Language & Intelligence

🧠 Language & Intelligence

Part 1 of 7 — The Building Blocks of Language

🔑 Key Concept: Language is a shared system of arbitrary symbols that is rule-governed and generative. Small units of sound combine into units of meaning, and rules for combining words let speakers produce an unlimited number of new sentences.

What makes something a language?

The units of language

Pragmatics, the social rules for using language in context, is excluded from the AP exam (CED 3.5.A); focus on the five units above.

Why This Matters for the AP Exam: Questions often ask you to count or identify units. Phonemes are about SOUND; morphemes are about MEANING. English uses about 40 phonemes, and those few sounds build every English word.

Concept Check — Units of Language 🎯

📚 Worked Examples: Counting and Comparing Units

Example 1 — Phonemes are sounds, not letters.

• "ship" has 4 letters but 3 phonemes: /sh/ /i/ /p/.
• "cats" has 4 phonemes (/k/ /a/ /t/ /s/) and 2 morphemes ("cat" + plural "-s").
• "box" and "rocks" end with the same sounds, /ks/. But "box" is 1 morpheme, while "rocks" is 2 (rock + plural "-s"). The same sound can be a morpheme in one word and not in another.

Example 2 — Morphemes can be words or word parts.

Example 3 — Syntax and semantics are different.

• "Colorless green ideas sleep furiously" follows English syntax (adjective, adjective, noun, verb, adverb) but makes no sense. It has correct syntax and broken semantics.
• "The dog bit the man" and "The man bit the dog" use the same words. Only the syntax (word order) differs, and that changes the meaning.

Example 4 — Generativity. With about 40 phonemes, thousands of morphemes, and a limited set of rules, a speaker can build sentences that have never been spoken before, such as "The purple giraffe borrowed my calculator during lunch." Rules can also repeat inside one another ("the friend of the sister of the coach..."), so there is no longest possible sentence.

Example 5 — Sign languages are full languages. Sign languages such as American Sign Language have all of the same features. Handshape, location, and movement combine like phonemes, signs carry meaning like morphemes, and a grammar orders signs into sentences. Language does not require sound.

Recall Practice ✍️

🎯 AP Exam Strategy: Language Structure Traps

Memory Trick: Phoneme = Pronunciation (sound). Morpheme = Meaning. Syntax = Sequence (word order). Semantics = meaning.

Identify the Unit or Feature 🔍

Applied Scenarios 🔬

🔑 AP Connection: These units also explain children's speech errors. A child who says "foots" has learned the plural morpheme "-s" and applied it too widely. Part 2 follows how children move from cooing to full sentences and why those errors are a sign of progress.

🧠 Language & Intelligence

Part 2 of 7 — How Language Develops

🔑 Key Concept: In every culture studied, children communicate with gestures such as pointing before they speak, then move through the same stages: cooing, babbling, the one-word stage, and telegraphic speech. Along the way they make predictable errors, such as overgeneralization, that show they are learning rules. Language develops most easily during a sensitive period in early childhood.

The stages of spoken language

The ages below are approximate. Individual children vary by months, but the order of the stages is the same across languages.

Babbling narrows. Early babbling includes sounds from many languages. By about 10 months, a baby's babbling uses mostly the sounds of the language spoken at home. Deaf infants who are exposed to sign language "babble" with their hands, repeating the building blocks of signs.

Telegraphic speech gets its name from old telegrams, where people paid by the word and left out every word that was not essential.

Concept Check — Stages 🎯

📚 Gestures Come First

Across all cultures, infants communicate with their hands before their words. Around 9 to 12 months, babies point to show an adult something interesting or to ask for an object, wave goodbye, and raise their arms to be picked up.

• Pointing shows that the child understands that another person can share attention with them, which is a foundation for words.
• Children often combine a word with a gesture, such as pointing at a cup while saying "drink," before they combine two words.
• Gestures do not disappear when speech begins. Adults in every culture gesture while they talk.

📚 Overgeneralization: Errors That Show Learning

Overgeneralization (also called overregularization) means applying a grammar rule to words that are exceptions to it.

Why these errors are evidence of rule learning:

1. No one models them. Adults do not say "goed," so the child is not simply copying.
2. They appear after correct forms. Many children first say "went" correctly (a memorized word), then switch to "goed" once they learn the "-ed" rule, then return to "went" as they learn the exceptions. The temporary dip is a sign of progress.
3. Children apply rules to new words. Show a child a made-up creature called a "wug," then two of them, and the child says "wugs," a word they have never heard.

🔑 AP Connection: If a question describes a child applying a regular rule to an irregular word, the answer is overgeneralization, a normal step in development, not a language delay.

Match the Example to the Concept 🔍

📚 Sensitive Periods for Language

Research suggests that critical or sensitive periods in infancy and childhood have strong effects on development, especially for language (CED 3.2.B.4).

Language fits the sensitive-period pattern best. Three lines of evidence:

1. Speech-sound tuning. Young infants can tell apart speech sounds from many languages. By around their first birthday, they become much better at the sounds of their home language and worse at sounds their language does not use.
2. Age of second-language learning. People who move to a new country as young children usually reach native-like grammar and accent. People who arrive as teens or adults can become fluent, but they usually keep an accent and make more grammar errors.
3. Late first-language exposure. Deaf children who are not exposed to any language, signed or spoken, until later childhood tend to have lasting difficulty with grammar, even after years of practice. Children exposed to sign language from birth reach milestones on the usual schedule.

🔑 AP Connection: A sensitive period does NOT mean learning stops. Adults learn new languages every day. It means early childhood is when complete, native-like language comes most easily.

Applied Scenarios 🔬

🎯 AP Exam Strategy: Language Development Traps

AP Tip: On data questions, look for whether a skill improves or declines with age of exposure. A steady decline in grammar scores as age of arrival rises is the classic sensitive-period pattern.

🧠 Language & Intelligence

Part 3 of 7 — Defining Intelligence: One Ability or Many?

🔑 Key Concept: Throughout history, psychologists have not agreed on how to define or measure intelligence, and both the definition and the measurement can be subject to bias. The central debate is whether intelligence is one general ability (g) or a set of multiple abilities (CED 2.8.A.1).

A working definition

Most definitions describe intelligence as the mental ability to learn from experience, solve problems, and adapt to new situations. Beyond that shared core, definitions differ:

• Across cultures. Some cultures stress quick, abstract reasoning and school knowledge. Others stress social responsibility, practical skill, or wisdom in getting along with others. A definition built in one culture may miss abilities another culture values.
• Across time. Early test makers often treated familiarity with their own culture's vocabulary and customs as a sign of intelligence, which gave an advantage to people raised in that culture.
• Across researchers. Whatever definition a researcher chooses decides what the test measures, and therefore who scores high.

The central debate

Charles Spearman proposed g after noticing that people's scores on very different school subjects and tests tended to rise and fall together.

Named theories that list specific types of intelligence are not tested on the current exam; know the g versus multiple-abilities debate and the evidence on each side.

Concept Check — g or Multiple Abilities? 🎯

📚 Evidence on Both Sides

The middle ground. Many researchers today describe intelligence as a general ability plus several more specific abilities. The two views disagree mostly about how much weight to give each.

Another way to divide abilities: crystallized and fluid

Topic 3.4 introduces a related distinction:

• Crystallized intelligence is accumulated knowledge and verbal skill, such as vocabulary. It tends to hold steady or grow through adulthood.
• Fluid intelligence is the ability to reason quickly and solve new problems. It tends to decline gradually in later adulthood.

Because the two follow different paths with age, this distinction is evidence that intelligence has more than one part.

Reading the Data 📊

Practice data invented for this lesson. A researcher gives 300 high school students four tests and reports the correlation between each pair of tests.

📚 Why Intelligence Is Hard to Define and Measure Fairly

The CED stresses that defining and measuring intelligence "continues to be elusive and can be subject to bias." Three reasons:

1. The definition shapes the test. If intelligence is defined as speed on abstract puzzles, a test will reward speed, and people from settings that do not emphasize timed tests may score lower without being less capable.
2. Test content reflects a culture. A question that asks what to do "if you lose a library book" assumes that the test taker grew up with libraries. Unfamiliar content lowers scores for reasons unrelated to reasoning ability.
3. A single number can hide strengths. Under the multiple-abilities view, reporting only one total score can misrepresent a person with a very uneven profile.

These problems connect directly to Part 4 (does a test measure what it claims to, which is construct validity?) and Part 5 (how test scores have been misused).

🎯 AP Exam Strategy: Defining-Intelligence Traps

Recall Practice ✍️

Applied Scenarios 🔬

🧠 Language & Intelligence

Part 4 of 7 — Intelligence Testing

🔑 Key Concept: Early intelligence tests produced an intelligence quotient (IQ) by dividing mental age by chronological age. Today, IQ scores are often used to identify students for educational services. Like every psychological assessment, an intelligence test is useful only if it is standardized, reliable, and valid (CED 2.8.B).

Where IQ came from

In the early 1900s, the French psychologist Alfred Binet was asked to find schoolchildren who needed extra help. He compared each child's performance with the typical performance of children at each age, giving a mental age. Binet warned that his test should be used to help children, not to rank them for life.

Why modern tests changed the formula. Mental age stops rising in adulthood. A 40-year-old who performs like a typical 40-year-old and a 20-year-old who does the same would get very different ratio IQs, which makes no sense. Comparing people with others their own age solves the problem.

Labeling or describing cognitive abilities and disabilities by score range is outside the scope of the AP exam (CED 2.8.B.1); focus on how scores are calculated and interpreted.

Concept Check — Calculating and Checking Tests 🎯

📚 The Three Psychometric Principles

All psychological assessments, including intelligence tests, should follow sound psychometric principles (CED 2.8.B.2).

Two types of reliability

Two types of validity

Reliability is necessary but not sufficient. A bathroom scale that always reads five pounds heavy is perfectly consistent (reliable) but wrong (not valid). A test cannot be valid if it is not reliable, because random noise cannot measure anything well.

Worked example: evaluating a new test

A company designs a 60-item "problem-solving test" for high school students.

1. Standardization: Every student takes it on a computer, with the same instructions and a 45-minute limit. ✔
2. Reliability: Scores on odd and even items correlate at +.91 (split-half), and scores a month apart correlate at +.87 (test-retest). ✔
3. Construct validity: Scores correlate +.80 with typing speed but only +.15 with teacher ratings of problem solving. ✘ The test seems to measure typing, not problem solving.
4. Conclusion: The test is standardized and reliable but not valid for its stated purpose.

Name the Principle 🔍

📚 Reading IQ Scores on the Normal Curve

Modern intelligence test scores are set so that they form a normal (bell-shaped) distribution with a mean of 100 and a standard deviation of 15.

Percentile rank tells you the percentage of people who scored at or below a given score.

• A score of 100 is at about the 50th percentile.
• A score of 115 (one SD above the mean) is at about the 84th percentile: 50% below the mean plus 34% between 100 and 115.
• A score of 85 (one SD below the mean) is at about the 16th percentile.

📚 Achievement Tests vs. Aptitude Tests (CED 2.8.D.1)

The difference is the purpose, not the format. The same kind of reading question can appear on both. An aptitude test is judged mainly by its predictive validity; an achievement test is judged by how well it covers what was taught.

🎯 AP Exam Strategy: Testing Traps

Memory Trick: Reliability = Repeatability. Validity = measures the right Variable.

Recall Practice ✍️

Applied Scenarios 🔬

🧠 Language & Intelligence

Part 5 of 7 — Intelligence Controversies

🔑 Key Concept: Intelligence scores are shaped by genes, environments, and the conditions of testing. Systemic issues such as poverty, discrimination, and unequal schooling affect scores, and test results have been misused to limit people's opportunities (CED 2.8.B.3, 2.8.C, 2.8.D.2).

🗣️ Real-World Example

In a well-known 1995 study, Claude Steele and Joshua Aronson gave Black and White college students a difficult verbal test. When the test was described as a measure of intellectual ability, Black students scored lower than White students with similar prior test scores. When the same test was described as a problem-solving exercise that did not measure ability, the gap largely disappeared. Same test, same kind of students; only the framing changed.

Why This Matters for the AP Exam: Heritability describes differences within a group, not any one person. Even a high heritability estimate does not mean that a person's own score is fixed by genes.

Concept Check 🎯

📚 Genes and Environment: The Evidence

Evidence for genetic influence (twin and adoption studies, CED 1.1.A.3)

Evidence for environmental influence

📚 Within-Group vs. Between-Group Differences (CED 2.8.C.2)

IQ scores vary more within a group than between groups. The range of scores inside any one group is far larger than the difference between two groups' averages.

Seed analogy: Plant seeds from the same varied batch in two pots, one with rich soil and one with poor soil. Within each pot, height differences reflect the seeds' genes, because the soil is the same. The difference between the pots comes entirely from the soil. In the same way, a trait can be highly heritable within groups while the gap between groups is caused by environment.

Personal and sociocultural biases can also affect how an individual's score is interpreted. A tester who expects low performance from a group may read the same score differently.

Match the Finding to the Concept 🔍

📚 Stereotype Threat and Stereotype Lift (CED 2.8.B.3)

Both effects can make test scores unequal for reasons that have nothing to do with ability. That is why researchers try to build assessments that are socio-culturally responsive: familiar content, neutral framing, and testing conditions that do not call attention to group stereotypes.

📚 How Test Scores Have Been Misused (CED 2.8.C.3)

Scores from intelligence tests have been used to limit access to:

• Jobs (screening applicants with tests unrelated to the work)
• Military ranks (assigning recruits to roles and promotions by test score)
• Educational institutions (keeping students out of programs or schools)
• Immigration to the US (early-1900s testers claimed that some immigrant groups were less intelligent, using tests given in English to people who had just arrived)

This history is an example of applying psychological concepts in inappropriate or discriminatory ways (skill 1.B).

📚 Fixed vs. Growth Mindset (CED 2.8.D.2)

People's beliefs about intelligence can affect academic achievement. Psychologist Carol Dweck described two beliefs:

• Fixed mindset: intelligence is set at birth. Students who hold this view tend to avoid challenges and give up after failure, because failure seems to reveal low ability.
• Growth mindset: intelligence can grow through effort and experience. Students who hold this view tend to persist and see mistakes as information.

In studies by Dweck's group, children praised for being smart were more likely to choose easy tasks afterward than children praised for their effort.

🎯 AP Exam Strategy: Controversy Traps

Applied Scenarios 🔬

🧠 Language & Intelligence

Part 6 of 7 — Problem-Solving Workshop and AAQ Practice

🔑 Key Concept: The exam tests whether you can APPLY language and intelligence concepts to new scenarios and to research studies. This part practices both: first with short scenarios, then with an Article Analysis Question (AAQ) built on a practice study.

🔧 Problem-Solving Framework

Quick Reference: Most Common Traps

Scenario Analysis 🎯

📚 Two Cases to Work Through

Case 1: The advanced-program test

A district uses a single intelligence test score to choose students for an advanced-learning program. Critics point out that the program enrolls far more students from wealthy neighborhoods.

Apply your knowledge:

• Systemic issues (2.8.C.2): Poverty and unequal schooling can lower scores, so a score gap may reflect opportunity rather than ability.
• Construct validity (2.8.B.2): Is the test measuring reasoning, or familiarity with vocabulary common in wealthier homes?
• Within vs. between groups: Scores vary far more within each neighborhood than between neighborhoods, so many qualified students from every neighborhood exist.
• Socio-culturally responsive testing (2.8.B.3): Neutral framing and familiar content reduce stereotype threat.
• g vs. multiple abilities (2.8.A.1): One total score can hide a student's strong specific ability.

Case 2: The bilingual toddler

Parents worry because their 2-year-old, who hears two languages, knows fewer words in each language than a cousin who hears one.

Apply your knowledge:

• Total vocabulary: Counting words across both languages gives a fairer comparison than counting words in one.
• Stages (3.5.B.1): Ask whether the child is moving through the usual sequence. A child using telegraphic phrases such as "more milk" is on track.
• Overgeneralization: Errors such as "I goed" in either language show rule learning.
• Sensitive period (3.2.B.4): Early exposure to both languages is an advantage, because language is learned most easily in early childhood.

Decision Guide

✍️ AAQ Practice

The Article Analysis Question gives you one study summary and asks six focused tasks (7 points, 6 parts): identify the research method, state an operational definition, describe what a statistic shows, identify an ethical guideline and how it was applied, explain generalizability with evidence, and explain how a finding supports or refutes the hypothesis.

Practice study summary

This practice study is invented for this lesson; it is not a real published study.

Introduction: The researchers hypothesized that students told a math test measures their ability would score lower than students told the same test is a practice puzzle activity.

Participants: 120 ninth graders (64 girls, 56 boys) from two public high schools in one Midwestern city volunteered. Parents gave written informed consent, and students gave written assent and were told they could stop at any time.

Method: Students were randomly assigned to one of two instructions. The "ability" group read that the test "measures your true math ability." The "puzzle" group read that the test is "a set of practice puzzles that do not measure ability." Everyone then took the same 20-item multiple-choice math test in a 25-minute session. Score was the number of items answered correctly (0 to 20). Test forms were labeled with ID numbers, and afterward all students were told the real purpose of the study.

Results: The ability group averaged 11.2 correct (SD = 3.1); the puzzle group averaged 13.6 correct (SD = 3.0). The difference was statistically significant, and the effect size was d = 0.6.

Model responses for the six AAQ tasks

A. Research method: "This is an experiment: the researchers manipulated the test instructions (the independent variable) and randomly assigned students to the two conditions."

B. Operational definition: "Math performance was operationally defined as the number of items answered correctly out of 20 on the test."

C. Statistic: "The means show that the ability group answered about 2.4 fewer items correctly on average (11.2 vs. 13.6). Because the difference was statistically significant, it is unlikely to be due to chance, and an effect size of 0.6 is a medium effect."

D. Ethical guideline: "The researchers used informed consent and assent: parents signed consent forms and students gave assent after being told they could stop at any time." (Confidentiality, through ID numbers, or debriefing, by explaining the real purpose afterward, would also work.)

E. Generalizability: "The findings generalize only to a limited extent. All 120 students were ninth graders from two schools in one Midwestern city who volunteered, so the results may not apply to older students, students in other regions, or students who would not volunteer."

F. Support or refute: "The results support the hypothesis. Students told the test measured their ability scored lower (11.2) than students told it was a puzzle activity (13.6). Since only the instructions differed and students were randomly assigned, the framing likely caused the difference, consistent with stereotype threat and test anxiety."

⚠️ Always use specific numbers and details from the summary. A general answer such as "the sample is too small" earns nothing unless the summary shows why.

AAQ Practice — Use the Practice Study Above 🎯

Match the Scenario to the Concept 🔍

Applied Scenarios 🔬

🧠 Language & Intelligence

Part 7 of 7 — Synthesis & AP Review

🔑 Key Concept: Language and intelligence share one big idea: both are complex mental abilities that grow through an interaction of biology and experience, and both are measured in ways that can be done well or badly.

Master Integration Table

Comprehensive Review 🎯

📚 Cross-Unit Connections

How the parts fit together

1. Language is built from units (Part 1) and develops in a fixed order (Part 2), with a sensitive period that makes early exposure matter.
2. Intelligence is hard to define (Part 3). The g vs. multiple-abilities debate shapes what tests try to measure.
3. Tests must be standardized, reliable, and valid (Part 4), and achievement and aptitude tests serve different purposes.
4. Scores reflect more than ability (Part 5): environments, stereotypes, and systemic inequities all affect them, and scores have been misused.

Final Recall Check ✍️

✍️ EBQ Practice

The Evidence-Based Question gives you three short sources and asks for an argument (7 points, 3 parts): (A) a specific, defensible claim; (B) evidence from one source plus an explanation that uses a psychological concept; (C) evidence from a DIFFERENT source plus an explanation that uses a DIFFERENT concept. Cite sources as "(Source 1)."

Prompt: Develop an argument about whether schools should rely on a single intelligence test score to decide which students enter advanced programs.

The three sources below are invented for this lesson; they are not real published studies.

Source 1: Researchers gave 200 high school students the same reasoning test. Half read that the test "measures natural intelligence"; half read that it was "a problem-solving warm-up." Students from groups stereotyped as weaker in reasoning scored significantly lower in the "intelligence" condition but not in the "warm-up" condition.

Source 2: A district compared students' placement-test scores with family income. Average scores were higher in high-income neighborhoods, but students in every neighborhood earned scores across the full range, and the spread of scores within each neighborhood was about ten times larger than the gap between neighborhood averages.

Source 3: A school tested students twice, one month apart, with its placement test. Scores correlated +.90 across the two sessions, but placement-test scores correlated only +.20 with students' later grades in the advanced program.

Model response

(A) Claim: "Schools should not rely on a single intelligence test score to place students in advanced programs, because scores are affected by testing conditions and may not predict success in the program."

(B) Evidence and explanation: "Students from stereotyped groups scored lower only when the test was described as measuring intelligence (Source 1). This shows stereotype threat: worry about confirming a negative stereotype about one's group lowered performance, so a single score can understate a student's real ability."

(C) Evidence and explanation: "The placement test was consistent across sessions (r = +.90) but barely related to later grades (r = +.20) (Source 3). The test is reliable but lacks predictive validity, so using it alone would admit and reject students based on a number that does not forecast success in the program."

⚠️ Part C must use a different source AND a different concept from Part B. Source 2 could support a third argument using within-group vs. between-group variation.

Final Exam Simulation 🔬