More Than Words: The Role of Multiword Sequences in Language Learning and Use

The ability to convey our thoughts using an infinite number of linguistic expressions is one of the hallmarks of human language. Understanding the nature of the psychological mechanisms and representations that give rise to this unique productivity is a fundamental goal for the cognitive sciences. A long-standing hypothesis is that single words and rules form the basic building blocks of linguistic productivity, with multiword sequences being treated as units only in peripheral cases such as idioms. The new millennium, however, has seen a shift toward construing multiword linguistic units not as linguistic rarities, but as important building blocks for language acquisition and processing. This shift—which originated within theoretical approaches that emphasize language learning and use—has far-reaching implications for theories of language representation, processing, and acquisition. Incorporating multiword units as integral building blocks blurs the distinction between grammar and lexicon; calls for models of production and comprehension that can accommodate and give rise to the effect of multiword information on processing; and highlights the importance of such units to learning. In this special topic, we bring together cutting-edge work on multiword sequences in theoretical linguistics, first-language acquisition, psycholinguistics, computational modeling, and second-language learning to present a comprehensive overview of the prominence and importance of such units in language, their possible role in explaining differences between first- and second-language learning, and the challenges the combined findings pose for theories of language.     

The Role of Multiword Building Blocks in Explaining L1–L2 Differences

Why are children better language learners than adults despite being worse at a range of other cognitive tasks? Here, we explore the role of multiword sequences in explaining L1–L2 differences in learning. In particular, we propose that children and adults differ in their reliance on such multiword units (MWUs) in learning, and that this difference affects learning strategies and outcomes, and leads to difficulty in learning certain grammatical relations. In the first part, we review recent findings that suggest that MWUs play a facilitative role in learning. We then discuss the implications of these findings for L1–L2 differences: We hypothesize that adults are both less likely to extract MWUs and less capable of benefiting from them in the process of learning. In the next section, we draw on psycholinguistic, developmental, and computational findings to support these predictions. We end with a discussion of the relation between this proposal and other accounts of L1–L2 difficulty.     

Maturational constraints on the recruitment of early processes for syntactic processing

An enduring question in the study of second-language acquisition concerns the relative contributions of age of acquisition (AOA) and ultimate linguistic proficiency to neural organization for second-language processing. Several ERP and neuroimaging studies of second-language learners have found that neural organization for syntactic processing is sensitive to delays in second-language acquisition. However, such delays in second-language acquisition are typically associated with lower language proficiency, rendering it difficult to assess whether differences in AOA or proficiency lead to these effects. Here we examined the effects of delayed second-language acquisition while controlling for proficiency differences by examining participants who differ in AOA but who were matched for proficiency in the same language. We compared the ERP response to auditory English phrase structure violations in a group of late learners of English matched for grammatical proficiency with a group of English native speakers. In the native speaker group, violations elicited a bilateral and prolonged anterior negativity, with onset at 100 msec, followed by a posterior positivity (P600). In contrast, in the nonnative speaker group, violations did not elicit the early anterior negativity, but did elicit a P600 which was more widespread spatially and temporally than that of the native speaker group. These results suggest that neural organization for syntactic processing is sensitive to delays in language acquisition independently of proficiency level. More specifically, they suggest that both early and later syntactic processes are sensitive to maturational constraints. These results also suggest that late learners who reach a high level of second-language proficiency rely on different neural mechanisms than native speakers of that language.     

Chunk‐Based Memory Constraints on the Cultural Evolution of Language

In the fields of linguistics and cognitive science, considerable attention has been devoted to the question of how linguistic structure emerged over evolutionary time. Here, we highlight the contribution of a fundamental constraint on processing, the Now‐or‐Never bottleneck. Language takes place in the here and now, with the transience of acoustic speech signals and our exceedingly limited memory for sound sequences requiring immediate processing. To overcome this bottleneck, the cognitive system employs basic chunking mechanisms to rapidly compress and recode incoming linguistic input into increasingly abstract levels of representation, thereby prolonging its retention in memory. Our suggestion is that these chunk‐based memory processes influence linguistic structure across multiple time scales. Chunk‐based memory constraints govern language acquisition and processing on the level of the individual. Through usage, linguistic structures that are more easily chunked will tend to proliferate, thus shaping the cultural evolution of language across generations of language users. This results in a selection of learnable structures, from individual words to multiword sequences that are optimally “chunkable,” so as to better squeeze through the Now‐or‐Never bottleneck. From this perspective, language can be thought of as an adaptive system that culturally evolves to fit learners’ cognitive capabilities, thereby resulting in the structure it bears today.     

Children and Adults as Language Learners: Rules,Variation, and Maturational Change

Here we overview our recent research investigating children and adults’ learning of rules and variation. In all these studies, our findings are that children and adults differ in how they acquire linguistic patterns that are productive, variable, inconsistently used, or lexically restricted. Some of our studies examine children's learning of natural languages; other studies expose learners to miniature languages and then ask them to produce novel sentences or judge their grammaticality. In every case there are important differences between learners as a function of their ages. Young children learn categorical rules and categorically follow patterns that are widespread in natural languages, even when their linguistic input exemplifies these patterns only probabilistically. In contrast, adult learners reproduce the probabilistic patterns of the input. Older children are in between, producing regular patterns somewhat more often than they appear in the input but also acquiring some probabilistic variation. These results suggest that the outcome of learning is quite different at different ages and that many of the properties of natural languages may be shaped by the behavior of children as they learn their native languages.     

Children’s Grammars Grow More Abstract with Age—Evidence from an Automatic Procedure for Identifying the Productive Units of Language

We develop an approach to automatically identify the most probable multiword constructions used in children’s utterances, given syntactically annotated utterances from the Brown corpus of CHILDES. The found constructions cover many interesting linguistic phenomena from the language acquisition literature and show a progression from very concrete toward abstract constructions. We show quantitatively that for all children of the Brown corpus grammatical abstraction, defined as the relative number of variable slots in the productive units of their grammar, increases globally with age.     

Cognitive Biases,Linguistic Universals,and Constraint‐Based Grammar Learning

According to classical arguments, language learning is both facilitated and constrained by cognitive biases. These biases are reflected in linguistic typology—the distribution of linguistic patterns across the world's languages—and can be probed with artificial grammar experiments on child and adult learners. Beginning with a widely successful approach to typology (Optimality Theory), and adapting techniques from computational approaches to statistical learning, we develop a Bayesian model of cognitive biases and show that it accounts for the detailed pattern of results of artificial grammar experiments on noun‐phrase word order (Culbertson, Smolensky, & Legendre, 2012). Our proposal has several novel properties that distinguish it from prior work in the domains of linguistic theory, computational cognitive science, and machine learning. This study illustrates how ideas from these domains can be synthesized into a model of language learning in which biases range in strength from hard (absolute) to soft (statistical), and in which language‐specific and domain‐general biases combine to account for data from the macro‐level scale of typological distribution to the micro‐level scale of learning by individuals.     

Explicit and implicit second language training differentially affect the achievement of native-like brain activation patterns

It is widely believed that adults cannot learn a foreign language in the same way that children learn a first language. However, recent evidence suggests that adult learners of a foreign language can come to rely on native-like language brain mechanisms. Here, we show that the type of language training crucially impacts this outcome. We used an artificial language paradigm to examine longitudinally whether explicit training (that approximates traditional grammar-focused classroom settings) and implicit training (that approximates immersion settings) differentially affect neural (electrophysiological) and behavioral (performance) measures of syntactic processing. Results showed that performance of explicitly and implicitly trained groups did not differ at either low or high proficiency. In contrast, electrophysiological (ERP) measures revealed striking differences between the groups' neural activity at both proficiency levels in response to syntactic violations. Implicit training yielded an N400 at low proficiency, whereas at high proficiency, it elicited a pattern typical of native speakers: an anterior negativity followed by a P600 accompanied by a late anterior negativity. Explicit training, by contrast, yielded no significant effects at low proficiency and only an anterior positivity followed by a P600 at high proficiency. Although the P600 is reminiscent of native-like processing, this response pattern as a whole is not. Thus, only implicit training led to an electrophysiological signature typical of native speakers. Overall, the results suggest that adult foreign language learners can come to rely on native-like language brain mechanisms, but that the conditions under which the language is learned may be crucial in attaining this goal.     

Large-scale cortical network properties predict future sound-to-word learning success

The human brain possesses a remarkable capacity to interpret and recall novel sounds as spoken language. These linguistic abilities arise from complex processing spanning a widely distributed cortical network and are characterized by marked individual variation. Recently, graph theoretical analysis has facilitated the exploration of how such aspects of large-scale brain functional organization may underlie cognitive performance. Brain functional networks are known to possess small-world topologies characterized by efficient global and local information transfer, but whether these properties relate to language learning abilities remains unknown. Here we applied graph theory to construct large-scale cortical functional networks from cerebral hemodynamic (fMRI) responses acquired during an auditory pitch discrimination task and found that such network properties were associated with participants' future success in learning words of an artificial spoken language. Successful learners possessed networks with reduced local efficiency but increased global efficiency relative to less successful learners and had a more cost-efficient network organization. Regionally, successful and less successful learners exhibited differences in these network properties spanning bilateral prefrontal, parietal, and right temporal cortex, overlapping a core network of auditory language areas. These results suggest that efficient cortical network organization is associated with sound-to-word learning abilities among healthy, younger adults.     

A Comparative Perspective on the Role of Acoustic Cues in Detecting Language Structure

Most human language learners acquire language primarily via the auditory modality. This is one reason why auditory artificial grammars play a prominent role in the investigation of the development and evolutionary roots of human syntax. The present position paper brings together findings from human and non‐human research on the impact of auditory cues on learning about linguistic structures with a special focus on how different types of cues and biases in auditory cognition may contribute to success and failure in artificial grammar learning (AGL). The basis of our argument is the link between auditory cues and syntactic structure across languages and development. Cross‐species comparison suggests that many aspects of auditory cognition that are relevant for language are not human specific and are present even in rather distantly related species. Furthermore, auditory cues and biases impact on learning, which we will discuss in the example of auditory perception and AGL studies. This observation, together with the significant role of auditory cues in language processing, supports the idea that auditory cues served as a bootstrap to syntax during language evolution. Yet this also means that potentially human‐specific syntactic abilities are not due to basic auditory differences between humans and non‐human animals but are based upon more advanced cognitive processes.     

Relationships amongst age,language and related skills in adults with Down syndrome

Studies into the effects of ageing on language in adults with Down syndrome (DS) have tended to rely on measures that lack sensitivity to change because they fail to explore across linguistic domains or rely on proxy reports. The study aim was to use measures of receptive and expressive language from studies of younger individuals with DS in exploring relationships across linguistic and associated skills, and age in young to older adults. Fifty-five adults (aged 19–58 years), 10 with a diagnosis of or signs of early stage Alzheimer's Disease (AD), provided data on measures of functioning associated with AD, non-verbal cognition, receptive language (which provided a measure of mental age), receptive and expressive language, and short term auditory and visual memory. The first order correlation between the measure of AD and CA was significant; but not when the 10 participants with AD were removed from the analysis. Significant negative correlations were obtained between CA and all other measures; small to large significant positive correlations were found amongst the other measures. Partial correlations were conducted to remove the potential effects of AD and IQ (the latter measured by a test of non-verbal cognition). Remaining significant correlations were between auditory short term memory and all other included measures, expressive language and all other included measures, and CA and auditory short term memory and expressive language. The results indicate that deterioration with age in this cross-sectional study was accounted for largely by the presence of AD. The exceptions were for auditory short term memory and expressive language. The findings may reflect an underlying deficit in auditory short term memory for adults with DS, as has been found in previous research of children and adolescents. The implications are discussed in terms of the importance of including comprehensive measures of receptive and expressive language and the need to account for the presence of AD in studies of cognitive decline associated with ageing in DS.     

Is the Linguistic Content of Speech Less Salient than its Perceptual Features in Autism?

Open-ended tasks are rarely used to investigate cognition in autism. No known studies have directly examined whether increased attention to the perceptual level of speech in autism might contribute to a reduced tendency to process language meaningfully. The present study investigated linguistic versus perceptual speech processing preferences. Children with autism and controls were tested on a quasi-open-format paradigm, in which speech stimuli contained competing linguistic and perceptual information, and could be processed at either level. Relative to controls, children with autism exhibited superior perceptual processing of speech. However, whilst their tendency to preferentially process linguistic rather than perceptual information was weaker than that of controls, it was nevertheless their primary processing mode. Implications for language acquisition in autism are discussed.     

The emergence of multiword utterances in children with Down syndrome

Morphosyntax is one of the most impaired aspects of language development in children with Down syndrome. The present study aimed to assess the emergence of multiword utterances in this population. Sixteen Italian-speaking children with Down syndrome were followed from 36 to 48 months of age. Data derived from an analysis of their spontaneous productions showed that although the mean productivity of multiword utterances increased over the three time points (36, 42 and 48 months), different growth patterns of early syntactic development could be identified: (1) null or marginal development; (2) a gradual increase in multiword production over time; (3) an increase in the production of more complex multiword utterances and a decrease or inverted U-shaped profile in the production of simpler multiword productions; (4) an inverted U-shaped profile in multiword productions. In addition, children showed an improvement in their ability to express different semantic functions by word combinations. Significant relationships were found between early syntactic skills and both the child’s vocabulary size and developmental age.

A better knowledge of the acquisition pace and content of word combinations could allow more effective rehabilitative treatment during the first steps of children’s syntactic development.     

Morphological processing in a second language: behavioral and event-related brain potential evidence for storage and decomposition

This study reports the results of two behavioral and two event-related brain potential experiments examining the processing of inflected words in second-language (L2) learners with Russian as their native language. Two different subsystems of German inflection were studied, participial inflection and noun plurals. For participial forms, L2 learners were found to widely generalize the -t suffixation rule in a nonce-word elicitation task, and in the event-related brain potential experiment, they showed an anterior negativity followed by a P600-both results resembling previous findings from native speakers of German on the same materials. For plural formation, the L2 learners displayed different preference patterns for regular and irregular forms in an off-line plural judgment task. Regular and irregular plural forms also differed clearly with regard to their brain responses. Whereas overapplications of the -s plural rule produced a P600 component, overapplications of irregular patterns elicited an N400. In contrast to native speakers of German, however, the L2 learners did not show an anterior negativity for -s plural overapplications. Taken together, the results show clear dissociations between regular and irregular inflection for both morphological subsystems. We argue that the two processing routes posited by dual-mechanism models of inflection (lexical storage and morphological decomposition) are also employed by L2 learners.     

No statistical learning advantage in children over adults: Evidence from behaviour and neural entrainment

Explicit recognition measures of statistical learning (SL) suggest that children and adults have similar linguistic SL abilities. However, explicit tasks recruit additional cognitive processes that are not directly relevant for SL and may thus underestimate children’s true SL capacities. In contrast, implicit tasks and neural measures of SL should be less influenced by explicit, higher-level cognitive abilities and thus may be better suited to capturing developmental differences in SL. Here, we assessed SL to six minutes of an artificial language in English-speaking children (n = 56, 24 females, M = 9.98 years) and adults (n = 44; 31 females, M = 22.97 years), using explicit and implicit behavioural measures and an EEG measure of neural entrainment. With few exceptions, children and adults showed largely similar performance on the behavioural explicit and implicit tasks, replicating prior work. Children and adults also demonstrated robust neural entrainment to both words and syllables, with a similar time course of word-level entrainment, reflecting learning of the hidden word structure. These results demonstrate that children and adults have similar linguistic SL abilities, even when learning is assessed through implicit performance-based and neural measures.     

Exceptions and anomalies: An ERP study on context sensitivity in autism

Several studies have demonstrated that people with ASD and intact language skills still have problems processing linguistic information in context. Given this evidence for reduced sensitivity to linguistic context, the question arises how contextual information is actually processed by people with ASD. In this study, we used event-related brain potentials (ERPs) to examine context sensitivity in high-functioning adults with autistic disorder (HFA) and Asperger syndrome at two levels: at the level of sentence processing and at the level of solving reasoning problems. We found that sentence context as well as reasoning context had an immediate ERP effect in adults with Asperger syndrome, as in matched controls. Both groups showed a typical N400 effect and a late positive component for the sentence conditions, and a sustained negativity for the reasoning conditions. In contrast, the HFA group demonstrated neither an N400 effect nor a sustained negativity. However, the HFA group showed a late positive component which was larger for semantically anomalous sentences than congruent sentences. Because sentence context had a modulating effect in a later phase, semantic integration is perhaps less automatic in HFA, and presumably more elaborate processes are needed to arrive at a sentence interpretation.     

Children with specific language impairment also show impairment of music-syntactic processing

Both language and music consist of sequences that are structured according to syntactic regularities. We used two specific event-related brain potential (ERP) components to investigate music-syntactic processing in children: the ERAN (early right anterior negativity) and the N5. The neural resources underlying these processes have been posited to overlap with those involved in the processing of linguistic syntax. Thus, we expected children with specific language impairment (SLI, which is characterized by deficient processing of linguistic syntax) to demonstrate difficulties with music-syntactic processing. Such difficulties were indeed observed in the neural correlates of music-syntactic processing: neither an ERAN nor an N5 was elicited in children with SLI, whereas both components were evoked in age-matched control children with typical language development. Moreover, the amplitudes of ERAN and N5 were correlated with subtests of a language development test. These data provide evidence for a strong interrelation between the language and the music processing system, thereby setting the ground for possible effects of musical training in SLI therapy.     

Thinking About Multiword Constructions: Usage-Based Approaches to Acquisition and Processing

Usage-based approaches to language hold that we learn multiword expressions as patterns of language from language usage, and that knowledge of these patterns underlies fluent language processing. This paper explores these claims by focusing upon verb–argument constructions (VACs) such as “V(erb) about n(oun phrase).” These are productive constructions that bind syntax, lexis, and semantics. It presents (a) analyses of usage patterns of English VACs in terms of their grammatical form, semantics, lexical constituency, and distribution patterns in large corpora; (b) patterns of VAC usage in child-directed speech and child language acquisition; and (c) investigations of VAC free-association and psycholinguistic studies of online processing. We conclude that VACs are highly patterned in usage, that this patterning drives language acquisition, and that language processing is sensitive to the forms of the syntagmatic construction and their distributional statistics, the contingency of their association with meaning, and spreading activation and prototypicality effects in semantic reference. Language users have rich implicit knowledge of the statistics of multiword sequences.     

Training with cognitive sequences improves syntactic comprehension in agrammatic aphasics

A major open question in cognitive neuroscience concerns the modularity of language: does human language rely, in part, on neural processes that are not language specific? Such reliance would predict that learning should transfer between non-linguistic and linguistic domains via this common neural basis. To test this prediction, we studied effects of non-linguistic cognitive sequence training on syntactic comprehension of six left-hemisphere damaged aphasic patients. Syntactic comprehension impairment was quantified before and after 10 weeks of training on a non-linguistic sequence processing task. This task used a transformational rule specifically corresponding to the transformation required for understanding a particular type of sentence referred to as relativised. Non-linguistic sequencing improved significantly with training (day effect: F(9,45)=3.7, <0.005). Moreover, a significant transfer of this improvement was observed for relativised, but not active nor passive sentences (pre-post x type interaction: F(2,10)=4.72, <0.05). The specificity of this transfer indicates that language relies partially on functional and neural processes that are not language specific.