A persistent mystery in language acquisition is how speakers are able to learn seemingly arbitrary distributional restrictions. This article investigates one such case: the fact that speakers resist using certain adjectives prenominally (e.g. ??the asleep man). Experiment 1 indicates that speakers tentatively generalize or CATEGORIZE the distributional restriction beyond their previous experience. Experiment 2 demonstrates that speakers are sensitive to STATISTICAL PREEMPTION—that is, speakers learn not to use a formulation if an alternative formulation with the same function is consistently witnessed. Moreover, they are able to generalize the restriction to apply to other members of the category as well. Finally, experiment 3 finds evidence that speakers DISCOUNT a pseudo-preemptive context, rationally ignoring it as uninformative.

Children (aged five-to-six and nine-to-ten years) and adults rated the acceptability of well-formed sentences and argument-structure overgeneralization errors involving the prepositional-object and double-object dative constructions (e.g. Marge pulled the box to Homer/Marge pulled Homer the box). In support of the entrenchment hypothesis, a negative correlation was observed between verb frequency and the acceptability of errors, across all age groups. Adults additionally displayed sensitivity to narrow-range semantic constraints on the alternation, rejecting double-object dative uses of novel verbs consistent with prepositional-dative-only classes and vice versa. Adults also provided evidence for the psychological validity of a proposed morphophonological constraint prohibiting Latinate verbs from appearing in the double-object dative. These findings are interpreted in the light of a recent account of argument-structure acquisition, under which children retreat from error by incrementally learning the semantic, phonological, and pragmatic properties associated with particular verbs and particular construction slots.

Noun-class prefixes are obligatory in most Bantu languages. However, the Sotho languages (Sesotho, Setswana, Sepedi) permit a subset of prefixes to be realized as null at the intersection of ‘unmarked’ phonological, syntactic, and discourse conditions. This raises the question of how and when the licensing of null prefixes is learned. Using longitudinal data from three Sesotho-speaking children, this article shows that the conditions needed to license null prefixes have been learned before the age of three, suggesting early abilities for grammatical generalization even at the intersection of different levels of linguistic structure. The implications for learnability theory and Bantu linguistic structure more generally are discussed.

The status of subject clitics in French has been heavily debated (Kayne 1975, Rizzi 1986, Roberge 1990, Auger 1994b, Miller & Sag 1997, De Cat 2007b, and many others). Distributional properties of French subject clitics have led Kayne (1975), Rizzi (1986), and others to analyze them as argument-bearing elements occupying canonical subject position, cliticizing to the verb only at the level of the phonology. While this hypothesis enjoys a wide following, a growing body of evidence suggests that it fails to capture patterns of subject-clitic use in colloquial French dialects/registers (Roberge 1990, Auger 1994b, Zribi-Hertz 1994, Miller & Sag 1997). Using new evidence from prosodic and corpus analyses, speaker judgments, and crosslinguistic typology, this article argues that (i) European Colloquial French exhibits differences from Standard French that impact how subject clitics are best analyzed, and more specifically (ii) subject clitics in European Colloquial French are affixal agreement markers, not phonological clitic arguments.

We investigate the development of word order in German children's spontaneous production of complement clauses. From soon after their second birthday, young German children use both verb-final complements with complementizers and verb-second complements without complementizers. By their third birthday they use both kinds of complement clauses with a variety of complement-taking verbs. Early in development, however, verb-final complements and verb-second complements are used with separate sets of complement-taking verbs, and they are used with separate sets of item-specific main-clause phrases. For example, initially phrases such as ‘I want to see’ were used exclusively with verb-final complements, whereas phrases such as ‘do you see’ and ‘you have to say’ were used exclusively with verb-second complements. Only later in development—when specific complement-taking verbs were used with both verb-second and verb-final complements, with a greater variety of main-clause phrases, and when specific main-clause phrases were used with both verb-second and verb-final complements—was there evidence for structural links between these various, item-based, complement-clause constructions.

In recent years linguists have gained new insight into human language capacities on the basis of results from linguistics and biology. The so-called biolinguistic enterprise aims to fill in the explanatory gap between language and biology, on both theoretical and experimental grounds, hoping to reach a deeper understanding of language as a phenomenon rooted in biology. This research program is taking its first steps, and it has already given rise to new insights on the human language capacity, as well as to controversies, echoing debates that go back to the earlier days of generative grammar. The present discussion piece provides a high-level characterization of biolinguistics. It highlights the main articulation of this research program and points to recent studies linking language and biology. It also compares the biolinguistic program, as defined in Chomsky 2005 and Di Sciullo & Boeckx 2011, to the view of the human language faculty presented in Jackendoff 2002 and Culicover & Jackendoff 2005, and to the discussion in Jackendoff 2011.

We discuss three points relevant to Kissine's (2021) target article on autism spectrum disorder (ASD) and its implications for theories of language acquisition. First, we argue that individuals with ASD do have social communication abilities and that these are linked to their language abilities; therefore, ASD does not provide evidence that language competence and language use are dissociable. Second, we argue that typically developing children show remarkable abilities to learn new words in noninteractive situations, and thus these abilities are not unique to ASD. Third, we point out that even noninteractive situations can serve as models for, and can implicate, children's social communication abilities. In sum, we agree with Kissine that children (with and without ASD) are skilled language learners, able to take diverse paths to learning and to benefit from many different kinds of learning situations. However, as to whether these abilities in themselves demonstrate a threat to constructionist approaches to language acquisition, we do not think that Kissine has accrued a compelling case.

Subjacency characterizes a set of phenomena whose acquisition must be explained by any proposal for human language learning. We take a broader perspective than previous responses to Ambridge, Pine, and Lieven (2014), arguing that they have not shown that this UG principle is ‘redundant’ because their proposed alternative does not take into account firmly established constraints on A-bar dependencies. We illustrate a range of challenges for theories hoping to reduce subjacency to independently motivated, primarily nonsyntactic constraints: they must include a way to account for attested crosslinguistic variation in island effects, the cross-construction generality of island effects, and the effects of resumption and of WH-in-situ on island behavior.

Pater's (2019) target article proposes that neural networks will provide theories of learning that generative grammar lacks. We argue that his enthusiasm is premature since the biases of neural networks are largely unknown, and he disregards decades of work on machine learning and learnability. Learning biases form a two-way street: all learners have biases, and those biases constrain the space of learnable grammars in mathematically measurable ways. Analytical methods from the related fields of computational learning theory and grammatical inference allow one to study language learning, neural networks, and linguistics at an appropriate level of abstraction. The only way to satisfy our hunger and to make progress on the science of language learning is to confront these core issues directly.

Given that nouns rarely appear in isolation in French, infants acquiring the language must often retrieve the underlying representation of vowel-initial lexical forms from liaison contexts that provide conflicting information about the initial phoneme. Given this ambiguity, how do learners represent these nouns in their lexicons, and how do these representations change as learners’ knowledge of liaison and the lexicon become more adult-like? To explore this question, we analyze the types of errors children make, in both naturalistic and elicited speech, and how these are affected by input frequency. In doing so, we evaluate two major proposals for how children's early representations of liaison develop. The first model, couched in a constructionist framework, predicts relatively late mastery of liaison (age five or older) and heavy dependence on the contexts in which a particular noun appears in the input. The second model takes an approach to liaison development that integrates it more closely with general phonological development and predicts relatively early mastery (by age three). The results of a corpus study reveal that by age three children are correctly producing liaison in the nominal domain and that their production errors are consistent with a phonological model of liaison acquisition. An elicitation task demonstrates that three-year-olds succeed at learning and correctly apply their knowledge of liaison to new nouns following brief exposure, though their productions continue to be influenced by nouns’ input distributions. Taken together, our findings suggest that by age three children are well on their way to adult-like representations of liaison. A phonologically based model, incorporating the effect of distributional context on early errors, provides a better overall fit to the data we present.

Why can't we say the asleep cat? There is a class of adjectives in English, all of which start with a schwa (e.g. afraid, alone, asleep, away, etc.), that cannot be used attributively in a prenominal position. A frequently invoked strategy for the acquisition of such negative constraints in language is to use indirect negative evidence. For instance, if the learner consistently observes paraphrases such as the cat that is asleep, then the conspicuous absence of the asleep cat may be a clue for its ungrammaticality (Boyd & Goldberg 2011). This article provides formal and quantitative evidence from child-directed English data to show that such learning strategies are untenable. However, the child can rely on positive data to establish the distributional similarities between this apparently idiosyncratic class of adjectives and locative particles (e.g. here, over, out, etc.) and prepositional phrases. With the use of an independently motivated principle of generalization (Yang 2005), the ungrammaticality of attributive usage can be effectively extended to the adjectives in question.

In this commentary, we emphasize the importance of the observations presented by Kissine (2021) in his target article for our understanding of the nonmonolithic nature of pragmatics. Our first aim is to complement Kissine's argument, discussing some critical cases of linguistic processes that demonstrate the need for a finer-grained characterization of pragmatic phenomena. In addition, we report some findings that suggest that perspective taking may emerge as atypical even in autistic individuals who appear to be able to pass the standard theory-of-mind tasks. Our second aim is thus to argue that, albeit difficult to spot in experimental settings, the atypical theory-of-mind profile of low- and high-functioning autistic individuals is mirrored in their difficulties in everyday sociocommunicative interactions. Moreover, we claim that subtle differences in perspective-taking abilities may explain the highly heterogeneous linguistic profile of autistic individuals. Ultimately, with this commentary we wish to highlight the need for an increased appreciation of the role of perspective taking in typical and atypical language acquisition. This is crucial to our understanding of the nature of language acquisition, and can shed more light on the interaction between language and other aspects of human cognition.

In his target article, ‘Autism, constructionism, and nativism’, Kissine (2021) argues that data from autism should be taken into consideration in the debate about L1 acquisition. This paper responds to Kissine's piece by pointing out several of its underlying assumptions and suggesting directions for future research on the topic. Traditional framings of autism as a deficit have recently been challenged in favor of an identity-based approach, the neurodiversity paradigm, which suggests that autistic speech should not be measured in terms of its resemblance to nonautistic speech and that literature on intercultural miscommunication may offer insights into autistic communication. There are some indications that distinct autistic discourse practices may be identifiable in communities of practice, and studies on autistic literacy could benefit from considering the theoretical perspectives found in literature on multimodality and translanguaging. Finally, research on language acquisition might be strengthened by the incorporation of holistic neurocognitive theories about autistic minds.

Previous research on the acquisition of noun classification systems (e.g. grammatical gender) has found that child learners rely disproportionately on phonological cues to determine the class of a new noun, even when competing semantic cues are more reliable in their language. Culbertson, Gagliardi, and Smith (2017) use artificial language learning experiments with adults to argue that this likely results from the early availability of phonological information during acquisition. Learners base their initial representations on formal features of nouns, only later integrating semantic cues from noun meanings. Here, we use these same methods to show that early availability affects cue use in children (six- to seven-year-olds) as well. However, we also find evidence of developmental changes in sensitivity to semantics; when both cue types are simultaneously available, children are more likely to rely on phonology than adults are. Our results suggest that both early availability and a bias favoring phonological cues contribute to children's overreliance on phonology in natural language acquisition.

This article investigates how children learn an infinitely expanding ‘universal’ system of classificatory kinship terms. We report on a series of experiments designed to elicit acquisitional data on (i) nominal kinterms and (ii) sibling-inflected polysynthetic morphology in the Australian language Murrinhpatha. Photographs of the participants' own relatives are used as stimuli to assess knowledge of kinterms, kin-based grammatical contrasts, and kinship principles, across different age groups. The results show that genealogically distant kin are more difficult to classify than close kin, that children's comprehension and production of kinterms are streamlined by abstract merging principles, and that sibling-inflection is learned in tandem with number and person marking in the verbal morphology, although it is not fully mastered until mid to late childhood. We discuss how the unlimited nature of Australian kinship systems presents unusual challenges to the language learner, but suggest that, as everywhere, patterns of language acquisition are closely intertwined with children's experience of their sociocultural environment.

This article examines the acquisition of noun classes in Tsez, looking in particular at the role of noun-internal distributional cues to class. We present a new corpus of child-directed Tsez speech, analyzing it to determine the proportion of nouns that children hear with this predictive information and how often this is heard in conjunction with overt information about noun class agreement. Additionally, we present an elicited production experiment that uncovers asymmetries in the classification of nouns with versus without predictive features and by children versus adults. We show that children use noun-internal distributional information as a cue to noun class out of proportion with its reliability. Children are biased to use phonological over semantic information, despite a statistical asymmetry in the other direction. We end with a discussion of where such a bias could come from.

We argue for an extension of the proposal that grammars are in part shaped by processing systems. Hawkins (2014) and others who have advanced this idea focus primarily on parsing. Our extension focuses on production, and we use that to explore explanations for certain subject/object asymmetries in extraction structures. The phenomenon we examine, which we term the mirror asymmetry, runs in opposite directions for within-clause and across-clause (long-distance) extraction, showing a preference for subject extraction in the former and for object extraction in the latter. We review several types of evidence suggesting that the mirror asymmetry and related phenomena are best explained by an account of the formation of grammars that assigns an important role to properties of sentence planning in production.

I completely agree with Ambridge, Pine, and Lieven (AP&L) that anyone proposing a learning-strategy component needs to demonstrate precisely how that component helps solve the language acquisition task. To this end, I discuss how computational modeling is a tool well suited to doing exactly this, and that it has the added benefit of highlighting hidden assumptions underlying learning strategies. I also suggest general criteria relating to utility and usability that we can use to evaluate potential learning strategies. As a response to AP&L's request for Universal Grammar (UG) components that actually do work, I additionally provide a review of one potentially UG component that is part ofa successful learning strategy for syntactic islands, and that satisfies the evaluation criteria I propose.

The nominal anaphoric element one has figured prominently in discussions of linguistic nativism because of an important argument advanced by C. L. Baker (1978). His argument has been frequently cited within the cognitive and linguistic sciences, and has provided the topic for a chain of experimental and computational psycholinguistics papers. Baker's crucial grammaticality facts, though much repeated in the literature, have not been critically investigated. A corpus investigation shows that his claims are not true: one does not take only phrasal antecedents, but can also take nouns on their own, including semantically relational nouns, and can take various of-PP dependents of its own. We give a semantic analysis of anaphoric one that allows it to exhibit this kind of freedom, and we exhibit frequency evidence that goes a long way toward explaining why linguists have been inclined to regard phrases like the one of physics or three ones as ungrammatical when in fact (as corpus evidence shows) they are merely dispreferred relative to available grammatical alternatives. The main implication for the acquisition literature is that one of the most celebrated arguments from poverty of the stimulus is shown to be without force.

Ambridge, Pine, and Lieven (AP&L) claim that the knowledge attributed to children by the proponents of UG does not account for language acquisition, bringing evidence from several domains. In this response, we take issue with their claims with respect to two domains. In the case of categories, where distributional learning plays an important role, we argue that AP&L fail to recognize recent analyses showing that abstract representations yield better quantitative models for early child data. In the case of subjacency, we provide several empirical arguments against their claim that it can be reduced to some general discourse-pragmatic principles.