLD]) shows strong language-specific effects. Topics are systematically early-acquired in their Germanic languages, whereas CLLD in Romance emerges significantly later-corroborating findings from monolingual studies. A second, supplementary corpus study on the development of cliticization furthermore shows that the late development of CLLD is not attributable to (similarly) late development of clitics; it is specific to CLLD. I explore potential explanations for this language-driven asymmetry, particularly their relative parametric complexity. Overall, I propose these results are problematic for approaches assuming crosslinguistically fixed and late development of CP, such as mainstream maturational approaches, favoring instead early CP development.
This study aimed to explore how classifier-based semantic skills and frequency of classifier input affect the learning of Mandarin shape classifiers among Mandarin-speaking children. A novel classifier learning task was developed, which allowed the examination of the influences of semantic skills and input frequency within the same task. Two novel classifiers were invented, each encoding distinct shape features not present in any real Mandarin classifiers. Typically-developing monolingual Mandarin-speaking children, including younger children (3-5 years old, n = 73) and older children (6-8 years old, n = 87), and adult controls (n = 62) completed the task. A learning phase was implemented immediately followed by a testing phase, assessing both production and comprehension of the two novel classifiers. Results showed that children and adults employed both the semantic strategy and the frequency strategy, while the weighting of these two interrelated strategies changed across ages. In particular, older participants showed more skilled semantic extraction and generalization related to classifiers than younger participants. Younger children were more likely to use input frequency to assist their learning of novel classifiers, which could be attributed to an immature development of semantic skills for classifiers. With developed semantic skills, adults may be able to extract the semantic features even with a low dosage of classifier input; as a result, the effect of frequency was not found for adults at the group level. Despite the overall tendency regarding the roles of semantic skills and input frequency in development, individual differences emerged within each age group, revealing varying degrees of use of the semantic or the frequency strategy.
This study examines whether Flemish students can learn the French (L2) passive through structural priming, which is the tendency to repeat grammatical structures. We conducted two experiments: a between-language (first language-second language [L1-L2]) and a within-language (L2) one. In Experiment 1, 50 Flemish secondary school students without prior exposure to the French passive described pictures (24 items) in French after hearing a Dutch prime: an active, a passive, or a baseline sentence. Notably, Dutch and French passives use different auxiliary verbs (e.g. "The cook is being carried by the butcher" is De kok wordt ["becomes"] gedragen door de slager and Le chef est ["is"] port & eacute; par le boucher, respectively). Midway through the experiment, we implemented a within-language intervention block of four French passive primes containing the same verb as the target. As hypothesized, no passives were produced before the intervention. Afterward, only a few participants formed passive sentences. Nevertheless, the intervention had a significant effect on overall passive production, and there was marginally significant passive priming. Additionally, both L2 proficiency and comprehension of the target structure correlated positively with passive production. To complement these findings, we conducted the same experiment with only French primes in Experiment 2, as it is often assumed that within-L2 priming precedes cross-linguistic priming. However, there was no passive priming and only a marginally significant intervention effect, with lower passive production and comprehension than in Experiment 1. Our results suggest that learners first need to understand L2 syntactic structures before they can be primed to produce them.
Motion verbs refer to a figure's path and manner of motion. Previous work on the acquisition of motion verbs has demonstrated that learners are sensitive to their language's distribution of motion verbs: learners that are acquiring languages which are more likely to encode the path of motion default to mapping path of motion for novel verbs; learners acquiring languages that are more likely to encode the manner of motion default to mapping manner. The current work adds to the literature on the acquisition of motion verbs by exploring the nature of children's sensitivity to two additional visual elements, agent identity and context or location of the event, while mapping novel motion verbs. In experiment 1, we find that children are more sensitive to a context change than an agent change when extending new motion verbs. We conclude that they are more likely to encode context as potentially relevant to verb meaning than agent identity. In a series of two follow-up experiments, we expand on the context cue. In experiment 2a, we gather adult judgments of images and videos in variety of contexts as accurately depicting specific nouns and verbs. We find that adults are influenced by context of an event when judging depictions of verbs, but not by an object's context when judging depictions of nouns. We also find that certain contexts are more influential for verb meaning than others. In experiment 2b, we use the findings from the adult judgment task to manipulate event context in a verb learning task with children. We find that children are more likely to encode more restrictive contexts as relevant to verb meaning. We hypothesize that children's sensitivity to context in mapping motion verbs reflects knowledge of the underlying adult representation of context as important for some motion verb meanings.
In this article, we present a study comparing the production of German requesting forms in adults and in children from 4 to 11 years, depending on different social scenarios that varied in social distance of the addressee and in the urgency of the request. We explored the variation of the elicited utterances along the two different dimensions of sentence type (declarative, imperative, and interrogative) and whether participants used a modal verb. If so, we further investigated whether they employed subject-internal (Can I & mldr;) or subject-external modality (Can you & mldr;). Our results show that urgency had no effect on the choice of utterances in both adult speakers and children. On the other hand, social distance had an impact on the linguistic forms used, but in different ways for adults and children: While adults varied their use of sentence types, children did not display variation in the domain of sentence types, but rather varied the use of subject-internal versus subject-external modality. Moreover, our study indicates the acquisition path that the use of subject-external modality increases with age, and we highlight that this dovetails with the development of general cognitive capacities.
Research on language acquisition emphasizes the pivotal role of early linguistic input in shaping children's development. However, most studies focus on English, limiting linguistic and cultural diversity. We extend prior work by characterizing the lexical and morphosyntactic profile of infant-directed speech (IDS) across the first year in Argentinian Spanish (3-5, 6-8, and 9-10 months). Sixty-four mother-infant dyads, native speakers of Argentinian Spanish with infants aged 3-10 months, participated. Semi-structured play sessions provided a corpus analyzed with CLAN software. Descriptive analyses indicated age-related patterns in Argentinian Spanish IDS. Infants aged 6-8 months received less diverse vocabulary, whereas older infants encountered greater diversity. Older infants heard more verbs, whereas younger ones heard fewer. Infants aged 3-5 months were exposed to fewer closed-class words, contrasting with those aged 9-11 months, who heard more of the closed-class words. There was a decline in the noun/verb ratio with increasing age, while the open-/closed-class word ratio peaked at intermediate ages. The present study both contributes to our understanding of how IDS structure changes during the preverbal period and extends the evidence base for non-English languages. Our results highlight the dynamic nature of maternal speech during the preverbal stage and underscore the value of examining diverse linguistic environments to capture the complexity of language-acquisition processes across different linguistic and cultural contexts.
This study investigates the sources of the possible interpretations of negative disjunctive statements (e.g. Girafa nu a cules ghiocelul sau trandafirul 'The giraffe did not pick the snowdrop or the rose') in Romanian children and adults. Previous research highlights cross-linguistic variation among adults: English-speaking adults allow both neither nor and not both interpretations, while Japanese-speaking adults only allow the not both interpretation. In contrast, children across a number of languages have been shown to prefer the stronger neither nor interpretation. We attempt to connect this area of research with studies showing that children interpret positive disjunctive statements (e.g. Girafa a cules ghiocelul sau trandafirul 'The giraffe picked the snowdrop or the rose') in an inclusive or conjunctive manner (and in particular with studies that argue that the conjunctive interpretation is the result of scalar strengthening). If the source of strong meanings across polarities is strengthening, then we expect the same children who prefer neither nor interpretations of negative disjunctive statements to be the same children who prefer conjunctive interpretations of positive disjunctive statements. We conducted a Truth Value Judgment Task (TVJT) with Romanian 5-year-old children and a group of adult controls to examine their behavior in negative and positive disjunctive statements and to investigate a possible positive correlation between strong interpretations across sentence polarities. Our results indicate that both groups behaved relatively similarly: in negative contexts, they predominantly favored the neither nor interpretation, while in positive contexts, they predominantly favored an inclusive interpretation. These results do not support a positive correlation between strong interpretations of positive and negative disjunctive statements. Instead, we argue that strong interpretations may have different sources in sentences of differing polarity, namely, strengthening in positive statements and scope in negative ones.
As the interest in language models (LMs) gains traction in academia, there has been an impulse to compare LMs with children- the idea seems to be that if these models perform well, perhaps this reveals something about the acquisition process. The problem with this line of reasoning is that it is often at odds with the nature of the input, as LMs are typically trained on billions of words extracted from written sources. In order to transcend the difficulties with making such a comparison, there have been attempts to scale down these models to better approximate the input that children are exposed to in real life. In this research, we focus on word learning and compare human and model performance in the adjectival domain. Here we compare humans and models directly by first running an experimental study with participants using a modified version of the Human Simulation Paradigm, in which we analyze human learners' abilities to generate frame-compliant responses to syntactic frames. We then run the same task with BabyBERTa, an LM that is trained specifically on Child Language Data Exchange System (CHILDES) data. For human learners, we find that frames trigger adjective responses, and responses shift as a result of the frame, and that additional semantic information enhances subcategorization patterns. Conversely, BabyBERTa diverges from our participants-at times failing to predict adjectives, and not improving when semantic information is included. More troubling, similar mismatches continue at different scales of LMs, suggesting a flaw with the claim that models are representative of acquisition. While LMs can generate grammatical results, they fail to capture human performance. Although the Human Simulation Paradigm (HSP) and BabyBERTa have both been argued to replicate the acquisition process, divergent results suggest that models are not learning in the same way as humans.
Perlmutter's unaccusativity hypothesis suggests that intransitive verbs can be classified into two verb classes: unergative and unaccusative. The arguments of unergatives exhibit syntactic and semantic properties similar to the subjects of transitive verbs, whereas the arguments of unaccusatives share resemblances in these aspects with the objects of transitive verbs. In Mandarin, the inversion of subjects serves as a diagnostic sentence for distinguishing between these two verb classes. Although previous studies employing this diagnostic have revealed semantics-driven gradients between unergatives and unaccusatives, this has not been extensively explored in the domain of child language acquisition. Our study examined whether Mandarin-speaking children's acquisition of unaccusativity is mediated by more articulated semantic event types, distinguished as core and non-core in Sorace (2000) hierarchy. Through a forced-choice task and an acceptability judgment task, we found that children under the age of 7 struggled to differentiate between change-of-state verbs (non-core unaccusatives) and unergatives using subject-inversion. However, they exhibited this ability with core unaccusatives, namely change-of-location verbs. Our findings suggest that Mandarin-speaking children develop an early understanding of unaccusativity for core unaccusatives, but not for non-core unaccusatives. We propose that the acquisition of Mandarin unaccusativity may be better understood through a more sophisticated framework, such as Sorace's hierarchy, where the semantic differences within unaccusatives can explain the trajectory of learning.
Children acquiring sign language (SL) form a heterogeneous group in which the hearing statuses of the children, parents, and family members, as well as the children's access to language and opportunities to use the languages acquired, vary across children. The context of language acquisition and features of the linguistic environment among these children are unique but rarely studied, especially in smaller SL communities. This research explores the features of language exposure, language use, and the multilingual nature of the linguistic environment among children acquiring Finnish Sign Language (FinSL). Altogether, 188 hearing and deaf children and adolescents (0;8-17;10) acquiring FinSL participated in this study, and the data were collected by a parental questionnaire. Children were placed into four groups based on the hearing status of the child and the parents (hearing children of deaf adults, deaf of deaf, deaf of hearing, and hearing of hearing). The features of children's language exposure, frequency of meeting other language users, and children's language use with different interlocutors were compared among the child groups. The results revealed that different SLs, spoken languages, and multilingualism play a considerable role in children's linguistic environment but display distinct patterns across the child groups. The observed differences and similarities in language exposure, frequency of interaction with other language users, and patterns of language use between groups help illustrate the complex, dynamic nature of children's multilingualism. This highlights the importance of acknowledging children's linguistic environment as multifaceted and suggests describing children's linguistic environment in a more nuanced, detailed manner in further studies.
Deaf children learning sign language must direct gaze to their caregivers to perceive linguistic input. To determine how much linguistic input children perceive from their parents, we analyzed naturalistic play sessions between signing deaf children (N=24) aged 9-60 months (M=35 months) and their deaf (n = 15) and hearing (n = 9) parents. Sessions were videotaped and coded for all ASL signs, gaze and attention-getting cues. Across all participants, 30% of sign tokens were fully perceived, and an additional 17% were partially perceived. Factors predicting sign perception included child age, having a deaf parent, and the sign's position in the utterance. Our findings suggest that parents using ASL are sensitive to their child's gaze and time their utterances so they are most likely to be seen.
This study examines the development of pragmatic competence in request formation among Russian-speaking monolingual and bilingual Russian-Hebrew children aged 5 to 8, for whom Russian is their Heritage Language (HL) and Hebrew is their Societal Language (SL). Using a pragmatic elicitation task, we analyzed responses from 56 children across formal and informal scenarios to assess differences in request strategies, including address forms, sentence types, modals, morphosyntactic particles, and lexical mitigators like 'please.' Results indicate that monolingual peers displayed a broader and more contextually appropriate range of strategies, while bilingual children demonstrated reduced sensitivity to formality, often relying on informal address forms and impersonal modals in contexts that typically demand formal requests. This pattern might stem from cross-linguistic influence (CLI) from Hebrew and limited exposure to formal Russian, which may shape both early-acquired pragmatic features, such as basic politeness markers, and later-developing phenomena, including nuanced formality distinctions and particle use. To conclude, we suggest that pragmatic development in bilingual children reflects more than simple transfer or developmental delay; it emerges from a dynamic interplay between linguistic competence, environmental input, and socio-cultural adaptation.
To determine whether morphological salience plays a role in the comprehension and production of grammatical aspect within a single language, this study tested the comprehension and production of the pre-verbal marker zai- and the post-verbal markers -le and -zhe in four- to six-year-old Mandarin-speaking typically developing (TD) children and age-matched children with developmental language disorder (DLD), while strictly controlling the verb types combined with them. The results showed that both TD children and children with DLD displayed similar performance in the comprehension of the three aspect markers. In the production task, however, children in both groups performed better on the pre-verbal marker than on the post-verbal ones, and more errors were produced with the latter than with the former. This selective effect of modality could arguably be related to different processing mechanisms engaged by language comprehension versus production: the facilitating effect of morphological salience would be more impactful in production as it arguably solicits more processing resources than in comprehension. In addition, the impact of lexical aspect on grammatical aspect was further confirmed in the comprehension and production tasks because better performance was found when aspect markers were combined with their prototypical verbs than with non-prototypical ones. The findings of this study call for explicit training on the less salient markers in children with DLD during therapeutic interventions targeting aspectual production, especially their combinations with non-prototypical verbs.
Children's difficulties with long-distance dependencies have been ascribed to intervention effects that arise from moving a noun phrase (NP) argument over another. This study investigates intervention effects in the comprehension of b & egrave;i-passives by Mandarin-speaking children. Some syntactic evidence suggests that short b & egrave;i-passives do not have a syntactically projected external argument (unlike long b & egrave;i-passives) and also cannot be interpreted as adjectival structures (unlike short passives in English). We explore the predictions of the Intervention Hypothesis by comparing children's performance on short and long passives and by testing whether Number and Shape features modulate these effects in Mandarin long passives. Although both features are encoded on classifiers, neither triggers movement or agreement in Mandarin. Results from our two-choice sentence-picture matching task, conducted with 80 Mandarin-speaking children aged 3 to 6, show that children exhibit significant difficulties comprehending long passives compared to short passives and actives, supporting the Intervention Hypothesis. Moreover, mismatches in Number and Shape do not ameliorate their performance on long passives, suggesting that not all linguistic features are equally relevant in the computation of syntactic dependencies in child grammar across languages. This supports a language-specific grammatical approach to intervention effects, wherein only features that trigger syntactic processes in a given language are represented in the computation of intervention in that language.
Scalar diversity, the phenomenon that different scalar expressions give rise to scalar implicatures (SI) at different rates, is well established in native speakers, but much less is known about it in second language (L2) SI derivation. This study investigates whether first language (L1)-Chinese-L2-English learners exhibit scalar diversity, how their patterns compare to native English speakers, and whether perceived distance between scale mates contributes to L2 diversity in the same way as in native speakers. Experiment 1 used an inference task to measure SI rates for 43 lexical scales and compared learners' responses with native speakers. Experiment 2 measured learners' perceived distance between scale mates for the same lexical scales and compared them with native ratings. Results showed that L2 learners exhibit scalar diversity, with SI rates varying considerably across lexical scales, and their rank ordering of scales largely aligns with that of native speakers. At the same time, our findings reveal divergences: variations in learners' SI rates are partly explained by structural properties such as semantic distance, though to a lesser extent compared to native speakers, and differences in SI rates between the two groups follow a heterogeneous rather than uniform pattern across scales. Proficiency and differences in perceived scale structure, reflected in distance ratings by L2 learners and native speakers, are potential contributors to learners' divergence from native speakers in SI behavior. These findings are discussed in relation to the Interface Hypothesis, which holds that both processing limitations and representational differences contribute to the challenges observed in phenomena at the external interface.
There is cross-linguistic variation in the strategies languages use to answer polar questions. Specifically, languages differ in the type of answers that can be produced: only a response particle, only a continuation, or a combination of both. Among languages that allow response particles, there is further variation in how these are used to answer negative polar questions. In Positive-Negative languages, which Italian has been claimed to belong to, positive response particles are used to express positive polarity, whereas negative response particles mainly express negative polarity. This article experimentally investigates these two aspects of answers to polar questions in adult and child Italian. We do so by eliciting answers to both positive and negative polar questions and analyzing the type of answers that were given and their form. To our knowledge, this is the first study investigating the answering system in both positive and negative polar questions in adult and child Italian. Our results from adults suggest that Italian is mostly, but not strictly, a Positive-Negative language. In addition, we show that 4- to 5-year-old Italian children are able to answer positive and negative polar questions, behaving in line with adults, although they have yet to fully acquire the ability to produce mandatory continuations in certain contexts. Lastly, as a corollary result, we show that our data experimentally support the validity of the markedness scale in adult and child Italian: some responses are more marked than others and are mapped to more specialized forms.
The process of word-learning is often thought of as mapping forms to meanings, but what about forms that lack meaning? This paper explores a class of such meaningless forms, English expletives it and there, which co-occur with referential forms. We investigate whether children distinguish these expletives from their referential counterparts and whether they favor referential interpretations. In two comprehension experiments, we examined how 4-year-olds interpret these forms in both ambiguous and unambiguous contexts. Children accessed the expected readings in unambiguous contexts and showed no referential bias in ambiguous contexts, suggesting that by age 4, children understand and readily access the expletive uses of these words.
