An Overview of Human Language

Getting Oriented

There are at least three ways to look at language:

• The mathematical view (Formal language theory): Languages are sets of strings over an alphabet.
• The evolutionary view (Human Language): Language is a product of evolution, shaped by natural selection and cultural transmission.
• The probabilistic view (Language models): Language is a system of probabilities and patterns, often modeled statistically and generated via neural networks.

We studied formal languages earlier in the course. Now let’s learn about human language.

How to Study Language

People have been studying language for thousands of years, looking into the question of what language is, where it came from, why we have it, what we can do with it, and what language might tell us about ourselves. We might never get a complete story. These questions are really complex.

Answering these questions requires research from archaeology, anthropology, psychology, linguistics, computer science, philosophy, cognitive science, genetics, ethology, neuroscience, and many other fields. Focusing on a single field will get you an incomplete picture.

The study of language is inherently interdisciplinary.

Steven Mithen presents the interdisciplinarity approach to language in his 2024 book The Language Puzzle. It is really good. You should read it. It looks at language as a puzzle with many fragments:

• The frame of the puzzle is made from:
  • Studying the last 6 million years of human history, from the time of the LCA of humans and chimpanzees to the present (Ch. 2)
  • The nature of language itself: words, rules, prosody, and the basis for linguistic diversity (Ch. 3)
• The interior fragments are assembled from:
  • Studying the calls and vocalizations of monkeys and apes as a foundation for the study of human language evolution (Ch. 4)
  • Fossil evidence for the development of the vocal tract (Ch. 5) and the size and shape of the brain (Ch. 11)
  • Archaeological evidence for symbolic and artistic behaviors (Ch. 15), and the use of fire (Ch. 10) and tools (Ch. 7)
  • Understanding words themselves, particularly the distinction between iconic words and arbitrary words (Ch. 6), and changes in word meanings, pronunciations, and roles (Ch. 13)
  • How infants acquire language (Ch. 9) and how language is transmitted between generations (Ch. 8)
  • How language impacts thought, such as how the use of metaphor shapes complex abstract concepts (Ch. 14)
  • Whether (and if so, how much of) a genetic predisposition exists for language acquisition (Ch. 12)

Mithen identifies and assembles these fragments to form a compelling, modern understanding of language evolution.

Why Study Language

Because it’s fun! Also for personal enrichment and intellectual curiosity.

Professional Reasons

Language plays a huge role in computer science, cognitive science, and artificial intelligence. A good understanding of human language is crucial for developing natural language processing systems and effective language models.

It’s also fun to contrast the precision of formal languages and programming languages with the messiness, ambiguities, redundancies, and exceptions of evolved human languages. LLMs see language as probabilities and patterns in a network not unlike the human brain (with its long-distance connections and distributed processing).

A technical study of language can also lead to rewarding careers in education, speech therapy, language translation, teaching reading and writing, building text-to-speech and speech-to-text systems, coaching communication and social skills, and other fields that can make a huge difference in people’s lives.

Personal Reasons

But we don’t study language just to get a job.

Wielding language well can enhance communication, persuasion, and understanding across different contexts and cultures. It empowers you to persuade, inform, and connect with others effectively. You might even make more friends and have stronger relationships. It provides a buffer against those who would use language against you.

Academic Reasons

A holistic study of language and linguistics is important, because we want to get closer to answering our questions about language and to deepen our understanding. Like all fields, powerful theories arise which over time are found to be lacking or become outdated as new evidence provides better explanations for the old speculations. The study of language is no different. Here are three examples of once widely accepted theories that have been strongly challenged as new research and evidence have come to light:

Topics in Human Language

Following is a brief overview of some key topics in the study of human language. After this overview, we’ll be taking deeper dives into many of these topics, including words, sounds, structure and meaning, language diversity, and the relationship between language, thought, cognition, and culture.

Language Evolution

We clearly use language differently than our nearest relatives the chimpanzees. They can communicate with gestures, facial expressions, and vocalizations, and can learn ASL signs but can’t seem to organize the signs recursively. Our LCA probably was very much like today’s chimpanzees.

The language we have today evolved with us. Mithen advocated the Hmmmmm hypothesis, which posits that language was preceded by “proto-language” that was Holistic, manipulative, multi-modal, musical, and mimetic.

But how did language evolve? Surely it was through a multitude of factors, all occurring over millions of years. A few historical highlights identified by Mithen in The Language Puzzle are:

• The LCA (6 mya) and other forest-dwelling primates had small brains and rough vocal tracts to make holistic calls, which they could do little more than make faster or slower, or louder or softer.
• Global climate became more arid 4 mya so hominins had to adapt to (predator-rich) grassland environments, where living in larger groups and learning signs (for communication and alliance building) helped. Genetic mutations led to more cognitive skills, statistical learning, tool usage, stone throwing, bipedalism to cover larger distances, and larger brains.
• The larger brains, smaller teeth, and flatter faces led to more efficient vocal tracts that could produce significantly many more sounds. Calls no longer had to be holistic: they could be made up of smaller meaningful units which became vowels and consonants, and iconic words arose.
• Larger brains + bipedalism led to babies being born earlier: childhood provided the cultural transmission bottleneck to allow syntax to emerge.
• 1.5 mya Broca’s area—shared neural real estate for language and manual dexterity such as tool making and use—appeared. Mirror neurons, which fire both when you act and when you observe action, also appeared.
• As humans spread though different environments, they encountered never before seen animals, plants, topographies, hazards, and weather conditions that needed describing. At first, all the new words were iconic but generalizations ultimately formed.
• Arbitrary words often arose due to repeated mishearings, mispronunciations, misunderstandings, and social conventions. The drive toward efficiency mattered too, simplifying long words and making some easier to say.
• More evolutionary changes caused our species, Homo sapiens, to develop more complex cognitive fluidity: a larger cerebellum, more globular brain shape, smaller occipital lobes, and larger neural networks led to words for abstract concepts, different categories of words, and metaphor. Humans began story telling and even to imagine a supernatural realm of ghosts and spirits and demons, and gain the ability to persuade and mislead.

Utterances

Modern language is characterized by combining words into utterances, according to various rules.

Utterances can be spoken, signed, or written.

Utterances have meanings which are influenced not only by the words themselves but also by prosody—intonation, stress, rhythm, tempo, and pauses in spoken language, and by body posture, movements, pauses, and facial expressions in signed languages.

Come up with utterances (sentences or phrases) using different forms of prosody to convey meaning changes such as statements vs. questions, emphasis, sarcasm, irony, etc.

If utterances are made up of words, then what, exactly, are words?

Words

Exactly what counts as a word is kind of hard to pin down, and varies between languages. It’s typically understood to be a unit of meaning that can stand alone in an utterance. There’s a lot of variation among languages, especially in the orthographic representations—think of agglutinative languages like Finnish, Hungarian, Turkish, Japanese, Swahili, Korean, Quecha, and Tamil.

There are different branches of linguistics that study words from various perspectives:

• Morphology: How words are made up of smaller pieces (called morphemes)
• Lexicology: The broader branch of linguistics that studies everything about words: their meaning, their history, their behavior, and how they relate to one another.
• Phonology: The study of the functional patterns of speech sounds (phonemes, e.g., vowels, consonants, etc.), and how phonemes create meaning

Putting words together into utterances leads to different dimensions of language study:

• Syntax: How words are structured to form phrases and sentences
• Semantics: What sentences mean
• Pragmatics: How language is used to communicate ideas and concepts

We’ll cover words in much more detail later, but by way of introduction, but for now, here’s a bit about what’s coming.

We’ll see how to classify words into lexical and grammatical words, and further into categories such as nouns, verbs, adjectives, prepositions, conjunctions, pronouns, and determiners. We’ll see word-like entities, like the English oops, shhh, hmm, uh, um, haha, tsk-tsk, yikes, oops. And (instinctive) screams, grunts, yawns, laughs, and other such vocalizations.

There’s another classification dimension: Some words are iconic, meaning their form resembles their meaning (e.g., buzz, bang, hiss, plop), and others are arbitrary (e.g., dog, cat).

The lexicon of a language is its inventory of words (and their meanings). Different languages have wide differences in the size of their lexicons! We’ll learn theories about why. Children acquire words at a remarkable pace—about 10 new words per day between ages 2–8—many of which are inferred from a single exposure in context. How? We’ll learn that too. And also how it is that new words are created every day.

Spoken language introduces a whole new set of elements to study. We’ll learn about phonemes (the smallest units of sound in a language that can distinguish meaning). Some spoken languages have only a few dozen phonemes, while others have over a hundred. The IPA (International Phonetic Alphabet) provides a standardized way to represent these sounds. And yes, we’ll learn all those crazy terms like bilabial, alveolar, velar, glottal, uvular, glottal, plosive, affricate, etc.

Then we’ll look a little at writing. A lot of languages are written with the Latin alphabet, even when the sounds don’t match the letters perfectly. Watch the following video to see if the Latin alphabet is a good fit for Tlingit.

We’ll also address the question of how words get their meanings. There are some theories, like Rosch’s Prototype Theory—that categories aren’t defined by certain conditions, but rather organized around prototypes (or best examples), such as a robin for the category "bird" (and not a penguin). Other members of the category are included based on their similarity to the prototype, leading to graded membership rather than a strict boundary. Then there’s Putnam’s Twin Earth Thought Experiment, which gives evidence for meaning not being entirely in one’s head (intensionally) but also (extensionally) within one’s environmental context. And another theory: Causal Theory of Reference, which suggests that words get their meaning anchored historically, through chains of use.

One more teaser for what is coming: Words are definitely related to each other. We’ll learn about synonyms, antonyms, hyponyms, hypernyms, homonyms, and polysemy.

Grammar

Words and word-like entities are combined (arranged) according to rules to form larger units like phrases and sentences. We need the rules because there’s a benefit to having not every arrangement of words be acceptable. The rules constitute a grammar. Where do the rules come from? How do we know if an utterance is well-formed?

Rules or Guidelines?

In programming languages, rules are explicit and prescriptive. In spoken natural language, we tend to treat them as guidelines, but in writing, they are often more rigidly followed.

A grammar will tell you where all the components of an utterance should go—based on the category (e.g., for English: noun, verb, adjective, noun phrase, verb phrase, subject, predicate, adverbial clause, prepositional phrase, etc.) of the word or phrase—of the components, but it won’t tell you what the utterance means. Here are two classic examples of perfectly grammatical (structurally sound) sentences:

• “The rat the cat the dog the cow tossed worried killed ate the malt” (incomprehensible)
• “Colorless green ideas sleep furiously” (semantically absurd)

Many (most?) human languages exhibit hierarchical structure, where smaller units combine to form larger units in a nested manner. They tend to all have recursion—the property of language that allows rules to be applied repeatedly, embedding structures within structures. This is what enables sentences to be infinitely long and complex, even with a finite set of rules and vocabulary.

We’ll cover these in much more detail later in the course.

Here’s a fascinating question: Where Does Grammar Come From? Chomsky thought language acquisition is too fast and input too impoverished for learning alone, putting forth the Poverty of the Stimulus argument: children know grammatical rules they’ve never seen exemplified. So he proposed a kind of innate Language Acquisition Device (LAD) for what became known as Universal Grammar. It was very popular for a while, but modern research, including the body of work by Tomasello and Christiansen, note that general learning mechanisms + social interaction are sufficient. The computer simulations of Kirby and colleagues support this view, showing that cultural transmission can lead to the emergence of structured language over generations.

There’s much more in the study of language acquisition that argues against the need for an innate Language Acquisition Device (LAD). Pidgins are rudimentary contact languages with no native speakers and minimal grammar. Creoles are pidgins that children acquire natively, with a spontaneously developed full grammar. And there’s a well-known case of deaf children in 1980s Nicaragua inventing a full sign language in one generation. Themselves. With no adults to mimic.

There seems to be a lot of evidence that given the right social conditions, grammar emerges.

Grammar is not innate. It emerges from the interaction of general learning mechanisms, social interaction, and cultural transmission.

Pragmatics

Pragmatics is the study of how context influences meaning, things like prosody, gestures, speaker attributes (age, gender, status), surroundings, relationship with the listener, context, innuendo, subtext, shared situational knowledge, and so on.

An interesting pragmatic concept isimplicature: a meaning that is implied but not explicitly stated. Implicature enables efficient and sometimes polite communication. Some classic examples:

• “If you could pass the guacamole that would be awesome.”
• “Nice store you have here. It would be a shame if anything were to happen to it.”
• “Crack the window.”
• “I was thinking we can take care of this here.” (spoken to a cop at a traffic stop with a $100 bill protruding from behind the license and registration)

Paul Grice proposed the following maxims of implicature for cooperative conversation:

• Quality: Your contribution should be true or something you have evidence for.
• Quantity: Make your contribution as informative as required.
• Relation: Be relevant.
• Manner: Be clear, brief, and orderly.

Efficient discourse is also aided by deixis: words or phrases whose reference (spatial, temporal, or personal) shifts with context, such as I, you, here, now, this, that, these, those, tomorrow, last week, the following.

Linguistic Diversity

There are ~7,000 languages in the world today, and many more that have gone extinct. They vary widely in their phonology (sound systems), morphology (word formation), syntax (sentence structure), semantics (meaning), and pragmatics (use in context). We’ll look later at many of these differences, and how they came to be. We’ll study vowel shifts, simplification, grammaticalization, borrowing, word shortening, reanalysis, contact, migration, splitting communities, and other mechanisms of language change.

Languages evolving into new languages is fascinating, and very often eye-opening. We’ll look at language family trees, such as this one:

The study of so many languages provides a good deal of evidence against Universal Grammar. We’ll be looking at many of the reasons why UG fell so out of favor. Here’s a quick sampling of observations that showed that several aspects thought to be universal simply aren’t so:

• Languages with an apparent lack of recursion
• Languages with an absence of numbers and counting
• Ergative-absolutive languages
• Topic-prominent languages
• Time without tense
• Tonal languages
• Sign languages

The World Atlas of Language Structures (WALS) is pretty cool, with many more examples. It shows that when linguists systematically catalogued features across thousands of languages, feature after feature that had been proposed as universal turned out to be highly variable (word order, case marking, use of morphology, how questions are formed, whether there’s a passive construction, tenses, subject-prominence, segmental phonology, and so on). The apparent “universals” mostly turned out to be tendencies or statistical preferences when real cross-linguistic data was gathered at scale.

Language and Thought

We can’t talk about language without considering its impact on thought. The question of whether and how much language influences thought (and vice versa) is complex and the subject of a lot of debate. Linguistic relativity asserts that language influences worldview or cognition. You’ll sometimes hear it called the “Sapir-Whorf hypothesis.” We’ll look at the weak and strong versions of this hypothesis.

We’ll also see how language may shape thought, by looking at:

• Whether space is conceptualized egocentrically or allocentrically
• Whether time is conceptualized linearly or cyclically
• Whether number is conceptualized precisely or approximately
• Where inner speech comes from
• Languages with immediacy of experience (e.g., Pirahã)

Signs and Symbols

Human language is overwhelmingly symbolic. This is extraordinary! No other species uses symbols at scale. We’ve found ornamental shell beads dating back to ~100,000 BCE, and geometric ochre engravings dating back to ~75,000 BCE. These are likely symbolic artifacts, indicating that our ancestors were using symbols long before the emergence of fully developed language. The cave paintings at Lascaux and Chauvet (30,000–17,000 BCE) also suggest symbolic thought, as they depict animals and scenes that may have had cultural or spiritual significance.

Symbols are one of three kinds of signs, the others being icons (signs resembling what they represent) and indices (causally connected to their referent). But symbols are fascinating because they have no intrinsic connection to their referent, just (a) shared convention, (b) memory, (c) the ability to hold “word” and “meaning” as linked but separate. What’s cool is you can make a symbol for anything, even abstractions: justice, infinity, recursion, love, LLM, the meaning of life, etc.

We’ll study symbols later, looking at where they come from and how they hook on to the world (what philosophers call the symbol grounding problem). There are a few theories about this, of course!

The concept of reference is really interesting. Before the word hooks onto its meaning, one has to learn or be taught what object is getting the word. Human infants develop a (proto-linguistic) joint attention at ~9 months. Without joint attention, ostensive definition (“that’s a dog”) doesn’t work. What’s fascinating is that pointing is surprisingly rare in non-human primates.

Connections to Formal Languages and LLMs

In the modern view of language, two things stand out:

• Language didn’t evolve for communication alone. It co-evolved with social complexity, material culture, symbolic thought, and music.
• Human language is not just words and grammar. It requires signs, symbols, social cognition, and the ability to manipulate and understand these elements in complex ways.

What about formal languages and language models?

Formal languages are just words and rules. We covered them earlier.

LLMs, though, feel like something more:

• They are trained on the output of evolution, namely, the languages we have today
• They learn statistical regularities of symbols, grammar, and various patterns from data alone, hinting that language is not innate, but learned
• They lack embodiment, grounding, joint attention, continuous world experience, but it is pretty amazing how far you can get without those things

When we study LLMs later, two key questions we’ll encounter are:

• Do LLMs truly understand language, or are they merely manipulating symbols without grounding?
• Words (tokens, actually) are just vectors. Is the vector actually the meaning? Is this how humans attach meanings to words?

The Theoretical Landscape

Theories are helpful! There are many, many “Theories of Language” out there.

Theories are organized bodies of knowledge with explanatory and predictive powers. Language theories aim to account for observed linguistic phenomena.

Before the modern scientific era, theories were often philosophical, speculative, and presented in rather fun dialogs or debates. The most famous of these is Plato’s Cratylus dialogue. Cratylus argues that words have a true origin and Hermogenes argues words are purely conventional (nomos). Socrates comes along, finds they are both somewhat right and somewhat wrong.

Here are some theories, roughly grouped. Please understand that none of these theories have to be 100% true or 100% debunked.

• Theories based on generative and formal grammars
  • Transformational-Generative Grammar (Chomsky): syntax generated by phrase-structure rules plus transformations, from an innate universal grammar.
  • Government and Binding Theory: later Chomskyan model organizing syntax into modular subtheories.
  • Minimalist Program: reduces grammar to a minimal operation set (chiefly Merge) driven by economy principles.
  • Generalized Phrase Structure Grammar (GPSG): non-transformational, context-free-based alternative.
  • Head-Driven Phrase Structure Grammar (HPSG): constraint-based, lexicalist grammar using feature structures.
  • Lexical Functional Grammar (LFG): parallel constituent-structure and function-structure representations.
  • Categorial Grammar: assigns functional types to words so syntax mirrors semantic composition.
  • Combinatory Categorial Grammar (CCG): categorical grammar extended with combinators.
  • Tree-Adjoining Grammar (TAG): builds sentences via substitution and adjunction of elementary trees.
  • Dependency Grammar: sentence structure as directed dependency links between words.
  • Optimality Theory: grammar as resolution of ranked, violable constraints.
  • Relational Grammar: grammatical relations (subject, object) as syntactic primitives.
  • Relational Network Theory (formerly Stratificational Linguistics): language as interconnected structural "strata."
  • Glossematics: structuralist theory of language as pure formal relations.
  • Tagmemics: grammatical units analyzed as slot-and-filler "tagmemes."
• Functional, cognitive, and usage-based theories
  • Systemic Functional Linguistics: language as a network of meaning-making choices in social context.
  • Cognitive Grammar: grammar as inherently meaningful, grounded in general cognition.
  • Construction Grammar: form-meaning pairings ("constructions") at every level.
  • Functional Discourse Grammar: models production top-down from intention to phonology.
  • Role and Reference Grammar: links syntax to semantic roles and discourse-pragmatics.
  • Prague Linguistic Circle Functionalism: early structuralist-functionalist emphasis on communicative function.
  • Usage-Based Models of Language: grammar built bottom-up from constructions encountered in use.
• Philosophy of language and meaning
  • Ideational Theory of Meaning: words get meaning by standing for mental ideas.
  • Truth-Conditional Semantics: sentence meaning is its truth conditions.
  • Use Theory of Meaning: meaning is constituted by use within a "language-game."
  • Private Language Argument: a language private to one speaker's sensations is incoherent, since meaning is public and rule-governed.
  • Speech Act Theory: utterances as conventional actions, not just statements.
  • Cooperative Principle: conversational meaning via implicature guided by shared maxims (quality, quantity, relation, manner).
  • Relevance Theory: pragmatics guided by a built-in expectation of relevance.
  • Montague Grammar: formal logic and model-theoretic semantics applied to natural language.
  • Descriptivist Theory of Names: names refer via an associated descriptive content.
  • Causal Theory of Reference: names refer via a causal chain traced back to an original "baptism."
  • Prototype Theory: categories organized around best examples ("prototypes") with graded, fuzzy-boundary membership rather than necessary/sufficient conditions: on your course page, and noted there as strikingly compatible with how embeddings capture graded similarity.
  • Symbol Grounding Problem: symbols in a closed formal system (like a dictionary defining words with more words) refer only to other symbols unless grounded in perception and action: the sharpest old philosophy-of-mind question for evaluating whether LLMs "understand" anything.
  • Semiotics: Peirce’s triadic sign (icon/index/symbol) theory: icons resemble their referent, indices are causally connected to it, symbols are purely conventional.
  • Semiotic Arbitrariness (Saussure): the sign's form-meaning link is fundamentally conventional; treated as foundational, with onomatopoeia/compounds as marginal exceptions.
  • Iconicity (post-Saussure): documents far more pervasive motivated form-meaning links (ideophones, sound symbolism, the bouba/kiki effect) than "marginal exception" suggests: Mithen's position in The Language Puzzle.
  • Linguistic Relativity (aka Sapir-Whorf Hypothesis): language shapes thought; strong version (linguistic determinism) is largely discredited, weak version (habitual-thought influence) has empirical support (e.g., color-term effects on discrimination speed).
  • Logical Positivism: a sentence is meaningful only if empirically verifiable or analytically true.
  • Picture Theory of Meaning: meaningful propositions "picture" possible states of affairs.
  • Deconstruction: meaning is inherently unstable, endlessly deferred through chains of difference.
• Mathematical and computational theories
  • Operator Grammar: language built from elementary operator-argument relations carrying information.
  • Chomsky Hierarchy: classifies formal grammars by generative power.
  • Distributional Hypothesis: (“You shall know a word by the company it keeps”) a word's meaning is captured by its contexts: the direct ancestor of, and arguably vindicated by, word embeddings.
  • Iterated Learning Model: computational/experimental model showing compositional structure emerges from repeated transmission through a learning bottleneck, without needing innate specification—one of the strongest formal cases against requiring innateness.
• Theories of language origin and evolution
  • Universal Grammar — innate, biologically given grammar shared by all humans.
  • Gestural Theory: language originated in gesture before vocal speech.
  • Hmmmmm Hypothesis: protolanguage was Holistic, manipulative, multi-modal, musical, and mimetic, later fracturing into language and music; explicitly argues against a gesture-first origin.
  • Bow-wow, Ding-dong, Yo-he-ho, etc.: 19th-century speculative cluster deriving language from sound-imitation, emotional cries, or work rhythms.
  • Hockett's Design Features: 1960s framework of 16 features (9 shared with other primates, 7 uniquely human) meant to characterize language
  • The Language Instinct: language as a Darwinian biological instinct shaped by natural selection for communication, analogous to echolocation in bats; criticized as too UG-sympathetic and for underweighting cultural/social factors and the pace of language change.
• Psycholinguistics and acquisition theories
  • Poverty of the Stimulus: children acquire grammar from too little input for it to be learned by general reasoning alone.
  • Principles and Parameters: innate fixed principles plus a small set of parameters set by exposure.
  • Behaviorist Theory: language learned through imitation, reinforcement, conditioning.
  • Critical Period Hypothesis: a biologically limited early window for native language acquisition.
  • Cognitive/Constructivist Theory: language as one expression of general cognitive development.
  • Cultural-historical psychology: language acquired through social interaction and scaffolded communication.
  • Connectionism: language acquisition as pattern-learning in neural-network-like associative systems.
  • Bootstrapping Theories: using one available layer of information (meaning, syntax, prosody) as a foothold into another.
• Writing systems
  • Grammatology: general science of writing systems, tracing pictographic $\longrightarrow$ logographic $\longrightarrow$ syllabic $\longrightarrow$ alphabetic evolution.
  • Rebus Principle: using a symbol for its sound rather than its meaning enables abstract representation.
  • Glottographic vs. Semasiographic Writing: writing tied to a specific language's sounds vs. writing that conveys meaning independent of any language.
  • Visible Speech Hypothesis: all mature writing is fundamentally phonetic at its core.
  • Derrida's Critique of Phonocentrism: Western thought wrongly privileges speech over writing as primary.
• Sign language
  • Stokoe's Structural Analysis of Sign Language: demonstrated ASL has its own phonology, analyzable into meaningless sub-lexical parameters.
  • Modality Effects Theory: distinguishes universal properties of language from artifacts of the vocal-auditory modality, using sign languages as the test case.
  • Iconicity in Sign Language Research: documents extensive iconic motivation in sign lexicon and grammar.
  • Bimodal Bilingualism Theory: studies simultaneous spoken and signed language acquisition.

The LLM era is having quite an impact on the many of the theories.

• Some research shows LLMs can acquire certain linguistic competencies previously thought to require innate knowledge, being another argument against the Poverty of the Stimulus. Some have pushed back saying LLMs don’t it effortlessly like humans, claiming a degree of innateness may still exist.
• The success of word embeddings and LLMs has led to a resurgence of interest in distributional and predictive theories of language, which had been somewhat overshadowed by generative grammar approaches. This paper found next-word-prediction skill, brain-encoding performance, and behavioral scores are pairwise correlated across dozens of models, possibly supporting predictive-processing accounts of comprehension. Others argue this might go too far and that prediction is only one way to discover useful linguistic features.

There appear to be at least two new theories (or at least research directions) emerging in the study of language in the context of LLMs.

• Formal vs. Functional Linguistic Competence: evaluation of LLMs' abilities in light of the brain's separation of knowledge about language (formal) from everyday use (functional); LLMs excel at the former but show spotty performance on the latter.
• Stochastic Parrots: theory that LLMs manipulate form without grounding in meaning or communicative intent, which is problematic.

Recall Practice

Here are some questions useful for your spaced repetition learning. Many of the answers are not found on this page. Some will have popped up in lecture. Others will require you to do your own research.

1. What are three views of language?
   Language as mathematics (formal language) • Language as an evolutionary communication system (human language) • Language as statistical patterns (LLMs)
2. What are some of the disciplines that contribute to the study of language?
   Linguistics, anthropology, archaeology, psychology, neuroscience, genetics, computer science, philosophy, ethology, and more.
3. What is Mithen’s central claim about the nature of human language?
   Language is not one thing, but rather a puzzle made of many pieces (words, grammar, symbols, sounds, social cognition, vocal anatomy, brain structure) that evolution assembled, imperfectly, over millions of years.
4. What are three popular theories of language that have been seriously challenged and are considered somewhat incompatible with modern research?
   Hockett’s design features, Chomsky’s Universal Grammar, and Pinker’s language instinct.
5. What is the Hmmmmm hypothesis?
   A theory of language evolution that posits a proto-language that was Holistic, Manipulative, Multi-modal, Musical, and Mimetic.
6. When did hominins start moving from trees to open grasslands and why?
   Around 4 million years ago, due to global climate change leading to more arid conditions.
7. What major physical evolutionary changes likely enabled the development of human language?
   Changes in brain size and structure, and especially the vocal tract.
8. Once the capacity for generating and using words emerged, how did some words end up changing?
   Mishearings, mispronunciations, misunderstandings, social conventions, and a drive toward efficiency.
9. What specific changes in brain size and structure enabled greater cognitive fluidity in our ancestors?
   A larger cerebellum, smaller occipital lobes, a more globular shape, enhanced connectivity between different brain regions through larger neural networks.
10. What is an utterance?
    A unit of language made up of words that is spoken, signed, or written, and has meaning influenced by prosody and context.
11. What are some types of prosody?
    Intonation, stress, rhythm, tempo, and pauses in spoken language; body posture, movements, pauses, and facial expressions in signed languages.
12. What is morphology?
    The study of the structure and formation of words, including the analysis of morphemes, which are the smallest units of meaning.
13. What is phonology?
    The study of the sound systems of languages, including the rules for combining and using phonemes, which are the smallest units of sound that can distinguish meaning.
14. What is the difference between syntax and semantics?
    Syntax deals with structure and semantics deals with meaning.
15. What is the difference between iconic and arbitrary words?
    Iconic words have a form that resembles their meaning (e.g., onomatopoeic words like "buzz"), while arbitrary words have a form that does not have a direct resemblance to their meaning (e.g., "dog").
16. What is really annoying about the use of the Latin Alphabet to write English?
    The Latin Alphabet has far too few letters to match English phonemes, leading to many irregularities and inconsistencies in spelling and pronunciation.
17. What is the origin of the sentence “The rat the cat the dog the cow tossed worried killed ate the malt”?
    It’s constructed from one of the stanzas of the poem This is the house that Jack built.
18. What is Chomsky’s poverty of the stimulus argument and why is it considered less convincing today?
    The argument is that since children acquire grammatical rules they have never directly experienced (an impoverished stimulus), we must have an innate Language Acquisition Device. Modern research has shown that general learning mechanisms and social interaction may be sufficient to explain language acquisition.
19. What do we now believe grammar emerges from?
    The interaction of general learning mechanisms, social interaction, and cultural transmission.
20. In seven words or less, what is pragmatics?
    The study of how context influences meaning.
21. What are Grice’s four maxims of cooperative conversation?
    Quantity (say enough, not too much), Quality (be truthful), Relation (be relevant), Manner (be clear).
22. What is deixis? Give three examples.
    Deixis refers to words whose reference shifts with context. Examples: I, here, now.
23. Approximately how many human languages are in use today?
    ~7,000
24. What kinds of languages stand as arguments against universal grammar?
    Languages with an apparent lack of recursion, languages with an absence of numbers and counting, ergative-absolutive languages, topic-prominent languages, time without tense, tonal languages, sign languages.
25. What is meant by linguistic relativity?
    The idea that language influences worldview or cognition.
26. What is Peirce’s distinction between icons, indices, and symbols? Give an example of each.
    Icons resemble their referent (a photo, a map). Indices are causally connected to their referent (smoke → fire). Symbols are arbitrary and conventional (the word "dog"; red meaning stop).
27. What is joint attention?
    The ability of two individuals to attend to the same thing simultaneously.
28. How does human language differ from formal languages?
    Human language is an evolved communication system with words, grammar, symbols, social cognition, and more, while formal languages are artificial systems of symbols and rules only.
29. How are language capabilities of LLMs not like human language?
    LLMs (currently, arguably) lack true understanding, intentionality, and grounding in the physical and social world.

Summary