Last 10 minutes of class on Wednesday, November 4, 2026.
Everyone takes the exam at the same time.
Overview of Human Language, Words
You will take the exam on paper. There is 10 minute time limit, or 15 minutes for those with time-and-a-half accommodations.
You may bring one sheet of paper with notes written on both sides.
Do each of the following to maximize your preparation:
Were you able to check off every box?
The learning objectives for the course up to this point in time include you having developed a familiarity with, and an ability to discuss:
The fact that active recall is better for acquiring long-term knowledge does not mean that outlines and concept maps are not useful. Learners should use multiple techniques—think “both and” rather than “either or.”
Overview of Human Language
Three views of language
Mathematical (formal language theory): sets of strings over an alphabet
Evolutionary (human language): shaped by natural selection and cultural transmission
Probabilistic (language models): statistical patterns learned by neural networks
How to study language
Inherently interdisciplinary
Archaeology, anthropology, psychology, linguistics, computer science,
philosophy, cognitive science, genetics, ethology, neuroscience
Mithen's The Language Puzzle (2024)
Language as a puzzle of many fragments
Frame: 6 million years since the LCA; nature of language itself
Interior: primate calls, fossils, archaeology, words, acquisition,
transmission, metaphor and thought, genetics
Why study language
Professional: CS, cognitive science, AI, NLP, LLMs, education,
speech therapy, translation, text-to-speech, speech-to-text
Personal: communication, persuasion, relationships, defense against manipulation
Academic: theories get challenged and replaced as evidence accumulates
Once-dominant theories now strongly challenged
Hockett's 16 design features (1960s)
9 shared with other primates, 7 uniquely human
A "non-starter" as a framework for evolutionary linguistics
Chomsky's Universal Grammar
Innate Language Acquisition Device (LAD), common deep structure
Poverty of the stimulus argument
Challenged by cross-linguistic variation and the Iterated Learning Model
Pinker's The Language Instinct (1994)
Language as a Darwinian instinct, like echolocation in bats
Critiques: UG-sympathetic, ignores social and cultural factors,
language changes faster than genes can
Language evolution
Chimpanzees: gestures, expressions, vocalizations, ASL signs, but no recursion
Mithen's Hmmmmm hypothesis: proto-language was
Holistic, manipulative, multi-modal, musical, mimetic
Highlights
LCA (6 mya): small brains, rough vocal tracts, holistic calls
Aridity (4 mya): grasslands, larger groups, bipedalism, tools, bigger brains
Smaller teeth, flatter faces → better vocal tracts → vowels, consonants,
iconic words
Earlier births → childhood → cultural transmission bottleneck → syntax
Broca's area and mirror neurons (1.5 mya)
New environments → new (iconic) words → generalizations
Arbitrary words from mishearings, mispronunciations, convention, efficiency
Homo sapiens: cognitive fluidity, abstraction, metaphor, storytelling,
the supernatural, persuasion, deception
Features of language (Hockett's, still useful as descriptors)
Arbitrariness
Displacement (spatial and temporal)
Duality of patterning (phonemes → morphemes)
Discreteness
Productivity
Cultural transmission
Utterances
Words combined into utterances according to rules
Spoken, signed, or written
Prosody
Spoken: intonation, stress, rhythm, tempo, pauses
Signed: posture, movements, pauses, facial expressions
Words (preview)
Phonology, morphology, lexicology
Syntax, semantics, pragmatics
Lexical vs. grammatical; word-like entities (oops, hmm, um)
Iconic vs. arbitrary
Lexicons; children learn ~10 words a day between ages 2 and 8
Phonemes, the IPA, writing systems (Latin alphabet fit for Tlingit, English)
Theories of meaning: Prototype Theory, Twin Earth, Causal Theory of Reference
Lexical relations
Grammar
Rules for combining words into phrases and sentences
Rules (programming languages) vs. guidelines (natural language)
Grammatical but incomprehensible or absurd sentences
"The rat the cat the dog the cow tossed worried swatted ate the malt"
"Colorless green ideas sleep furiously"
Hierarchical structure and recursion
Where does grammar come from?
Chomsky: poverty of the stimulus, LAD, Universal Grammar
Tomasello, Christiansen: general learning + social interaction suffice
Kirby et al.: simulations of cultural transmission (Iterated Learning)
Pidgins and creoles
Nicaraguan Sign Language
Grammar is not innate; it emerges
Pragmatics
How context influences meaning
Implicature
Grice's maxims: quality, quantity, relation, manner
Deixis: I, you, here, now, this, that, tomorrow
Linguistic diversity
~7,000 languages today
Variation in phonology, morphology, syntax, semantics, pragmatics
Mechanisms of change: vowel shifts, simplification, grammaticalization,
borrowing, shortening, reanalysis, contact, migration, splitting
Language family trees, Glottolog
Evidence against UG
Apparent lack of recursion, no numbers or counting,
ergative-absolutive, topic-prominent, time without tense,
tonal languages, sign languages
WALS: "universals" turn out to be tendencies
Language and thought
Linguistic relativity (Sapir-Whorf), weak and strong versions
Space (egocentric vs. allocentric), time (linear vs. cyclic),
number (precise vs. approximate), inner speech,
immediacy of experience (Pirahã)
Signs and symbols
Shell beads (~100,000 BCE), ochre engravings (~75,000 BCE),
Lascaux and Chauvet (30,000–17,000 BCE)
Peirce: icons, indices, symbols
Symbols: convention, memory, separable word and meaning, abstraction
Symbol grounding problem
Reference and joint attention (~9 months); pointing rare in other primates
Connections to formal languages and LLMs
Language did not evolve for communication alone
Human language is more than words and grammar
Formal languages are just words and rules
LLMs: trained on evolved language, learn patterns from data alone,
lack embodiment, grounding, joint attention
Do LLMs understand? Is a vector a meaning?
The theoretical landscape
Theories: organized knowledge with explanatory and predictive power
Plato's Cratylus: natural origin vs. convention
Groups of theories
Generative and formal grammars
Functional, cognitive, and usage-based
Philosophy of language and meaning
Mathematical and computational
Language origin and evolution
Psycholinguistics and acquisition
Writing systems
Sign language
Impact of the LLM era
LLMs acquire competencies once thought to require innateness
Resurgence of distributional and predictive theories
Formal vs. functional linguistic competence
Stochastic parrots
Words
What are words?
Units of meaning that can stand alone in an utterance
Hard to pin down; agglutinative languages
Tokens in language processing; Byte-Pair Encoding
Basic terminology
Spoken word, phoneme, vowel, consonant, syllable
Phonology, morpheme, morphology
Syntax, semantics, pragmatics
Chereme and cherology (now phoneme and phonology for sign language)
Grapheme
Lexical vs. grammatical words
Lexical: carry meaning, stand alone, open class, very many
Nouns, verbs, adjectives
Concrete vs. abstract
Grammatical: function words, glue of syntax, closed class, very few
Prepositions, conjunctions, pronouns, determiners
Fuzzy boundary (auxiliary verbs)
Word-like units and vocalizations
Iconicity
Iconic (sound-symbolic, imagistic) vs. arbitrary words
Onomatopoeia: buzz, bang, hiss, plop, honk
Saussure: l'arbitraire du signe (signifier vs. signified)
Ideophones marginalized; Pinker's "quaint curiosity"
Comeback: mil/mal, maluma/takete, bouba/kiki
Consistent across languages; cross-modal perception
Wescott's Linguistic Iconism
High front vowels: small, fast, light
Low back vowels: large, slow, heavy
Laterals: fluid motion
Labials: roundness, fullness
Stops: abrupt actions
Dentals: small, sharp, precise
Phonesthemes (language-specific): sl-, fl-
Most words are still arbitrary
Lexicons
A language's inventory of words and meanings
Adult English: ~50,000–100,000 passive, ~20,000–30,000 active
Children: ~10 words per day, often from a single exposure
LLMs: trillions of tokens, weaker novel-word generalization
Morphemes
Smallest units of meaning
Free vs. bound, lexical vs. grammatical
Free lexical: dog, happy, walk
Bound lexical: bio-, tele-
Free grammatical: and, but, the, of
Bound grammatical: -s, -ed, -ing, -ment
Affixes: prefix, suffix, infix, circumfix
Derivational: change meaning or category (teach → teacher)
Inflectional: mark grammatical features (walk → walks)
Isolating, synthetic, polysynthetic languages
Verb features: person, number, tense, aspect, mood, voice, gender
Word meanings
Rosch's Prototype Theory: prototypes, graded membership
Is a hot dog a sandwich?
Works well for LLMs: vector similarity
Putnam's Twin Earth: meaning not entirely in the head
Causal Theory of Reference: meaning anchored by chains of use
Lexical relations
Synonymy
Antonymy: gradable, complementary, converse
Hyponymy: hyponyms and hypernyms
Polysemy: one lexeme, related senses
Homonymy: distinct lexemes, coincidental shared form
Word vectors: king − man + woman ≈ queen
Evolution of words (Mithen, Ch. 13)
English as a case study
Semantic change
Broadening, narrowing
Amelioration, pejoration
Weakening, strengthening
Word conversion
Metonymy
Causes: misunderstanding, metaphor, social tactics, efficiency
New words
Coinage, conversion, borrowing, compounding, derivation,
backformation, reduplication, blending, clipping, acronyms
Law of Brevity (Zipf's Law of Abbreviation) vs. Zipf's Law
goodbye, Imma, n't, lord, lady, won't
Sound change (covered under phonology)
Grammaticalization
will, gonna, finna, have (perfect), let's, because
This is a timed mini-quiz which tests for immediate understanding of topics and not your ability to work out problems over an extended duration of time. There will be 10 questions. Some may be multiple choice, multi-select, matching, and very short answer.
Here are what the 10 questions will cover:
All content on the assigned readings is fair game for questions, so do not neglect the readings, and by all means do the recall questions!