Psycholinguistics: The Language Bioprogram Hypothesis | ENG511 Lecture 20

In the late 1970s, deaf children at a school in Managua, Nicaragua, spontaneously created Nicaraguan Sign Language despite receiving no formal sign language instruction. Older students initially used different homesign systems, but converged on a shared communication system resembling a pidgin. Remarkably, younger children then expanded this system, adding grammatical rules and spatial markers for agreement, creating a fully functional language. The youngest children used the most complex structures, demonstrating that they invented the language rather than learning it from adults or older peers. This case provides direct evidence that children can create complete languages when given the opportunity.

The most compelling single example of spontaneous language creation in the 20th century is Nicaraguan Sign Language, which emerged in the late 1970s and early 1980s when deaf children across Nicaragua were brought together in schools for the first time. The adults in these schools taught signing based on Spanish words. The children interacting with each other created something substantially different - a gestural language with consistent grammatical structures, including spatial grammar and verb agreement systems that the adults signing lacked entirely. Crucially, each successive cohort of children who entered the schools learned from the previous cohort and produced a version of the language with richer and more consistent grammatical structures than their teachers had used. The language was being genuinely created generation by generation in real time, driven specifically by children, not adults. This suggests that the grammatical structuring capacity that Chomsky attributed to an innate language acquisition device is specifically active in childhood in a way that it is not in adults.

Nicaraguan Sign Language provides a unique case study of language emergence. After the revolution, deaf children from isolated backgrounds attended schools together for the first time. Without existing communication systems, they began gesturing with each other, repeating and modifying each other's gestures. This repetitive, periodic interaction created the conditions for language emergence. Younger children developed more fluent signing that fed back into the community, creating successive generations of acquirers. The language that emerged drew on gesture as raw materials but transformed these elements into a fully grammatical system with hierarchical structure and syntactic rules.

Nicaraguan Sign Language (NSL), which emerged in Nicaragua's deaf community since the 1970s, provides unprecedented evidence that all languages originate from child learners who possess a special capacity to create, organize, and change language over time; unlike deaf children in earlier generations who were language-less and had to develop communication systems independently, modern deaf children can immediately learn a fully developed language, allowing researchers to document language evolution across generations.

In the 1980s, deaf children in Nicaragua were brought together for the first time in schools and created Nicaraguan Sign Language (NSL) spontaneously. Starting with simple gestures, younger children rapidly developed it into a fully complex language with its own grammar, without any existing model to copy. This demonstrates that grammar can be generated rather than learned through imitation, and that there is a critical period for language acquisition during childhood.

Modules are systems of data input that process information with specific characteristics: automatic processing, obligatory processing, informational encapsulation, fixed neural architecture, and domain specificity. Domains are sets of representations about an area of knowledge. Modules can be grouped into domains. The distinction is important because modules represent the computational units of cognitive processing, while domains represent the organized knowledge structures that modules process.

Karmiloff-Smith distinguishes between domain-specificity and modular encapsulation. Domain-specificity refers to cognitive processes that handle particular types of information (like language or numbers), while modular encapsulation means these processes are isolated from general cognitive systems. She argues that some domain-specific processes may not be fully encapsulated, allowing for more flexible cognitive development and integration across domains.

The new DoGMA (Domain-General Modular Architecture) of empiricism proposes that cognition involves domain-general faculties (perception, memory, imagination, attention, empathy) governed by domain-general inductive biases implemented through architectural constraints, rather than requiring innate domain-specific representations; this framework channels historical empiricist traditions (Locke, Hume, Ibn Sina, Adam Smith) into modern deep learning architectures, demonstrating how neural networks can implement classical empiricist abstractions like composition, subtraction, and representation through hierarchical processing that controls for nuisance variation in sensory input.

There is considerable consensus about the species specificity and domain specificity of the language faculty. Verbal behavior constitutes a separate domain different from other events and noises. Dissociation of language and other cognitive abilities provides evidence—cases where language abilities are intact and other cognitive abilities are not, and vice versa. Developmental dissociations show that a human child and kitten develop differently despite similar experiences, with the child's language faculty functioning while the kitten's does not.

Cognitive modularity theory, established by Fodor (1983) and rooted in 19th-century phrenology, proposes that the mind/brain consists of distinct functional units each competent for specific tasks. Two core principles define this framework: domain specificity and encapsulation. Domain specificity holds that modules work on proprietary vocabulary—syntax handles number/person/animacy while phonology handles larity/starkness/voicing, with these vocabularies in complementary distribution. Encapsulation means modules access only limited information during computation; optical illusions demonstrate this by showing visual systems ignore conscious knowledge. The MECR effect illustrates that phonology receives input from multiple sources (vision and audition), yet maintains computational independence. Intermodular communication requires translation mechanisms because modules cannot interpret each other's vocabulary. The morphosyntax-phonology interface employs two distinct translation mechanisms: morphemic information (roots, affixes) undergoes lexical translation through dictionary-like matching, while non-morphemic information (syntactic boundaries) undergoes computational translation through algorithms. However, computational translation is incompatible with modularity because it would require modules to process two vocabularies simultaneously, violating domain specificity. Lexical translation is supported by evidence throughout cognition: phonetics-phonology mapping, categorical perception across senses, and all modular communication involving finite lists.

Creole languages emerge when speakers of different languages create a common pidgin, which then becomes the first language of children. Gullah (South Carolina/Georgia) and Krio (Sierra Leone) show remarkable structural similarities despite being separated by the Atlantic Ocean, demonstrating how English-based pidgins spread globally through maritime trade and slavery. All creoles worldwide share surprising grammatical, phonological, and lexical similarities because they originated from the same source: English mariners and traders. This phenomenon reveals how human migration and commerce create unexpected linguistic connections across vast distances, with creoles developing remarkably similar structures regardless of their geographic distribution.

Despite diverse substrates, Creoles show remarkable structural similarities. They are typically very low on inflectional morphology, often lacking contextual inflection entirely, with tense and aspect indicated by particles. Certain structural features recur across Creoles regardless of specific substrates, suggesting common mechanisms when language starts again. Substratum influence is important in almost every Creole, as creators transfer structures from native languages. However, the consistent simplification patterns across geographically and linguistically diverse Creoles indicate that substratum alone cannot explain their structural uniformity. This suggests universal principles at work in Creole genesis, possibly related to the challenges adults face in acquiring complex linguistic features.

A major debate exists about whether creoles form a distinct language type. McWhorter's prototype theory identifies creoles by three structural features: no inflection (expressing grammar through separate words), no tone, and no lexicalization (words don't have parts that no longer mean what they used to). However, this approach is controversial because non-creole languages like Mandarin Chinese and Vietnamese also lack inflection. Critics argue defining creoles by absence of features echoes problematic colonial-era discourse that framed them as 'primitive.' The debate centers on whether structural criteria or sociohistorical conditions (plantation colonies with enslaved people) should define creoles.

Creoles are mixed languages emerging from colonial encounters and forced population displacement, concentrated in the Caribbean, West Africa, and Southeast Asia. Pidgins are unstable contact languages used by adults for practical communication, characterized by simplified grammar, phonology, and morphology. Creoles emerge when children born into pidgin environments automatically impose grammatical coherence, creating fully developed languages. This process provides evidence for Universal Grammar—the hypothesis that humans possess innate cognitive structures for language acquisition. English-based creoles dominate the Caribbean and Africa (Jamaican Creole 3.2M, Haitian Creole 10M), French-based creoles spread across the Francophone world, Portuguese-based creoles include Cape Verdean Creole (oldest documented creole, 16th century), German-based creoles like Unserdeutsch in Papua New Guinea are nearly extinct, and Arabic-based creoles like Arabic Juba serve as national languages in Sudan. Haitian Creole demonstrates unique features including article-noun fusion, pronoun system with identical subject-object forms, and aspect-based verb system. Papiamento exemplifies complex multilingual creole formation with Portuguese base vocabulary but substantial Spanish, Dutch, and indigenous Caribbean influences. Cape Verdean Creole preserves Portuguese vocabulary while developing distinctive grammatical patterns including progressive aspect with 'tá' and permanent-temporary possession distinction ('ter' vs. 'tê').