1장
The Innate Symphony: How Language Reveals Our Humanity
Have you ever wondered why a three-year-old can master complex grammar that eludes the most sophisticated computer programs? Or why a baby born in Tokyo can effortlessly learn Japanese while her identical twin raised in Stockholm becomes fluent in Swedish? Steven Pinker's groundbreaking book "The Language Instinct" revolutionized our understanding of humanity's most remarkable ability. When published in 1994, it became an instant classic, praised by luminaries from Richard Dawkins to Oliver Sacks, who called it "an exhilaratingly brilliant book." Even today, Bill Gates lists it among his all-time favorites. Beyond academic circles, the book's influence reached popular culture, inspiring everything from linguistic references in films like "Arrival" to deeper public conversations about human nature. What makes this exploration so compelling is Pinker's central argument: language isn't something we learn like chess or mathematics-it's a biological adaptation wired into our brains, as intrinsic to human nature as spinning webs is to spiders.
2장
The Biological Foundations of Human Language
Imagine exploring the highlands of New Guinea in 1930, as Australian prospector Michael Leahy did, and discovering a million Stone Age people completely isolated from modern civilization for 40,000 years. What would strike you most about these people? Despite their isolation, they possessed fully developed languages-complex systems capable of expressing abstract concepts and sophisticated reasoning. This universal presence of language across all human societies provides our first clue that language is not merely a cultural invention but something deeper.
Every human group ever encountered already possessed language-no mute tribe has ever been discovered. These languages aren't primitive jabber but sophisticated systems. Even Stone Age societies have incredibly complex languages like Kivunjo, a Bantu language with seven verbal affixes, fourteen tenses, and sixteen genders. As linguist Edward Sapir noted, "When it comes to linguistic form, Plato walks with the Macedonian swineherd."
The belief that working-class or minority dialects are grammatically deficient is a pernicious myth. Black English Vernacular, for instance, follows systematic grammatical rules as complex as Standard English-it's simply a different dialect with its own consistent patterns. William Labov's research showed that working-class speech actually contained higher percentages of grammatical sentences than middle-class speech, with academic conferences producing the most ungrammatical utterances.
The most compelling evidence for language as an instinct comes from how children acquire it. When speakers of different languages must communicate without learning each other's tongues, they develop pidgins-choppy, grammatically limited jargons. But Derek Bickerton discovered something remarkable: when children are exposed to pidgins during their language-learning years, they spontaneously transform these rudimentary systems into complete languages called creoles. This happened in Hawaiian sugar plantations, where immigrant workers spoke simplified pidgin while their children developed a grammatically complex Hawaiian Creole with consistent rules for marking tense, relationships, and complex structures.
Even more striking evidence comes from sign languages. In Nicaragua, when deaf children were first brought together in schools in 1979, they spontaneously developed their own sign language. The older children created a basic pidgin (LSN), but younger children transformed it into a grammatically sophisticated language (ISN) with standardized rules and inflections. Similarly, a deaf boy named Simon acquired perfect American Sign Language despite his parents' imperfect, pidgin-like signing. Without explicit instruction, Simon extracted consistent patterns from inconsistent input, reorganized them logically, and created grammatical structures his parents never used.
The final evidence for language as instinct comes from neurology. Broca's aphasia, resulting from damage to the left frontal lobe, specifically impairs grammar while leaving other cognitive functions intact. Further evidence comes from Specific Language Impairment (SLI), which runs in families and appears to have genetic origins. The most compelling evidence comes from linguistic savants like Denyse, a severely intellectually disabled girl who speaks fluently and grammatically despite being unable to read, write, or handle money. This double dissociation-language without intelligence and intelligence without language-proves language isn't merely a function of general intelligence but a specialized instinct.
3장
The Language of Thought: Beyond Words
George Orwell's dystopian novel "Nineteen Eighty-four" envisioned Newspeak, a language designed to make unorthodox thoughts impossible by eliminating words for forbidden concepts. But Orwell himself equivocated on whether thoughts depend on words or exist in some "language of thought" merely clothed in words for communication. This fundamental question-do we think in English, Cherokee, or some brain-based "mentalese"-lies at the heart of understanding the language instinct.
The idea that language determines thought (linguistic determinism) pervades our culture. The Sapir-Whorf hypothesis suggests our thoughts are constrained by our language's categories. But this conventional absurdity defies common sense-we've all experienced searching for words to match our thoughts, remembering the gist rather than exact wording, and recognizing when our words don't match our intended meaning. If thoughts depended on words, how could new words be coined or children learn language?
Color perception provides a compelling test case for linguistic determinism. Languages differ in their color vocabularies-Latin lacks generic "gray" and "brown," Navajo collapses blue and green into one word. But physiologists know our retinas contain three types of cones wired to respond to specific color contrasts. Studies show humans worldwide (plus babies and monkeys) perceive colors similarly regardless of language. When Eleanor Rosch tested the Dani people, who have only black and white terms, they still learned categories based on primary colors more easily than off-shades-proving perception shapes language, not vice versa.
Whorf famously claimed the Hopi language had "no words, grammatical forms, constructions, or expressions that refer directly to time" and that Hopi speakers lacked our conception of time as a flowing continuum. Yet anthropologist Ekkehart Malotki documented Hopi sentences rich with temporal references, including tenses, time units, dating methods, and time-related vocabulary. The Hopi actually possess sophisticated timekeeping tools including sun calendars and ceremonial day sequences.
Contrary to popular belief, Eskimos don't have hundreds of words for snow. The myth grew like an urban legend-Franz Boas casually mentioned four unrelated word roots, Whorf embellished it to seven, and subsequent retellings inflated the numbers astronomically. Counting generously, experts find about a dozen Eskimo snow terms, comparable to English words like snow, sleet, slush, blizzard, and powder.
Several lines of evidence confirm that thought exists independently of language. Languageless adults like Ildefonso, a Mexican immigrant studied by Susan Schaller, demonstrate abstract thinking through abilities with numbers, locks, money, and pantomimed narratives. Babies as young as five months can perform simple mental arithmetic, tracking hidden objects being added or removed behind screens. And many creative people-from poets like Coleridge to scientists like Einstein, Faraday, and Tesla-report that their most profound insights come not through words but through mental images.
To understand how thinking can occur without language, we must grasp the computational theory of mind. Alan Turing made the idea of mental representation scientifically respectable by showing how reasoning could be implemented through physical symbol systems. In this view, reasoning involves representations-physical patterns whose elements correspond to concepts and whose arrangements represent relationships between concepts. A processor with fixed reflexes manipulates these representations according to consistent rules, producing new representations that correspond to logical conclusions.
Natural languages like English are fundamentally unsuited to serve as our internal computational medium for several reasons. Words are ambiguous (as newspaper headlines like "Stud Tires Out" demonstrate), while our thoughts about these concepts aren't ambiguous. English lacks logical explicitness-when processing statements like "Ralph is an elephant" and "Elephants have tusks," we understand Ralph has his own tusks, distinctions not explicit in the language itself. Mentalese must be both richer than spoken language (with symbols for disambiguating concepts) and simpler (lacking pronunciation rules or conversation-specific words). Ultimately, knowing a language means knowing how to translate between mentalese and words.
4장
The Architecture of Language: Grammar as Mental Software
Language's essential function is conveying news-telling us who did what to whom. This remarkable ability relies on two fundamental principles. First is Saussure's "arbitrariness of the sign"-the conventional pairing of sounds with meanings (like "dog" having no dog-like qualities yet meaning "dog"). Second is what Humboldt called making "infinite use of finite media"-using generative grammar to combine words in different orders to create distinct meanings, like distinguishing "Dog bites man" from "Man bites dog."
Grammar functions as a discrete combinatorial system where finite elements (words) combine to create larger structures with distinct properties. This system creates a crucial distinction between grammaticality and meaning-we recognize ungrammatical sentences despite their interpretability, showing grammar operates as a fixed code independent of meaning. Conversely, grammatical sentences can make no sense (like Chomsky's "Colorless green ideas sleep furiously") while still being recognized as well-formed.
A sentence is not a chain but a tree. Words are grouped into phrases like twigs joined to branches, with each phrase given a mental symbol that can connect to larger structures. This hierarchical organization is captured in phrase structure grammar, where rules define how words combine into phrases and phrases into sentences. The tree structure solves the word-chain problem through modularity-like telephone jacks or hose couplers. A symbol like "NP" (noun phrase) acts as a connector that allows components to snap into various positions within larger structures.
Phrase structure determines meaning by showing how words relate to each other. This explains why sentences with identical words can have different meanings, as in Groucho Marx's joke: "I once shot an elephant in my pajamas. How he got into my pajamas I'll never know." Newspaper headlines like "Tonight's program discusses stress, exercise, nutrition, and sex with Dick Cavett" are ambiguous because the words can be joined in different tree structures.
All phrases across all languages share a common structure. In a noun phrase like "the cat in the hat," the noun "cat" is the head that determines the phrase's core meaning and grammatical properties. Similarly, verb phrases are defined by their head verbs. The striking similarity between noun phrases and verb phrases-both having heads, role-players, modifiers, and subjects in the same order-suggests a standard design. This pattern extends to prepositional phrases and adjective phrases too.
These super-rules work not just for English but for all languages, with one modification: removing left-to-right order. Languages like Japanese are mirror images of English-verbs follow objects, postpositions replace prepositions, and question markers appear at the end. But the underlying structure remains consistent. This remarkable discovery suggests that grammar consists of universal principles plus parameters that vary between languages. Chomsky's "principles and parameters" theory proposes that children are born knowing these super-rules and only need to learn whether their language is "head-first" (like English) or "head-last" (like Japanese).
Deep structure (now called "d-structure") addresses a fundamental problem in grammar: verbs have requirements for specific elements, yet sometimes these elements appear in unexpected places. For example, "put" requires a subject, object, and prepositional phrase, but we can say "The car was put in the garage" without violating grammar. The solution is that sentences have two phrase structures: deep structure (where all required elements appear in their expected positions) and surface structure (where we actually hear them). Transformational operations "move" phrases from their deep positions to surface positions, leaving behind inaudible "traces" that maintain the connection to their original roles.
5장
The Lexical Landscape: Words and Their Creation
Though the mental dictionary lacks the glamour of mental grammar, the world of words is equally wondrous. People are infinitely creative with words, not just retrieving them from memory but generating new forms through rules. English morphology may seem simple compared to languages like Kivunjo (where verbs can have half a million forms), but it's rich in "derivational" morphology-creating new words from old ones with suffixes like -able, -ize, or -ness.
The mental dictionary isn't simply a memorized list but involves creative rule application. Even preschoolers who pass the "wug-test" demonstrate they can generate new words using morphological rules like "add -s to form plurals." Words have a delicate anatomy consisting of morphemes that fit together according to specific rules, forming a hierarchical structure similar to syntax. For example, "dogs" consists of the stem "dog" and plural inflection "-s"-a simple but powerful computational operation that recognizes abstract mental symbols rather than specific sounds or meanings.
At the most microscopic level of word structure, roots combine with affixes in less predictable ways. Irregular verbs in modern English are fossils of ancient Proto-Indo-European rules that once systematically replaced vowels to form past tenses, just as we now add -ed. These dead rules left behind patterns like sing-sang, swim-swam, and blow-blew that people memorize individually. Though irregular forms can't be productively applied to new verbs, speakers recognize the patterns and extend them humorously, as in "praught," "wunk," and "thunk."
Irregular forms reveal fascinating patterns in word formation. Paul Kiparsky noticed that compounds can be formed with irregular plurals but not regular ones-we say "mice-infested" but "rat-infested" (not "rats-infested"), "men-bashing" but "gay-bashing" (not "gays-bashing"), and "teethmarks" but not "clawsmarks." This distinction arises from the grammar of irregularity itself. Irregular plurals must be stored as roots in the mental dictionary, making them available for the compounding rule that joins stems together. Regular plurals, however, are assembled on the fly by inflectional rules, too late to be available for compounding.
The average person knows far more words than commonly believed. While Shakespeare used about 15,000 distinct words in his works, psychologists estimate that an average high school graduate knows about 45,000 words-and counting proper names, numbers, and foreign words, the figure approaches 60,000. This means learning roughly ten new words daily since age one, or a new word every ninety waking minutes. Even six-year-olds command about 13,000 words, absorbing a new word every two waking hours.
Words are pure symbols whose relationship between sound and meaning is entirely arbitrary. This arbitrariness means children face Quine's "gavagai" problem-when hearing a new word like "gavagai" as a rabbit scurries by, how do they know it means "rabbit" rather than "scurrying thing," "that particular rabbit," or countless other logical possibilities? Children solve this problem because they're innately designed to carve the world into discrete objects and actions, and expect languages to contain words for kinds of things (nouns) and kinds of actions (verbs).
6장
The Miracle of Speech Perception
Language perception is like a sixth sense, allowing us to extract linguistic meaning from sound waves in ways that transcend ordinary hearing. We hallucinate word boundaries that don't physically exist in the continuous sound stream. Even the phonemes we think we hear are illusions-cutting up recordings of words doesn't yield discrete sounds corresponding to letters. Speech perception is a biological miracle that allows humans to process 10-15 phonemes per second in casual speech, up to 40-50 per second for sped-up speech, far exceeding what the human auditory system should theoretically be able to distinguish.
Speech production begins in the lungs, where we override our normal breathing patterns to time exhalations with our intended phrases. Air travels through the trachea to the larynx, where the vocal folds can either close completely or vibrate to produce "voicing." The sound passes through various chambers-the throat, mouth, and potentially the nose-each with particular resonances. The tongue, "the most important of the speech organs," functions as three organs in one and shapes these resonances. Speech sounds aren't single gestures but combinations executed simultaneously, with each phoneme defined by selecting one of six speech organs, choosing a manner of articulation, and configuring other speech organs.
Phonological rules operate on features, not phonemes. When we pronounce words like "writing" versus "riding," the flapping rule makes the t/d sound identical, but the vowels remain different because the vowel-change rule applies before flapping. These rules target entire classes of sounds sharing features (like voicing) rather than individual phonemes. For instance, the past-tense suffix "-ed" becomes "t" after voiceless consonants (walked, slapped) and remains "d" after voiced ones (jogged, sobbed). Phonological rules aren't mere sloppiness but systematic adjustments that balance ease of articulation for speakers with clarity for listeners.
Speech recognition is extraordinarily difficult because no two voices are alike in vocal tract shape or articulation habits, and phonemes sound different depending on stress and speaking rate. The biggest challenge is coarticulation-our tongues anticipate upcoming sounds, placing themselves optimally for smooth transitions between phonemes. This creates context-dependent sound variations where each phoneme's acoustic signature blends with its neighbors. In words like "Cape Cod," the two "k" sounds use different tongue positions; in "horseshoe," the first "s" becomes "sh." These subtle adjustments wreak havoc on the speech signal, making it impossible to isolate clean phoneme boundaries.
English spelling isn't as deranged as it appears. While written language isn't instinctive like spoken language-it was invented only a few times in history and must be laboriously taught-it taps into the language system at logical points. All writing systems encode either morphemes, syllables, or phonemes, but never actual sound units identifiable on spectrograms. English spelling corresponds not to sounds but to the phonemes in our mental dictionary. For 84% of English words, spelling follows regular rules. Even when letters get different pronunciations across related words (like "electric-electricity" or "nation-national"), they serve to identify words as sharing the same root morpheme.
7장
The Computational Challenge of Understanding
Human sentence comprehension is remarkably fast and powerful, working in real time with only a half-second lag between speaker and listener. Understanding how this process works has practical applications for writing clear prose and interpreting legal language. The first step in comprehension is parsing-unconsciously grouping words into phrases and determining subjects, verbs, and objects. While grammar is a static database of language rules, the parser is the mental program that analyzes sentence structure during comprehension, applying these rules step by step as words arrive.
To understand a sentence like "The dog likes ice cream," the parser processes each word sequentially, building a syntactic tree. It identifies "the" as a determiner, which must be part of a noun phrase. The parser predicts upcoming structures while holding incomplete branches in memory. When "dog" arrives, it completes the noun phrase. As "likes" enters, it's identified as a verb forming part of a verb phrase, which requires an object. Finally, "ice cream" completes the noun phrase object and the sentence structure. When all dangling branches are resolved, we experience the mental "click" of comprehending a complete grammatical sentence.
Human sentence processing faces two computational challenges: memory for tracking dangling phrases and decision-making for resolving ambiguities. Computers excel at memory tasks but struggle with decisions, while humans show the opposite pattern. Our short-term memory bottleneck (limited to about seven items) makes "top-heavy" sentences difficult to process-those where we must hold incomplete phrases in memory too long. This explains why languages offer alternative constructions that allow heavy phrases to be moved to the end of sentences, easing the listener's burden.
While humans struggle with memory limitations, we excel at decision-making during parsing-something computers find challenging. Most words have multiple possible interpretations, creating "local ambiguities" at nearly every step. These ambiguities multiply exponentially when considering how phrases fit together. Computer parsers meticulously track all grammatical possibilities, finding legitimate but absurd interpretations humans would never consider. The famous example "Time flies like an arrow" yielded five different syntactic trees to a computer parser, including the nonsensical interpretation about "time-flies" being fond of arrows.
At the individual word level, our brains briefly entertain multiple meanings of ambiguous words-even unlikely ones. In Swinney's ingenious experiment, people heard sentences containing ambiguous words like "bug" (insect or surveillance device). When tested immediately after hearing "bug" in an insect context, they recognized both "ant" and "spy" faster than unrelated words, showing the brain activates all meanings regardless of context. This priming effect disappeared after just three syllables, explaining why we're unaware of considering inappropriate meanings.
Parsing ambiguities have life-or-death consequences in legal contexts. Insurance contracts, criminal statutes, and jury instructions often hinge on structural ambiguities. Courts resolve these using "canons of construction" that mirror psycholinguistic parsing principles like the Last Antecedent Rule (equivalent to minimal attachment). Despite these linguistic principles, judges often find ways around the most natural interpretations to reach outcomes they feel are just.
Parsing is only the first step in understanding language. Real speech differs from simple sentences because speakers rarely articulate "the truth, the whole truth, and nothing but the truth." Conversation relies on shared background knowledge, efficient references, and organizing information with topics early in sentences (often using passives) and new information at the end. Communication depends on mutual cooperation between speakers and listeners, with tacit expectations that speech will be informative, truthful, relevant, clear, and brief.
8장
The Tower of Babel: Language Diversity and Universals
The biblical story of Babel describes God confounding human language. In 1957, linguist Martin Joos claimed languages could "differ from each other without limit," while Chomsky countered that a Martian would see Earth's 4,000-6,000 languages as essentially one language with different vocabularies. Languages differ from English in various ways: some use case affixes rather than word order; some allow free word order; some have ergative systems where intransitive subjects pattern with transitive objects; some are topic-prominent rather than subject-prominent; some use different basic word orders (SVO, SOV, VSO); and some use classifier systems for nouns.
Despite these differences, Joseph Greenberg's 1963 study of 30 languages from five continents revealed 45 universal patterns in word order, and subsequent surveys have documented hundreds more. Some universals are absolute (no language forms questions by completely reversing word order), while others are statistical (subjects typically precede objects). Many universals are implicational: if a language has feature X, it will also have Y-like the connection between SOV order and postpositions versus SVO order and prepositions.
If language structure is innate and universal, why do we have thousands of mutually unintelligible dialects? Darwin himself noted the parallel between language differentiation and biological evolution-languages, like species, show homologies due to common descent and can be classified into related groups. Linguistic diversity emerges through three processes: variation (linguistic innovation), heredity (learning), and isolation (geographical or social barriers). Linguistic innovation occurs when speakers reanalyze speech they hear, interpreting it as having come from different rules than intended. Changes cascade through language systems: phonological rules arise from reanalyzing rapid speech, morphological rules from reinterpreting pronunciation patterns, and syntactic constructions from preferred word orders becoming obligatory.
English evolved dramatically over centuries, from its origins in Anglo-Saxon to today's global language. What became standard English was simply the London dialect of the seventeenth century. English originated with Germanic tribes (Angles, Saxons, Jutes) who invaded Britain after Rome's withdrawal, completely displacing the Celtic languages. The Norman Conquest in 1066 introduced French influence, creating the distinctive dual vocabulary system of Anglo-Saxon and Latinate words that characterizes modern English. Middle English (1100-1450) saw final syllables reduced or eliminated, forcing fixed word order and grammatical auxiliaries to compensate. The mysterious Great Vowel Shift (1450-1700) revolutionized pronunciation, creating the mismatch between English spelling and pronunciation we struggle with today.
We face an impending linguistic catastrophe. Linguist Michael Krauss estimates that 90% of the world's languages may disappear in the next century-including 80% of North American Indian languages, 90% of Australian languages, and thousands more worldwide. Only about 600 languages with over 100,000 speakers are reasonably safe. Languages die through habitat destruction, genocide, forced assimilation, demographic submersion, and media bombardment-what Krauss calls "cultural nerve gas." The extinction of languages impoverishes our understanding of the language instinct's scope and limits. Each language represents "a supreme achievement of a uniquely human collective genius," containing unique cultural treasures.
9장
The Miracle of Language Acquisition
A tabloid headline from 1985 claimed a newborn baby spoke fluently about heaven immediately after birth. Such stories remain firmly in the realm of fiction because real language development follows a much slower trajectory. Most children don't speak until age one, don't combine words until eighteen months, and don't produce fluent grammatical sentences until two or three years old. This raises profound questions about what's happening during those early years of language development.
Infants enter the world with remarkable linguistic abilities. Experiments using sucking responses show one-month-olds can distinguish phonemes like "ba" and "pa," responding not just to raw sounds but to speech-specific features. These abilities are innate-Kikuyu and Spanish babies can distinguish English phonemes their parents cannot perceive. While babies initially function as universal phoneticians, by ten months they've specialized to their native language's sounds. During the first year, babies' vocal tracts physically transform from a mammal-like configuration to the human arrangement that enables diverse vowel production. After minimal vocalization in the first months, babies begin playful sound exploration at 5-7 months, followed by syllabic babbling at 7-8 months.
Around their first birthday, babies begin understanding and producing words, initially in isolation. These first words show remarkable consistency across cultures: about half name objects (food, toys, animals), while others describe actions, properties, or social routines. Around eighteen months, language development accelerates dramatically-vocabulary grows at a new-word-every-two-hours pace, and children begin combining words syntactically. These two-word combinations already follow proper word order 95% of the time and express universal meanings: appearances, disappearances, properties, actions, and requests. Between ages two and three, language abilities explode exponentially, with sentence complexity doubling monthly, reaching thousands of syntactic types before the third birthday.
By age three, children's sentences become not only longer but structurally more complex, with deeper embedding of constituents. Three-year-olds master an astonishing range of sentence types-questions, relative clauses, comparatives, negations, complements, and passives. Though many sentences contain errors, careful analysis reveals children obey grammatical rules over 90% of the time. Even the notoriously complex English auxiliary system with its billions of possible combinations generates virtually no errors in thousands of children's utterances. Their errors aren't random but follow grammatical logic, like overregularizing past tense ("holded" instead of "held") when memory fails them, or extending causative rules ("Don't giggle me!") in ways that would be grammatical in other languages.
Children's language development requires exposure to human speech, but not all aspects of language input are equally important. Grammar acquisition doesn't depend on practice or parental correction. Children with articulation problems still develop excellent grammatical comprehension. Parents rarely correct grammar, focusing instead on truthfulness, and children ignore corrections when offered. Roger Brown's research showed parents responded identically to grammatical and ungrammatical sentences, providing no useful feedback. This creates a profound learning challenge: children must generalize to infinite sentences without feedback about incorrect generalizations.
10장
The Biological Foundations of Language
Children acquire language efficiently through innate constraints, using mental labels for parts of speech to generate infinite sentences from finite rules. Universal Grammar helps them focus on relevant grammatical features while ignoring irrelevant possibilities, making language acquisition manageable.
Babies aren't born talking due to both developmental sequence and brain development. Human infants are born with incomplete brains due to birth canal constraints - if we gestated like other primates, we'd begin talking at eighteen months. After birth, rapid brain development occurs, with synapses peaking between nine months and two years, and metabolic activity surpassing adult levels around age four.
Language acquisition is a one-shot skill that becomes unnecessary after learning the local language. The brain may dismantle this circuitry to conserve resources, as it consumes significant metabolic energy. Computer simulations demonstrate that a critical period in early childhood is evolutionarily inevitable.
Language is predominantly controlled by the left hemisphere, first discovered by Paul Broca in 1861. Multiple evidence supports this: aphasics often have right-side paralysis, normal people recognize words better in their right visual field, and brain imaging shows language activity primarily in the left hemisphere. This asymmetry exists because language processing benefits from concentration in one hemisphere, as it coordinates sequential information rather than spatial directions.
Brain damage can cause specific language deficits. Pure Word Deafness affects spoken word recognition while preserving other language abilities. Some aphasics struggle with particular word types or sentence structures. The mental lexicon can fragment in specific ways - patients might handle concrete but not abstract nouns, or living versus nonliving things.
Language disorders demonstrate genetic components. Stuttering, dyslexia, and Specific Language Impairment (SLI) run in families, with identical twins showing higher concordance than fraternal twins. The K family, with three generations of SLI, provides strong evidence for genetic transmission, suggesting a single dominant autosomal gene based on its inheritance pattern.
11장
The Evolution of Language: A Uniquely Human Trait
The elephant's trunk-with its 60,000 muscles, remarkable dexterity, and unique existence in nature-serves as a metaphor for human language. Just as elephants alone possess this extraordinary organ, humans alone possess language. And just as biologists don't make a fuss about the trunk's uniqueness, we should perhaps reconsider our approach to language's uniqueness.
Human language differs fundamentally from animal communication systems. Animal communication follows three designs: finite repertories of calls, continuous analog signals, or variations on themes. Human language, by contrast, has a discrete combinatorial system called grammar that makes it infinite, digital, and compositional. Despite popular claims about apes learning language, careful scientific analysis reveals these animals never truly acquired language. Even Kanzi, a pygmy chimp often cited as more advanced, shows only marginally better abilities than common chimps. His "sentences" are merely fixed formulas without internal structure.
The apparent uniqueness of human language poses no paradox for evolution. Critics mistakenly imagine evolution as a ladder with humans at the top, expecting a gradual "fade-in" of language in our closest relatives. But evolution created a bush, not a ladder-we and chimps evolved from a common ancestor, now extinct. Language could have emerged after our lineage split from chimps, with approximately 5-7 million years and roughly 350,000 generations for language to evolve gradually. Brand-new neural circuits could have arisen through genetic changes that duplicate brain maps, reroute connections, and refine internal wiring.
When language first evolved remains uncertain. Archaeologists try to infer language abilities from artifacts, assuming complex tools reflect complex minds capable of language. However, this approach likely underestimates language's antiquity, as many hunter-gatherer technologies made from perishable materials would leave no trace. Modern Homo sapiens emerged around 200,000 years ago with brains like ours, almost certainly possessing language. This contradicts the common textbook date of 30,000 years ago (based on cave art), since all modern human populations have identical language abilities despite diverging much earlier.
While Darwin argued that instincts evolve through natural selection like physical traits, Chomsky has been surprisingly skeptical about applying evolutionary theory to language. He suggests that "little is known" and that answers "may well lie not so much in the theory of natural selection as in molecular biology." Natural selection would be falsified by traits designed purely for beauty, complex organs without useful intermediate forms, organisms not produced by replication, or traits evolved solely to benefit other species.
Language shows unmistakable signs of adaptive complexity with its many precisely structured components-syntax, morphology, lexicon, vocal tract modifications, phonological rules, speech perception, and learning algorithms. These are physically realized as intricate neural circuits that enable the extraordinary ability to communicate infinite thoughts between minds. This complex design bears "the unmistakable stamp of nature's designer, natural selection."
While there are genuine challenges in reconstructing language evolution, they're resolvable. Critics ask who grammar mutants could talk to, but neighbors could decode their speech using general intelligence, just as we can understand phrases like "skid crash hospital" despite their lack of grammar. When grammar mutants make important distinctions that others decode with effort, this creates pressure for listeners to evolve matching systems for reliable parsing. Contrary to critics' claims, intermediate grammar forms are easy to imagine-systems with narrower symbols, less reliable rules, or fewer modules.
Human language is our species' most remarkable achievement-a biological instinct wired into our brains that allows us to communicate an infinite range of thoughts using a finite set of elements. It reveals the extraordinary complexity of human cognition and the power of natural selection to create seemingly miraculous adaptations. By understanding language as an instinct rather than a cultural invention, we gain profound insights into human nature itself-our shared cognitive architecture, our evolutionary history, and the biological foundations of our most distinctly human trait.
