Questions: Universal Grammar and the Innateness Hypothesis

Question 1

Children who have never heard sentences like 'Is the man who is tall happy?' spontaneously produce the correct form rather than '*Is the man who tall is happy?' — without instruction and without making errors. What does this pattern most directly support?

• A. Children generalize this rule by pattern-matching from simpler yes/no question structures they hear frequently
• B. Parents reliably correct children's errors with auxiliary fronting in relative clauses, providing the needed learning signal
• C. The grammatical constraint on structure-dependent movement is innate, not learned from the available input data
• D. English is simpler than other languages, so children in English acquire question formation faster

Answer: (C)

This is the classic Poverty of Stimulus argument. Children apply the correct structure-dependent rule without: (a) hearing such complex sentences in the input, (b) being explicitly taught the rule, and (c) making errors that would trigger parental correction. Option 0 is the usage-based counter-argument, but inferring abstract structure-dependence from surface patterns requires such strong statistical learning biases that they themselves look innate. Option 1 is empirically false — parents rarely correct syntactic errors. The POS conclusion: the grammatical principle must be part of the child's innate initial endowment.

Question 2

The head-directionality parameter in UG determines whether a language is head-initial (verb before object) or head-final (verb after object). When a child identifies this parameter setting from the input, the 'parameter cascade' predicts which outcome?

• A. The child learns only the specific word-order patterns that directly appear in the input
• B. The child must re-learn dozens of grammatical rules independently, each requiring its own evidence
• C. Many related grammatical properties fall into place simultaneously because they are linked to the same parameter
• D. The child initially sets the parameter incorrectly and gradually revises it based on explicit feedback

Answer: (C)

The parameter cascade is a key prediction of UG theory: a single parameter controls a cluster of related grammatical properties. Setting head-directionality also sets expectations about the positions of auxiliaries, adpositions, complementizers, and other heads — not just main verbs. When the child identifies the setting from limited input, a large number of related properties are simultaneously determined. This explains why language acquisition is so fast: the child is adjusting a small number of pre-specified switches, not inductively learning thousands of individual rules.

Question 3

Universal Grammar claims that most human languages have essentially identical surface grammar, which is why children can acquire any language with equal ease.

• T. True
• F. False

Answer: False

UG claims that all languages share deep structural PRINCIPLES (universal constraints on possible grammatical operations) while varying along PARAMETERS (e.g., head direction, pro-drop, null subject). Surface grammars differ enormously — word order, morphology, phonology, and agreement systems vary dramatically across languages. The universality is in the abstract constraining principles and the parameterized space of possibilities, not in surface form. UG explains both the cross-linguistic universals (shared principles) AND the diversity (parameters). Children acquire whatever language they are exposed to by setting parameters, not by recognizing familiar surface patterns.

Question 4

Children who acquire creole languages from structurally impoverished pidgin input tend to produce grammars more elaborate than the pidgin, suggesting they supplied grammatical structure from innate resources when the input underspecified it.

• T. True
• F. False

Answer: True

Creolization studies are cited as supporting evidence for UG. Pidgins are contact languages with minimal grammar. When children acquire them as native languages, they produce creoles — richer languages with systematic morphology and syntax absent from the input pidgin. The children appear to be 'filling in' grammatical structure that the input did not provide, consistent with innate grammatical resources. This is also observed in homesign: deaf children not exposed to a sign language spontaneously develop gestural systems with systematic grammatical properties, drawing on innate structure rather than input.

Question 5

What is the Poverty of Stimulus argument, and why does it support the claim that grammatical knowledge is innate?

Model answer: The Poverty of Stimulus argument observes that children acquire grammatical knowledge that exceeds what their input data could logically support. Children apply complex, abstract rules — such as structure-dependent auxiliary fronting in questions — correctly and spontaneously, despite never hearing such sentences in input, never being taught the rule, and never receiving corrections. The input is too sparse and too surface-level to reliably support induction of the abstract principle involved. The most parsimonious explanation is that children arrive with the grammatical constraint already in place — it is innate, part of Universal Grammar. The input triggers and refines UG-consistent options rather than teaching grammar from scratch.

The POS argument is essentially an argument from underdetermination: the relationship between input data and acquired grammar is too loose for pure learning without strong innate constraints. It echoes Plato's Meno — how does the learner know more than they were taught? In linguistics, the answer UG proposes is that they were born knowing the deep principles. The debate with usage-based theories turns on whether the statistical regularities in the input, combined with domain-general learning mechanisms, are sufficient to explain acquisition — or whether those mechanisms themselves require innate grammatical biases to work.