6.1 Language Structure, Psycholinguistics, and Language Acquisition Theories

Language Structure, Psycholinguistics, and Language Acquisition Theories

Language represents one of the most sophisticated, generative cognitive capacities in the human repertoire. Psycholinguistics—the interdisciplinary study of the psychological and neurobiological mechanisms enabling humans to acquire, comprehend, produce, and mentally represent language—examines this faculty across multiple levels of analysis, ranging from microscopic acoustic phonetic boundaries to high-level pragmatic discourse.

1. The Hierarchical Architecture of Language

Human natural language is structured hierarchically. At each successive tier, smaller units are combined according to systematic, rule-governed combinatorial systems:

                         [ Pragmatics ]
                  (Social context, implicature)
                               │
                               ▼
                         [ Semantics ]
                 (Literal meaning of morphemes/words)
                               │
                               ▼
                          [ Syntax ]
                (Grammatical phrase rules & trees)
                               │
                               ▼
                        [ Morphology ]
                (Morphemes: Free & Bound units)
                               │
                               ▼
                        [ Phonology ]
               (Phonemes: Minimal sound contrasts)

Phonology: The Sound System

• Phoneme: The smallest structural unit of sound in a spoken language that distinguishes one word from another (e.g., altering the initial phoneme /b/ to /p/ changes bat into pat). English uses approximately 40 to 44 phonemes, whereas natural languages globally range from roughly 11 to over 140.
• Allophones: Non-contrastive acoustic variations of a single phoneme that do not alter lexical meaning (e.g., the aspirated [pʰ] in pot versus the unaspirated [p] in spot).
• Categorical Perception: The perceptual phenomenon whereby continuous acoustic variations in Voice Onset Time (VOT)—the temporal interval between the release of a plosive vocal-tract closure and the onset of vocal-cord vibration—are perceived as discrete, qualitative phonetic categories. For example, stimuli along a continuum from /ba/ (VOT ~0 ms) to /pa/ (VOT ~60 ms) are not heard as intermediate sounds; listeners report a sharp, sigmoidal perceptual boundary around +25 to +30 ms.
• Infant Phonemic Sensitivity: Peter Eimas and colleagues (1971) employed the high-amplitude sucking (HAS) paradigm to demonstrate that one-month-old human infants exhibit adult-like categorical perception for VOT. Crucially, young infants (0–6 months) function as "universal phoneticians," capable of discriminating phonemic contrasts across all human languages (e.g., the Hindi retroflex vs. dental /d/ or Japanese-non-differentiated English /r/-/l/). Janet Werker and Richard Tees (1984) showed that this universal sensitivity declines between 8 and 12 months due to perceptual narrowing, tuning the infant's phonological inventory specifically to the ambient native tongue.

Morphology: The Units of Meaning

• Morpheme: The smallest unit of language that carries grammatical or semantic meaning. A single word can contain one or multiple morphemes (e.g., dog [1 morpheme], dog-s [2 morphemes], un-break-able [3 morphemes]).
• Free Morphemes: Units that can stand alone as independent words:
  • Lexical (Content) Morphemes: Nouns, verbs, adjectives, adverbs that carry core referential content (run, table, bright).
  • Functional (Grammatical) Morphemes: Articles, conjunctions, prepositions (the, and, in).
• Bound Morphemes: Units that cannot stand alone and must attach as affixes (prefixes, suffixes, infixes) to a root or stem:
  • Derivational Morphemes: Affixes that transform the syntactic grammatical category or fundamentally alter the lexical meaning of the base word (e.g., adding -ness converts the adjective dark into the noun darkness; adding un- converts happy to unhappy).
  • Inflectional Morphemes: Affixes that alter grammatical tense, number, aspect, person, or degree without altering the core part of speech (English possesses exactly eight inflectional affixes: noun plural -s, noun possessive -’s, verb third-person singular -s, past tense -ed, present participle -ing, past participle -en, comparative -er, and superlative -est).
• Overregularization Errors: During morphological acquisition (typically ages 2.5–5), children pass through a characteristic U-shaped developmental trajectory. Initially, they produce irregular past-tense verbs correctly (went, came, held) through rote episodic retrieval. As they induce the general inflectional morphological rule (add -ed for past tense), they overextend it to irregular roots, producing systematic overregularization / overgeneralization errors (goed, comed, foots, holded). Adult-like competence re-emerges later as lexical memory blocks overregular rule application.

Syntax: Combinatorial Grammatical Rules

Syntax governs the lawful arrangements of words into phrases, clauses, and sentences:

• Phrase Structure Rules: Describe the hierarchical constituency of sentences (e.g., $S \rightarrow NP + VP$; $NP \rightarrow (Det) + (Adj) + N$; $VP \rightarrow V + (NP)$). Sentences are not linear chains of probabilistic word transitions (Markov chains), but nested hierarchical tree structures.
• Chomsky's Transformational Generative Grammar: In Syntactic Structures (1957) and Aspects of the Theory of Syntax (1965), Noam Chomsky demonstrated that phrase structure rules alone cannot account for relationships between structurally diverse sentences with identical meaning, or structurally identical sentences with ambiguous meaning:
  • Deep Structure: The abstract, underlying semantic and syntactic representation specifying the foundational grammatical relations between constituents.
  • Surface Structure: The actual linear, phonological, or orthographic sequence of words that is articulated or written.
  • Transformational Rules: Systematic mental operations that map deep structures onto surface structures (e.g., passivization transformations convert the active deep structure "The dog chased the cat" into the surface passive "The cat was chased by the dog").
  • Structural Ambiguity: Occurs when a single surface structure maps onto two distinct deep structures (e.g., "Visiting relatives can be boring" can mean either "It is boring to visit relatives" or "Relatives who are visiting are boring").

Semantics and Pragmatics

• Semantics: The study of literal linguistic meaning at the lexical and sentential levels. Evaluates how truth conditions, sense, reference, and compositional semantic rules combine lexical items into meaningful propositions.
• Pragmatics: The study of language in communicative context, focusing on intended meaning, social conventions, speech acts, and conversational implicature:
  • Speech Act Theory (J.L. Austin & John Searle): Utterances do not merely state facts; they execute actions. An utterance contains a locutionary act (literal utterance), an illocutionary act (intended communicative force, e.g., promising, ordering, warning), and a perlocutionary effect (the actual psychological impact on the listener).
  • Grice's Cooperative Principle (Paul Grice, 1975): Communication succeeds because interlocutors adhere to an overarching Cooperative Principle operationalized through four Conversational Maxims:
    1. Maxim of Quantity: Provide as much information as required; do not provide more or less than is informative.
    2. Maxim of Quality: Do not state what you believe to be false or that for which you lack adequate evidence.
    3. Maxim of Relation (Relevance): Be relevant to the current conversational topic.
    4. Maxim of Manner: Be perspicuous: avoid obscurity, avoid ambiguity, be brief, and be orderly.
  • Conversational Implicature: When a speaker deliberately flouts or violates a maxim, listeners generate an inferential implicature (e.g., responding to "Are you coming to the party tonight?" with "I have an exam at 8:00 AM tomorrow" flouts the maxim of relation to implicitly convey "No").

2. Speech Perception and Auditory-Visual Integration

Speech perception is remarkably rapid, decoding roughly 15 to 25 phonemes per second from a continuous acoustic waveform devoid of physical silence between lexical word boundaries (the segmentation problem). Perceptual processing relies on dynamic interactions between bottom-up acoustic signals and top-down cognitive expectations:

3. Neurobiology of Language: Classical Syndromes and Cortical Networks

The classical neuropsychological model of language is the Wernicke-Geschwind Model, which maps distinct language faculties onto specific perisylvian structures of the left cerebral hemisphere:

               [ Primary Auditory Cortex (A1) ]
                              │
                              ▼
             [ Wernicke's Area (BA 22) ] ──> Auditory Comprehension
                              │
                (Arcuate Fasciculus Tract)
                              │
                              ▼
              [ Broca's Area (BA 44/45) ] ──> Motor Speech Planning
                              │
                              ▼
            [ Primary Motor Cortex (M1) ] ──> Vocal Articulation

Clinical Aphasia Syndromes

• Broca's Aphasia (Expressive / Motor / Non-Fluent Aphasia):
  • Lesion Site: Left posterior inferior frontal gyrus (Brodmann Areas 44 [pars opercularis] and 45 [pars triangularis]).
  • Clinical Profile: Effortful, halting, slow, non-fluent speech output. Severe agrammatism—speech is stripped of functional morphemes and grammatical affixes, resulting in "telegraphic speech" dominated by isolated content nouns and verbs (e.g., "Son... school... hospital... bad"). Dysprosody and articulatory groping (apraxia of speech) are prominent.
  • Comprehension: Preserved for simple canonical subject-verb-object structures, but severely compromised when understanding depends exclusively on complex syntax or passive transformations (e.g., failing to identify the agent in "The horse was kicked by the cow").
  • Metacognition: Patients possess acute insight into their expressive deficit, resulting in profound frustration, depression, and catastrophic emotional reactions.
• Wernicke's Aphasia (Receptive / Sensory / Fluent Aphasia):
  • Lesion Site: Left posterior superior temporal gyrus (Brodmann Area 22).
  • Clinical Profile: Fluent, effortless, rapid speech exhibiting normal melodic intonation, inflection, and syntactic prosody. However, speech is devoid of substantive meaning (word salad or paragrammatism).
  • Linguistic Errors: Frequent paraphasias (phonemic: "papple" for apple; semantic: "fork" for spoon) and bizarre, uninterpretable neologisms ("stremflig").
  • Comprehension: Severely devastated for both spoken and written language.
  • Metacognition: Patients exhibit anosognosia—they are characteristically unaware of their communicative deficits, remaining oblivious to their unintelligible output.
• Conduction Aphasia:
  • Lesion Site: The arcuate fasciculus, the subcortical white-matter association pathway connecting Wernicke's and Broca's areas, often involving the supramarginal gyrus (Brodmann Area 40).
  • Clinical Profile: Both spontaneous speech production and auditory comprehension remain intact. The hallmark pathognomonic symptom is disproportionately severe impairment in verbal repetition (e.g., repeating "No ifs, ands, or buts").
  • Behavior: Patients make repeated, unsuccessful approximations to reach a target word (conduite d'approche), demonstrating intact internal target representation alongside defective sensorimotor transmittal.
• Transcortical Aphasias (Preserved Repetition):
  • Transcortical Motor Aphasia: Non-fluent, halting output resembling Broca's aphasia, but verbal repetition is remarkably preserved and fluent. Caused by lesions in prefrontal cortex anterior/superior to Broca's area, isolating Broca's from motor initiation.
  • Transcortical Sensory Aphasia: Fluent, empty speech and poor comprehension resembling Wernicke's aphasia, but verbal repetition is preserved, often manifesting as involuntary echoing (echolalia). Caused by lesions in the parieto-occipital border zone isolating Wernicke's area from conceptual cortical networks.
  • Mixed Transcortical Aphasia (Isolation of Speech Area): Non-fluent speech and zero comprehension, yet echolalic repetition remains intact.
• Global Aphasia:
  • Lesion Site: Massive infarction encompassing the entire left perisylvian cortex (middle cerebral artery occlusion).
  • Clinical Profile: Complete devastation of all expressive, receptive, repetition, reading, and writing faculties.
• Angular Gyrus Syndromes:
  • Lesion Site: Left angular gyrus (Brodmann Area 39), situated at the parietal-temporal-occipital boundary.
  • Deficits: Mediates visual-orthographic to phonological-auditory cross-modal translation. Damage produces alexia (inability to read) and agraphia (inability to write). Damage within this locus can also trigger Gerstmann's Syndrome, characterized by the classic tetrad: finger agnosia, right-left disorientation, agraphia, and acalculia.

4. Theories of Language Acquisition

The ontogeny of language has served as the battleground for classic debates between extreme empiricist/behaviorist positions and nativist/biological modularity paradigms:

Behaviorist / Empiricist       Interactionist / Cognitive       Nativist / Innatist
(B.F. Skinner - 1957)          (Bruner, Saffran, Tomasello)     (Noam Chomsky - 1959)
────────────────────────       ──────────────────────────       ─────────────────────
- Operant conditioning         - LASS scaffolding               - Poverty of the stimulus
- Imitation & shaping          - Child-directed speech          - Universal Grammar (UG)
- Mand, tact, echoic units     - Statistical learning           - Innate LAD module
- Reinforcement schedules      - Joint attention                - Sensitive/Critical period

Skinner's Operant Account

In Verbal Behavior (1957), B.F. Skinner conceptualized language as overt verbal behavior acquired entirely via general-purpose operant conditioning, associative learning, imitation, and differential reinforcement from the linguistic environment:

• Operant Operands: Spoken utterances are shaped into functional response classes: mands (demands/requests reinforced by characteristic consequences, e.g., saying "milk" to receive milk), tacts (verbal labels emitted in response to non-verbal discriminative environmental stimuli, e.g., saying "dog" upon seeing a canine), and echoics (imitative vocal behaviors).
• Mechanisms: Parents reinforce phonological approximations and shape grammatical combinations through praise and compliance, while unreinforced or punished erroneous constructions extinguish.

Chomsky's Nativist Revolution and Critique of Skinner

In his landmark 1959 review of Skinner's Verbal Behavior, Noam Chomsky dismantled the behaviorist account, demonstrating that operant conditioning is mathematically and psychologically inadequate to explain human language acquisition:

1. Poverty of the Stimulus: The ambient verbal input infants receive is finite, noisy, degenerate (full of false starts, slips of the tongue, and fragmented sentences), and essentially devoid of systematic negative evidence (parents routinely correct factual inaccuracies rather than grammatical errors, e.g., approving "Doggie runned away" if true, but correcting "That's a nice horse" if the animal is a cow). Yet every neurologically intact child effortlessly induces the complete, infinite generative grammar of their language within a few short years.
2. Generativity (Productivity): Human language is endlessly creative. Speakers continuously produce and comprehend novel sentences they have never previously uttered, heard, or had reinforced.
3. Language Acquisition Device (LAD): Chomsky posited an innate, biologically pre-wired neurocognitive mechanism specifically dedicated to language acquisition.
4. Universal Grammar (UG): An innate, species-specific body of structural principles (e.g., structure dependence, phrase-structural hierarchy) and parameters (e.g., the head-directionality parameter, pro-drop / null-subject parameter) common to all human natural languages. Language learning consists of setting parameters based on exposure to minimal environmental triggers, not learning grammar de novo.

The Critical / Sensitive Period Hypothesis

Formulated by neurologist Eric Lenneberg in Biological Foundations of Language (1967), the Critical Period Hypothesis asserts that language acquisition must occur within a biologically delimited maturation window extending from infancy until puberty. Beyond this window, the loss of neural plasticity and progressive cerebral lateralization permanently impede native-level grammatical fluency:

• Wild / Isolated Children: The case of Genie, a girl subjected to severe environmental deprivation, isolation, and abuse from 20 months until age 13.5. Following intensive intervention, Genie acquired substantial lexical vocabulary, semantic comprehension, and pragmatic desire to communicate; however, she was permanently incapable of acquiring syntax, functional morphemes, or transformational grammar, speaking only in rudimentary telegraphic fragments. Dichotic-listening and evoked-potential testing suggested that her language processing relied on the right hemisphere, corroborating Lenneberg's thesis that the left hemisphere's syntactic machinery atrophies without timely maturational input.
• Deaf Sign Language Acquisition: Elissa Newport and colleagues evaluated deaf adults exposed to American Sign Language (ASL) at varying ages (native from birth, early childhood [4–7], or late [12+]). Despite decades of continuous daily ASL usage, late learners exhibited persistent, intractable deficits in complex morphology and syntactic production, proving that maturational timing, rather than total duration of practice, dictates linguistic mastery.
• Second Language Acquisition: Studies by Jacqueline Johnson and Elissa Newport (1989) demonstrated a linear decline in ultimate grammatical attainment for immigrants arriving between ages 3 and 17, flattening into a low, variable plateau for post-pubescent adult arrivals.

Interactionist and Statistical Learning Approaches

Modern developmental psycholinguistics integrates biological readiness with social and cognitive computational learning:

• Jerome Bruner's Language Acquisition Support System (LASS): Posits that the biological LAD is matched by an environmental social scaffold. Caregivers provide child-directed speech (CDS) (historically termed motherese—characterized by higher fundamental pitch, exaggerated prosodic contours, slowed speech rate, hyper-articulated vowels, and simplified sentence structures), establish joint attention, and engage in structured conversational turn-taking (scaffolding).
• Statistical Learning in Infancy: Jenny Saffran, Richard Aslin, and Elissa Newport (1996) presented 8-month-old infants with an unbroken continuous auditory stream of synthetic speech consisting of four three-syllable nonsense words (bidaku, padoti, golabu, tupiro) presented in random sequence (bidakupadotigolabu...) without pauses, stress, or intonational cues. The only cue to word boundaries was transitional probability ($P(Y|X)$): within words, syllable transitions were 1.0 (e.g., bi was always followed by da), whereas across word boundaries, transitional probabilities dropped to 0.33. Using the head-turn preference procedure, they demonstrated that infants rapidly computed these statistical probabilities, parsing novel words after merely two minutes of passive exposure.

5. Linguistic Relativity vs. Linguistic Determinism

The proposition that language fundamentally influences or constrains cognitive processes is known as the Sapir-Whorf Hypothesis (formulated by Edward Sapir and his student Benjamin Lee Whorf):

• Strong Version (Linguistic Determinism): Language determines thought. Cognitive categories and perceptual capacities are strictly constrained by the formal structure and vocabulary of one's native language; concepts absent from a linguistic lexicon cannot be conceptualized by its speakers. Status: Broadly rejected by modern cognitive science. Pre-linguistic infants and non-human animals engage in sophisticated categorization, concept formation, and reasoning without language.
• Weak Version (Linguistic Relativity): Linguistic categories and grammatical distinctions influence, bias, or facilitate non-linguistic cognitive tasks, such as memory recall, perceptual categorization speed, and spatial reasoning. Status: Supported by a growing body of rigorous cross-cultural psycholinguistic research.

Empirical Investigations of Linguistic Relativity

                               [ Cross-Cultural Color Categorization ]
                                                  │
                  ┌───────────────────────────────┴───────────────────────────────┐
                  ▼                                                               ▼
[ Eleanor Rosch / Heider (1972) ]                               [ Winawer et al. (2007) ]
- Dani tribe (Papua New Guinea)                                 - English vs Russian speakers
- Only two color terms: 'mili' (dark/cool)                      - Russian has distinct mandatory basic
  and 'mola' (bright/warm)                                        words for light blue ('goluboy')
- Result: Matched English speakers on focal                       and dark blue ('siniy')
  color memory; supported universal physiology                  - Result: Russian speakers faster at cross-boundary
  of visual trichromacy over linguistic determinism               discriminations; effect vanishes under verbal interference

• Cross-Cultural Color Categorization:
  • Universalist Findings (Eleanor Rosch / Heider, 1972): Rosch tested the Dani people of Papua New Guinea, whose language possesses only two basic color terms: mili (dark/cold shades) and mola (bright/warm shades). Despite lacking terms for primary spectral hues, Dani participants discriminated, recognized, and remembered universal "focal colors" (prototypical red, yellow, green, blue) just as accurately as English speakers possessing eleven basic color categories, demonstrating biological, physiological constraints on color perception (Brent Berlin & Paul Kay, 1969).
  • Relativist Nuance (Winawer, Witthoft, Frank, Wu, & Boroditsky, 2007): Russian obligatorily distinguishes light blue (goluboy) from dark blue (siniy) as distinct basic colors. Russian speakers were significantly faster at discriminating color patches that crossed this linguistic boundary than within-category patches. Crucially, this advantage was eliminated by a concurrent verbal shadowing task (verbal interference), but preserved under a spatial interference task, proving that linguistic categorical representations directly interface with online visual perception.
• Spatial Framing and Coordinate Systems (Stephen Levinson): Languages utilize fundamentally different spatial reference systems:
  • Egocentric: Relative to the viewer's body (left, right, in front of), common in English and Indo-European languages.
  • Geocentric (Absolute): Cardinal directions (north, south, uphill, seaward), utilized by languages such as Guugu Yimithirr (an Australian Aboriginal language). Speakers of Guugu Yimithirr constantly compute global orientation, successfully pointing toward cardinal directions in windowless rooms with flawless precision, reorganizing non-verbal memory tasks to maintain absolute rather than relative spatial alignments.