Language Acquisition: Innateness, Competence vs Performance, and Theories
Overview
Focus of the lecture: language is uniquely human, and how children acquire language from their environment. Two key questions guiding the discussion:
How is the knowledge of language developed in kids?
What is the role of language input in language acquisition?
By age years, children are said to know much of the syntax of their language, and by around years they can communicate intents, make requests, and express themselves with other speakers. However, this rapid grammar development is debated because kids typically have limited vocabulary and pragmatic knowledge by age and may struggle with pronunciation or producing all words perfectly.
The main problem: how do children develop a complex grammatical system within the first years, given the rapid changes in brain development (brain plasticity) during early life?
Three broad scholarly perspectives on language learning discussed:
Innate (nativist) approach: language is largely innate, with universal grammar guiding acquisition.
Empiricist / behaviorist approach: language learning is driven by experience and environmental input.
Interactionist approach: language learning is a product of social interaction and communicative goals.
The debate is framed as nature vs nurture, with evidence cited from animal learning, human neural development, and linguistic data.
Distinction between two concepts of language:
Competence: linguistic competence (mental grammar, knowledge of grammar rules).
Performance: linguistic performance (actual use of language in context).
Second language learning is used to illustrate performance differences even when grammar knowledge (competence) exists in textbooks or instruction.
Three major theoretical frameworks historically discussed:
Behaviorism (Skinner): language as learned through reinforcement; emphasis on external input and conditioning.
Universal Grammar / Nativist (Chomsky): language has an innate, domain-specific faculty with universal principles and parameters.
Interactionist / Cognitive-Functional (e.g., Piaget-inspired or social interaction emphasis): language learning is driven by social interaction and communicative needs.
The speaker argues that language learning is multidimensional and involves multiple interacting systems; all three approaches offer compelling evidence.
Key Questions and Concepts in Language Acquisition
Two broad questions about language knowledge:
How is linguistic knowledge developed in children?
What is the role of linguistic input in the development of language?
By early childhood, children show strong syntactic knowledge, but the extent and pace of grammar learning remain debated.
Core concept: internal linguistic knowledge vs observable usage (grammar vs performance).
The nature-nurture debate persists, with some arguing for innate linguistic structure, others for experience-based learning, and many arguing for a hybrid view.
The uniqueness of human language is contrasted with nonhuman primate learning (limited language-like capabilities in animals).
Brain Development and Language Learning
Language development occurs within a brain undergoing extensive changes during childhood (brain plasticity).
The rapid acquisition of native language grammar in the presence of changing neural substrates is a central challenge for theories of language learning.
The human brain supports sophisticated parsing, word meaning activation, and cross-language generalizations, pointing to powerful underlying cognitive mechanisms.
Developmental Milestones and Limitations
Typical milestones discussed:
By years: early syntax is acquired; rapid grammar development is observed in many children.
By years: more mature pragmatic and conversational abilities surface, though vocabulary breadth and pronunciation coverage remain variable.
Limitations noted:
Limited vocabulary by age
Imperfect pronunciation and pragmatic nuance
Variability across languages and individual learners
Debates about whether observed milestones imply complete mastery of target grammar or ongoing refinement and generalization.
The Nature of Language Structure: Competence vs Performance
Distinction:
Linguistic Competence: mental grammar, the internalized rules and representations a speaker has about their language.
Linguistic Performance: how language is actually used in real situations, including mistakes, hesitations, and planning.
Evidence from second language learners shows robust grammar knowledge (competence) may not always translate into native-like performance in real contexts.
In research, to study competence, experiments are designed to minimize performance factors (e.g., asking children to judge or produce sentences in controlled conditions).
The generative approach to language acquisition is primarily focused on competence rather than performance.
Innatist (Nativist) Framework: Universal Grammar, Principles & Parameters
Core claim: language is domain-specific and endowed with an innate set of knowledge (UG) that constrains learning.
Universal Grammar (UG): rules that are universal across all human languages, forming the core structure of language in the mind.
Language acquisition as a process starting from an input state and moving toward a mature grammar via a Language Acquisition Device (LAD).
Schematic idea: input leads to activation of LAD, which under practice and exposure yields the target mental grammar.
Principles: the core constraints common to all languages; they constrain the space of possible grammars (hypotheses) children consider.
Parameters: language-specific settings that override universal constraints, acting like switches that can be set in different ways depending on the language environment.
Concept: once principles constrain hypothesis space, parameters specify how a language can vary (e.g., word order, pro-drop vs explicit subject, wh- movement, etc.).
Formalization idea (LaTeX-style): Let \Theta = (p1, p2, …, pn), where each pi ∈ {0,1}, representing binary parameter settings for a given language. Different languages instantiate different parameter values.
Illustrative concepts:
Principles vs parameters help explain cross-language variation while preserving a shared innate endowment.
Parameters reduce the learning burden by narrowing the space of possible grammars the child must consider.
Classic elements in the UG framework discussed:
Language acquisition is aided by immutable constraints (principles) that apply to all languages.
Children are like scientists who form hypotheses about language within the restrictions set by principles and parameters.
Parameter setting allows the system to adapt to specific language environments.
The Grammar-building process in the UG view:
Input triggers activation of the innate language faculty.
Practice with language exposure tunes the grammar toward the target language, guided by principles and parameter settings.
Examples and clarifications:
Pro-drop vs non-pro-drop languages (e.g., Mandarin Chinese is pro-drop; English generally requires explicit subjects).
OSV/SVO word order variations and how parameters would describe them.
The wh-question formation parameter (placement of question words) as a switch affecting sentence structure.
Principle C (a specific UG constraint) example:
A pronoun cannot refer to the same entity as a coindexed name when the pronoun is in certain positions in the tree; i.e., coreference constraints depend on syntactic configuration.
Configurations a and b illustrate how coreference is permitted in one structure but not the other due to tree configuration.
This principle is considered universal and operative early in child language, not just a product of surface input.
The role of evidence: experimental linguistics tests UG principles by observing whether children show knowledge consistent with the principles across languages.
Critical concepts: recursion, long-distance dependencies, and the hierarchical nature of syntax are central to the UG view because they are hard to derive from simple linear input alone.
Empiricist / Behaviorist Frameworks
Behaviorism (Skinner): language learned through operant conditioning and reinforcement; verbal behavior is strengthened or weakened by consequences.
Example: correct sentences reinforced by social feedback (e.g., praise) and incorrect forms diminished by lack of reinforcement.
The idea of language being learned through environmental input and feedback, not through an innate grammar.
Classical human learning analogies used (e.g., discrimination and generalization) to explain how children might generalize from specific input to broader rules.
Problems with pure behaviorism for language:
Recursion and the generation of novel sentences (never heard before) are difficult to account for with simple reinforcement models.
The speed and universality of early grammar acquisition across diverse languages are hard to reconcile with reinforcement alone.
The argument that learning language likely involves domain-general cognitive processes in addition to domain-specific mechanisms.
Interactionist / Social-Contextual Perspectives
Interactionist claim: language development is driven by social interactions and communicative goals.
Key role of joint attention and communicative intent: children learn language effectively when caregivers’ input is tied to shared attention and purposeful communication.
Input effects: parent-child interactions (recasts, expansions, and adjusted feedback) influence language development, but the extent to which grammar is learned from input alone remains debated.
Special input (child-directed speech, motherese): while it may help phonology and semantics, its direct effect on grammar learning is contested.
The non-domain-specific view that social context supplies structure and meaning that guide language use and possibly some aspects of grammar development, but not entire grammar by itself.
Four Types of Linguistic Knowledge (illustrative examples about principle-based learning)
Knowledge about the parts into which a sentence can be divided (syntax categories like NP, VP, etc.).
Knowledge about which sentences are ambiguous (e.g., ambiguous scope, pronoun reference, etc.).
Knowledge about which sentences cannot be expressed in a given way (grammaticality constraints).
Knowledge about what particular sentences cannot mean even if their form is acceptable (semantic/pragmatic constraints).
These aspects of knowledge are argued to be present early and influence language acquisition beyond what input alone can provide.
Principle C in Depth: Ambiguity, Coreference, and Tree Structures
Example sentence: "The Ninja Turtle said he likes pizza." This sentence is ambiguous regarding whether "he" refers to the Ninja Turtle or someone else.
Principle C is a cross-linguistic constraint limiting coreference in certain configurations, ensuring pronouns do not refer to the same entity as a closer antecedent in some structures.
Tree-structure illustrations (configurations a and b):
Configuration a forbids coreference between pronoun and name.
Configuration b allows coreference between pronoun and name.
The claim: these coreference constraints are universal and operative early in language development, not solely derived from surface input.
The claim that such principles constrain hypothesis space in a way that is not solely reducible to data-driven generalization from input.
The Cake Analogy: Principles, Parameters, and Settings
Analogy to cooking a cake: ingredients are universal (sugar, flour, etc.), but the specific recipe (order, substitutions) varies by language via parameter settings.
Parameters are like switches that toggle different language properties (e.g., word order, subject presence, question formation).
As children acquire language, they set these parameters to match the linguistic environment, guided by underlying principles.
This analogy emphasizes that language learning is facilitated by an underlying architecture, not a complete blank slate.
Recursion and Long-Distance Dependencies
Recursion allows the construction of potentially infinite sentences by embedding phrases within phrases (e.g., possessives: "the mother\'s hat"; prepositional phrases: "in the black cabinet, in the storage closet, …").
Recursion poses a challenge for simple linear input models and supports a hierarchical structure view of grammar.
Demonstrations show that human speakers can generate syntactically and semantically coherent long-distance dependencies that would be difficult to learn through simple stimulus-response reinforcement alone.
Illustrative sentences discussed (and why they test grammatical vs semantic plausibility) include long-distance subject-verb agreement and other dependencies, which highlight underlying tree-based grammar.
Corrective Feedback and Its Limitations for Grammar Learning
Two types of feedback discussed:
Positive reinforcement (e.g., social praise) reinforces correct utterances.
Punishment or correction for errors (negative feedback) is not consistently observed in natural language learning, especially for grammar.
The claim is that parents rarely correct children's grammar directly; grammar learning may rely on implicit cues, structured input, and internalized constraints rather than explicit grammatical corrections.
Input and Special Input: The Environment’s Role
Interactionist claims emphasize that the environment and social context shape language acquisition through exposure and communicative goals.
Some scholars acknowledge “special input” (e.g., parental prosody, heightened attention cues) but deny that this alone explains how grammar is learned.
Even with identical environments, differences in input across cultures and schools can lead to different language development trajectories.
The argument for a domain-general learning process alongside domain-specific language faculties is common in interactionist accounts.
Recap: The Three Broad Approaches and Their Intersections
All three approaches contribute meaningful insights:
Nativist / UG: language has an innate, domain-specific core with universal principles and language-specific parameters.
Empiricist / Behaviorist: language learning is shaped by environmental input and feedback mechanisms.
Interactionist: language learning is driven by social interaction, communicative goals, and joint attention, with input playing a supportive role.
The speaker emphasizes that language is complex and involves multiple cognitive and social systems; no single account fully explains all aspects of language acquisition.
Practical and Philosophical Implications
For education and speech pathology: understanding the balance between innate constraints and environmental input can inform approaches to language intervention and therapy.
Ethical and philosophical implications: recognizing that language development is influenced by biological endowment, cognitive processing, and social context supports a holistic view of human cognition and communication.
The ongoing debate remains: how much of language is hard-wired vs. learned from experience, and how do universal constraints shape the course of development across diverse languages?
Key Terms and Concepts (Glossary)
Competence vs Performance
Universal Grammar (UG)
Principles (UG constraints)
Parameters (language-specific switches)
Language Acquisition Device (LAD)
Pro-drop language
OSV/SVO word order variations
Principle C (coreference constraint)
Recursion and long-distance dependencies
Joint attention
Recasts and expansions (interactionist input strategies)
Imitative vs intentional or goal-oriented language learning
Domain-specific vs domain-general cognition
Phonology, Lexicon, Syntax, Semantics (levels of linguistic knowledge)
Sound-to-meaning mapping in language learning
Input vs output in L1/L2 acquisition
Prototypical vs non-prototypical grammatical constructions
Cross-linguistic variation and the continuity hypothesis
Equations and Notation (illustrative)
Parameter settings across languages can be formalized as a parameter vector:
Language development can be viewed as moving from a general input state toward a specific final grammar under constraints:
Input I activates a Language Acquisition Device (LAD) that, through principles and parameters, yields a mature mental grammar G for a given language.
Connections to Previous Content and Real-World Relevance
The discussion connects to prior lectures on hierarchical sentence processing and sentence parsing, showing how deep structure and syntactic trees influence interpretation.
Real-world relevance: understanding how children acquire grammar informs educational strategies, second-language teaching, and clinical approaches to language disorders.
The interplay of cognitive, social, and linguistic factors mirrors broader themes in cognitive science about how complex skills emerge from multiple interacting systems.
If you want, I can tailor these notes to focus more on a particular section (e.g., Universal Grammar details, or the interactionist literature) or add more explicit examples and diagrams to accompany each section.