A toddler who has never once heard the word "goed" will look up from a knocked-over cup and announce, "I goed potty." Nobody taught her that. In fact, every adult around her says "went," constantly — and she ignores all of it.
This is one of the strangest facts in psychology. If kids learned language purely by copying what they hear, "goed" should be impossible. You can't imitate a word nobody around you ever says. Yet children all over the world, in every language, invent errors exactly like this around age two or three.
What's actually happening is that your toddler isn't copying — she's running a hidden experiment. She has unconsciously cracked a rule ("add -ed to make past tense") and is now overapplying it to a word that breaks the rule. The mistake is a fingerprint of a mind doing statistics on grammar. By the end of this lesson, you'll understand why a wrong word like "goed" is one of the clearest windows we have into how YOUR brain built language out of noise.
Language is a system of symbols — spoken, written, or signed — combined by rules to communicate meaning. The exam loves to test the building blocks, in order from smallest to largest, so learn the ladder.
A phoneme is the smallest distinct unit of sound in a language. English has roughly 40 phonemes. The word cat has three: /k/, /a/, /t/. Phonemes carry no meaning by themselves — /k/ doesn't mean anything. Different languages slice sound differently: English distinguishes /r/ and /l/, but Japanese does not, which is why the contrast is famously hard for native Japanese speakers to hear.
A morpheme is the smallest unit of meaning in a language. Crucially, a morpheme is not the same as a syllable or a word. The word cats has two morphemes: cat (an animal) and -s (meaning "more than one"). Unbreakable has three: un- (not), break, -able (capable of). Some morphemes stand alone (cat), others must attach (-s, un-). This phoneme-vs-morpheme distinction is the single most tested vocabulary pair in this lesson: sound vs. meaning.
Grammar is the whole system of rules that lets us communicate. It has two big branches:
Try This. Take the sentence "Colorless green ideas sleep furiously" (linguist Noam Chomsky's famous example). It is perfectly syntactically correct — noun, verb, adverb all in the right slots — yet semantically nonsense. Now scramble it: "Furiously sleep ideas green colorless." Same words, but now the syntax is broken too. Feeling the difference between "grammatically fine but meaningless" and "meaningless and disordered" is feeling the difference between syntax and semantics.
Across every culture and language ever studied, children climb the same developmental staircase, on roughly the same timetable. The timing varies, but the sequence does not.
Then comes the tell-tale error. As children master grammatical rules, they begin overgeneralization (also called overregularization): applying a regular grammatical rule to words that are exceptions. "Goed" instead of "went," "foots" instead of "feet," "holded" instead of "held." Here's the counterintuitive part the exam loves: a child who says "goed" is often demonstrating more grammatical sophistication than one who says "went." The younger child may have memorized "went" as a chunk. The older child has extracted the -ed rule and is now overapplying it. Overregularization is evidence of rule-learning, not a sign of getting worse.
This is the central debate of the lesson, and a guaranteed FRQ-style contrast.
B.F. Skinner (1957) gave the behaviorist / operant account: children learn language the way they learn any behavior — through association, imitation, and reinforcement. Babies babble; adults reward sounds that resemble real words with smiles and attention; those sounds get strengthened; the rest extinguish. Language is shaped, response by response, by its consequences.
Noam Chomsky demolished this in a famous critique. His core objection is the poverty of the stimulus: children produce sentences they have never heard, and make systematic errors (like "goed") that nobody reinforced — so imitation and reinforcement can't be the whole story. Chomsky argued for nativism: humans are born with an innate, biologically wired capacity for language. He proposed a language acquisition device (LAD) — a hypothetical built-in mechanism that primes children to absorb the grammar of whatever language surrounds them. Underlying all human languages, he claimed, is a shared universal grammar: a deep set of grammatical principles common to every language, which is why children acquire any language so fast and in the same stages.
The modern consensus borrows from both: we're born with a powerful innate readiness (Chomsky was right that pure reinforcement fails), but environment, exposure, and yes, some reinforcement and statistical learning still shape which language and which vocabulary a child ends up with.
Chomsky's innate machinery seems to need switching on early. There appears to be a critical period (more cautiously, a sensitive period) — a window in early childhood during which the brain is primed to acquire language, after which full native mastery becomes difficult or impossible. Children who learn a language early reach fluency; people who first encounter language (or a second language) after puberty rarely achieve flawless grammar and accent. The most haunting evidence is the case of Genie (Spotlight below).
Learning two languages in childhood does not confuse or delay children, despite an old myth. Bilingual children hit milestones on schedule (counting words across both languages). Bilingualism is associated with a bilingual advantage in executive function — particularly the ability to suppress one language while using the other, which appears to strengthen attentional control. The sensitive-period logic applies here too: second languages learned early tend to be spoken more natively.
Benjamin Lee Whorf proposed linguistic relativity (the Whorf hypothesis / Sapir-Whorf hypothesis): the language you speak shapes the way you think. There are two versions, and the difference is everything on the exam:
Bottom line to memorize: the strong/deterministic version is wrong; the weak/influence version is defensible.
Animals communicate richly — bees dance, vervet monkeys give different alarm calls for different predators, apes like Kanzi and Koko learned hundreds of symbols. The debate is whether this is true language in the human sense. Skeptics argue animals lack generativity (the ability to combine a finite set of symbols by rules into infinite novel sentences) and rarely show real syntax. The safe AP position: animals clearly communicate and can learn symbols, but evidence that they command human-like grammar and generativity remains contested.
Genie — the "forbidden experiment" nobody would ever run on purpose (discovered 1970).
Who & when: Genie (a pseudonym) was discovered by Los Angeles authorities in 1970 at age 13. From roughly 20 months old, she had been strapped to a chair in a silent room by an abusive father, almost never spoken to, and punished for making noise. She had been deprived of language for nearly her entire childhood.
What researchers did: A team of psychologists and linguists, including Susan Curtiss, worked intensively to teach Genie language and documented her progress for years. She had, in effect, missed the entire proposed critical period.
What they found: Genie acquired a substantial vocabulary and could communicate her needs with strings of words. But she never mastered grammar — syntax in particular. Her speech stayed telegraphic and rule-poor ("Applesauce buy store"), and she could not consistently form normal sentences despite years of instruction.
Why it matters: Genie is the most cited real-world support for the critical-period hypothesis for language. The dissociation is the key: vocabulary (learnable late) came, but the deep grammatical machinery (apparently dependent on early exposure) did not. The case also raises wrenching ethics questions — researchers were later accused of prioritizing study over Genie's welfare — which makes it a favorite for FRQ ethics prompts. For the exam: Genie = late language deprivation = words but not grammar = evidence for a sensitive/critical period.
Scenario 1. A 20-month-old points at the door and says, "Daddy go." She uses only the essential content words and skips "the," "is," and "to."
Which stage and concept? This is telegraphic speech, characteristic of the two-word stage. The hallmark is content words only — nouns and verbs — with grammatical function words (articles, prepositions) dropped, like an old telegram. Notice the word order is already correct, showing emerging syntax.
Scenario 2. A 4-year-old who used to say "I held the cat" now says "I holded the cat." His worried parents think his language is regressing.
What's actually happening? The opposite of regression. This is overgeneralization / overregularization — the child has extracted the regular "-ed" past-tense rule and is overapplying it to an irregular verb. It is evidence of rule-learning, a sign of growing grammatical competence, and it resolves on its own. This pattern is also a classic argument against Skinner's pure-imitation account: the child says "holded," a form he never heard and nobody reinforced.
Scenario 3. A researcher finds that speakers of a language with separate basic words for "light blue" and "dark blue" are slightly faster at telling those two shades apart than speakers of a language with one word for both. They are equally able to see the difference; they're just quicker to categorize it.
Which version of which hypothesis does this support? This supports the weak version of linguistic relativity (the Whorf hypothesis) — language influences perception and categorization without determining it. It does not support linguistic determinism (the strong version), because both groups can still perceive the color difference; the speakers without two words are not blind to the distinction, just slightly slower to sort it.
Phoneme vs. morpheme. The number-one trap. A phoneme is the smallest unit of sound (/k/); a morpheme is the smallest unit of meaning (-s = plural). Mnemonic: phoneme = phone = sound; morpheme = "more meaning." Cats = three phonemes (/k/ /a/ /t/ /s/ — actually four sounds) but two morphemes (cat + s). They almost never match up, which is exactly why the exam pairs them.
Skinner vs. Chomsky. Skinner = nurture: language is learned through imitation and reinforcement (operant conditioning). Chomsky = nature/nativism: an innate language acquisition device and universal grammar. The single best piece of evidence against Skinner and for Chomsky is overregularization ("goed") — an error no one models or rewards. Mnemonic: Skinner = Skills are learned; Chomsky = Children are born ready.
Overregularization is NOT a deficit. Students hear "goed" and call it a language delay or error of imitation. On the exam it is the reverse: it signals the child has acquired a grammatical rule and is overapplying it. It's progress wearing the costume of a mistake.
Linguistic determinism (strong) vs. linguistic relativity (weak). Both are "the Whorf hypothesis," so students treat them as one claim. The strong version (determinism) — language controls/limits thought — is rejected. The weak version — language influences thought — is supported. If an answer choice says language "determines," "controls," or "makes it impossible to think," it's describing the discredited strong version; "influences," "nudges," or "affects" describes the defensible weak version.
Four-choice MCQs in current AP format. Answers and explanations in section (h).
Respond to all six parts (A–F) in complete sentences using appropriate psychological terminology. Part F is worth 2 points; all others are worth 1 point each (7 total).
Stimulus — summarized study
Introduction. Researchers investigated whether the number of basic color words a language provides influences how quickly speakers categorize colors. They hypothesized that having two distinct everyday words for two shades (rather than one shared word) would speed categorization of those shades, consistent with a weak version of linguistic relativity.
Participants. 120 adults volunteered and were paid a small fee: 60 native speakers of Language A (which has separate basic words for "light blue" and "dark blue") and 60 native speakers of Language B (which uses one basic word covering both). Groups were matched on age (M = 24.1, SD = 3.2) and on years of education. All participants gave written informed consent, were told they could withdraw at any time without losing payment, and were identified in the dataset only by a code number. Color vision was screened so that no participant had a color-vision deficiency.
Method. Each participant completed a computer task. On each trial, two color patches appeared side by side, and participants pressed one key if the patches were the same color and another key if they were different. On the critical trials, the two patches were a light-blue and a dark-blue shade. The computer recorded reaction time in milliseconds for correct "different" responses. Accuracy was near ceiling (above 97%) for both groups, so the analysis focused on speed.
Results. On the critical light-blue/dark-blue trials, Language A speakers (two color words) averaged 610 ms to correctly respond "different," while Language B speakers (one color word) averaged 685 ms — a difference of 75 ms. On control trials comparing two clearly different colors (e.g., red vs. green), the two groups did not differ (both ~600 ms).
A. Identify the research method used in this study.
B. State the operational definition of the dependent variable as used in this study.
C. Describe what the results indicate about the difference between Language A and Language B speakers on the critical trials. (Cite the numbers.)
D. Identify one ethical guideline the researchers applied, and describe how they applied it.
E. Explain the extent to which the findings are or are not generalizable, using specific evidence from the study.
F. Explain whether the findings support or refute the strong version of linguistic relativity (linguistic determinism). In your explanation, apply the distinction between the strong and weak versions of the Whorf hypothesis.
A. The study used a quasi-experiment (a non-experimental / correlational design). Participants were grouped by their native language, which is a pre-existing characteristic that cannot be randomly assigned, so the researchers compared naturally occurring groups rather than manipulating an independent variable. (1 pt — "quasi-experiment," "correlational," or "non-experimental comparison" all earn it; "true experiment" does not, because there is no random assignment.)
B. The dependent variable was operationally defined as the reaction time in milliseconds to correctly press the "different" key on the critical light-blue/dark-blue trials. (1 pt)
C. The results indicate that speakers of Language A, which has two separate color words, categorized the two blue shades as "different" faster than speakers of Language B, which has one word — 610 ms versus 685 ms, a difference of about 75 ms — even though both groups were highly accurate. (1 pt — direction plus the cited numbers)
D. The researchers applied informed consent: participants gave written informed consent before taking part. (Alternatively, they applied confidentiality by identifying participants only with a code number, or the right to withdraw, since participants were told they could stop at any time without losing payment.) (1 pt)
E. Generalizability is somewhat limited because all participants were adults with a mean age of about 24, so the findings may not extend to children or older adults; the two language groups were, however, matched on age and education and screened for normal color vision, which strengthens the internal comparison and makes it more likely the speed difference reflects language rather than a confound. The reliance on speakers of just two specific languages also limits how far the pattern can be generalized to all languages. (1 pt — a clear, evidence-backed statement of limited generalizability)
F. The findings refute the strong version of linguistic relativity (linguistic determinism) while supporting the weak version. Linguistic determinism claims that language determines thought — that without a word for a concept, a person cannot perceive or think it. But here, Language B speakers, who lack a separate word for the two blues, were still highly accurate (above 97%) at telling the shades apart; they could clearly perceive the difference, just slightly more slowly (685 ms vs. 610 ms). Because the speakers without two words were not unable to make the distinction, language did not determine their perception, which contradicts the strong version. The data instead fit the weak version: having two color words influenced the speed of categorization without controlling whether the distinction could be made at all. (2 pts — states the support/refute position clearly AND correctly applies the strong-vs-weak distinction with study evidence.)
1. (B) Phoneme. The smallest unit of sound. (A) morpheme is the smallest unit of meaning; (C) a syllable is larger and built from phonemes; (D) "semantic marker" is not the term.
2. (D) four. un- + help + -ful + -ly = four morphemes. Students often miscount by counting syllables or stopping at help/ful/ly. Each affix that adds meaning counts.
3. (A). Early babbling is universal and includes sounds from languages the infant has never heard; only later does it narrow to the native phonemes. (B) and (D) overstate environment/imitation; (C) the one-word stage starts near 12 months, not 5.
4. (C) Telegraphic speech. Two content words, function words dropped — the two-word stage. (A) cooing is vowel sounds in infancy; (B) holophrastic is one word standing for a whole thought; (D) overregularization is a grammatical-rule error like "goed."
5. (B). Chomsky's poverty-of-the-stimulus argument: children generate novel sentences and unmodeled errors, so reinforcement/imitation can't fully explain acquisition. (A) misstates the critique; (C) is sociocultural, not Chomsky; (D) is false — deaf infants do coo/babble early.
6. (C). "Runned" shows the child has extracted the regular "-ed" rule and overapplied it — overregularization, a sign of rule-learning. (A) and (D) misread progress as loss; (B) is the discredited reinforcement explanation — the child invented a form no one modeled.
7. (A) Language acquisition device. Chomsky's hypothesized innate mechanism. (B) Broca's area is a brain region for speech production, not Chomsky's construct; (C) and (D) are not real terms in this context (operant chamber is Skinner's apparatus).
8. (B). Overregularization errors ("goed") are produced and not modeled or reinforced, directly undermining a pure imitation/reinforcement account. (A) is consistent with behaviorism; (C) and (D) don't bear on the acquisition mechanism.
9. (B) Critical/sensitive period. Genie and second-language research both show that full native mastery — especially grammar and accent — is hard to achieve after the early window closes. (A) determinism is about language shaping thought; (C) is about cognitive benefits of bilingualism; (D) is a developmental stage, not the relevant principle.
10. (B) Generativity. Producing infinite novel rule-governed combinations from finite symbols is generativity — necessary for language but, the scientist argues, not sufficient for understanding. (A), (C), (D) are unrelated to the "infinite novel combinations" idea.
11. (B). The bilingual advantage is linked to strengthened executive function from constantly inhibiting one language while using the other. (A) is false (each single language may have a smaller vocabulary); (C) and (D) are unrelated mechanisms.
12. (B). Current consensus: the strong version (determinism) is rejected; the weak version (language influences thought) has empirical support. (A) reverses it; (C) overclaims; (D) is false.
13. (A). The table shows an ordered sequence over the first ~2 years. (B), (C), and (D) all contradict the listed onset ages (babbling precedes one-word, which precedes two-word; cooing and babbling are not simultaneous at birth).
14. (B). The U-shaped curve is the textbook overregularization pattern: memorize "went" → acquire the "-ed" rule and overapply it ("goed") → learn the exception again. (A) misreads it as loss; (C) is the reinforcement explanation, which the unmodeled "goed" undercuts; (D) misuses the critical period.
15. (B). Damage to Broca's area impairs effortful production while comprehension is relatively spared, showing production and comprehension can dissociate across brain regions. (A) reverses Broca's role (comprehension is more tied to Wernicke's area); (C) and (D) are false.
| Part | Point(s) | Earned if the response… |
|---|---|---|
| A | 1 | Identifies the method as a quasi-experiment / correlational / non-experimental comparison of pre-existing language groups (not a true experiment). |
| B | 1 | Operationally defines the DV as reaction time in ms on the critical "different" trials. |
| C | 1 | States Language A was faster AND cites the numbers (610 vs. 685 ms). |
| D | 1 | Names a real ethical guideline (informed consent / confidentiality / right to withdraw) AND ties it to a specific detail in the study. |
| E | 1 | Commits to a generalizability judgment AND backs it with specific study evidence (e.g., adults M ≈ 24; only two languages; matched/screened groups). |
| F | 2 | (1) Clearly states the findings refute determinism / support the weak version; (2) correctly applies the strong-vs-weak distinction using study evidence (Language B's high accuracy shows perception was not determined by language). |
PsyIQ · Lesson 14 of 30 · Unit 2: Cognition. Includes AAQ (Article Analysis Question) FRQ practice. Q1-style and AAQ practice modeled on the redesigned (2025+) AP Psychology exam. Not affiliated with the College Board. AP is a registered trademark of the College Board. Content pending external psychology QC.
▶ Next up: Mock Exam 1 — Units 1–2 diagnostic (38 MC + 1 AAQ). Use it to find your weak spots before Phase 3.