ALL CARDS (So far)

sound wave

A traveling pattern of pressure changes in the air that moves outward from the source (e.g., the speaker's vocal tract) to the listener's ear. In the space between speaker and listener, air molecules only oscillate locally and pass the disturbance along - the molecules themselves don't travel from speaker to listener.

amplitude

The size of the pressure difference in a sound wave. Perceived as LOUDNESS (bigger pressure difference = louder). Shows up as darker shading on a spectrogram.

frequency

How quickly air molecules bunch up/spread out (waves passing a fixed point per second; measured in Hz). Perceived as PITCH (higher frequency = higher pitch). It is the y-axis of a spectrogram.

spectrogram

A visual representation of the frequency of sound waves over time.

spectrogram axes and shading

x-axis = time; y-axis = frequency (Hz). Darker areas = more acoustic energy (higher amplitude) at that frequency at that moment; lighter areas = less energy.

consonant articulatory parameters (the 3)

PLACE of articulation + MANNER of articulation + VOICING.

vowel articulatory parameters (the 3)

Tongue HEIGHT (high/mid/low = how open the jaw/mouth is) + tongue BACKNESS (front/central/back) + lip ROUNDING (rounded/unrounded).

place of articulation

Where in the vocal tract the airstream is modified (e.g., bilabial, dental, alveolar, palatal, velar, glottal).

manner of articulation

How much the airstream is obstructed. Stop/plosive = complete obstruction then release; fricative = narrow constriction that makes turbulent, hissy airflow; affricate = stop followed by a fricative release; nasal = mouth closed, air escapes through the nose; approximant (glide/liquid) = very little obstruction.

voicing

Whether or not the vocal folds vibrate during a sound's production.

shared vs. different parameters: /p/ vs. /b/

Same place (bilabial) and same manner (stop); they differ only in VOICING (/b/ voiced, /p/ voiceless).

shared vs. different parameters: /t/ vs. /s/

Same place (alveolar) and same voicing (both voiceless); they differ in MANNER (/t/ stop, /s/ fricative).

shared vs. different parameters: /i/ vs. /u/

Both are HIGH vowels; they differ in BACKNESS (/i/ front, /u/ back) and ROUNDING (/i/ unrounded, /u/ rounded).

formant

A range ('band') of amplified frequency in a spectrogram; appears as a dark horizontal band during vowels. Different vowels have different formant patterns because tongue/lip position changes the shape of the vocal tract.

F1 (first formant)

Reflects tongue/jaw height and openness.

INVERSELY related to tongue height:

low/open vowels (e.g., /ɑ/) have HIGH F1;

high vowels (/i/, /u/) have LOW F1.

F2 (second formant)

F2 tracks tongue backness/frontness (fronter tongue → higher F2, backer tongue → lower F2

FRONT vowels (/i/) have HIGH F2; BACK vowels (/u/) have LOW F2. Lip rounding also lowers F2.

F3 (third formant)

Reflects tongue tip position and lip rounding.

/i/ ('beet') on a spectrogram

High front vowel: LOW F1 (high tongue) + HIGH F2 (front tongue), so F1 and F2 are far apart (roughly F1 ~300 Hz, F2 ~2300 Hz).

/u/ ('boot') on a spectrogram

High back rounded vowel: LOW F1 (high tongue) + LOW F2 (back tongue + rounding), so F1 and F2 sit close together low on the plot (roughly F1 ~300 Hz, F2 ~900 Hz).

/ɑ/ ('father') on a spectrogram

Low back vowel: HIGH F1 (low/open tongue) + LOW-to-mid F2 (back tongue), so F1 and F2 are close together but higher than /u/ (roughly F1 ~750 Hz, F2 ~1200 Hz).

predicting a vowel's formants from its articulation

Tongue height -> F1 (higher tongue = lower F1). Tongue frontness -> F2 (more front = higher F2). Rounding lowers F2. Use the IPA chart to read off height/backness/rounding, then map to F1/F2.

stop consonant on a spectrogram

A GAP (silence/white space while the airway is closed) followed by a BURST (brief vertical spike of energy at the release).

aspiration on a spectrogram

A noisy, fuzzy stretch between the burst and the start of voicing (long VOT). English voiceless stops at the start of a word (/p t k/) are aspirated.

voicing on a spectrogram

A voicing bar (dark band at very low frequency) plus regular vertical striations while the vocal folds vibrate; absent for voiceless sounds.

voiced vs. voiceless stop on a spectrogram

Voiceless (/p t k/): gap -> burst -> aspiration (long VOT) -> vowel formants. Voiced (/b d g/): gap (voicing bar may continue) -> burst -> vowel formants with a short VOT and little or no aspiration.

% fricative on a spectrogram

Aperiodic noise (frication): a diffuse smear of energy, strongest at high frequencies for /s/. Voiced fricatives (/z/) also have a voicing bar.

why L2 speakers have accents

L1 articulatory patterns become deeply automated by the motor cortex,

so the articulatory system is not a 'blank slate' when learning an L2.

Perception of L2 contrasts is also filtered through L1 categories so learners may not hear or produce the difference.

epenthesis (example)

Japanese speakers, whose language follows strict (C)V syllable structure, may insert vowels into English consonant clusters (e.g., 'street' /striːt/ becomes /sɯtɯriːto/).

% why speakers of the same language have different accents

Speech patterns drift over time across communities due to both physical and social boundaries.

% schwa (ə)

A reduced vowel, very brief, occurring in unstressed positions (e.g., blossom, Sbarro).

% caret/wedge (ʌ)

A non-reduced vowel in a stressed position (e.g., does, sun).

STRESSED NON REDUCED VOWEL

% diphthong

A long, complex vowel that changes quality during articulation (e.g., bite, boy, bough).

grey matter

has most of brain's neuronal cell bodies

is the location of synapses.

white matter

Contains most of the bundles of fibers coming from the cell bodies.

cerebral cortex

The outer layer of the brain (the grey matter).

gyrus (gyri, plural)

A visible ridge of the cortex.

sulcus (sulci, plural)

A folded-in groove of the cortex.

frontal lobe

Associated with thought, planning, and movement.

parietal lobe

Associated with touch and spatial relations.

temporal lobe

Associated with hearing and memory.

occipital lobe

Associated with vision.

primary auditory cortex (A1)

The core region of cortex that first processes auditory input. Located in the temporal lobe (superior temporal gyrus/Heschl's gyrus).

V1

Primary visual cortex; occipital lobe.

M1

Primary motor cortex; frontal lobe (precentral gyrus).

S1

Primary somatosensory cortex (touch); parietal lobe (postcentral gyrus).

A1 / V1 / M1 / S1 summary

A1 = primary auditory (temporal lobe); V1 = primary visual (occipital lobe); M1 = primary motor (frontal lobe); S1 = primary somatosensory (parietal lobe).

transduction

The action or process of converting something into another form; in hearing, converting waves of air pressure into neural firing patterns.

where does transduction happen?

In the cochlea: hair cells on the basilar membrane bend as the fluid/membrane vibrates and convert the mechanical vibration into neural signals.

outer ear function

Sound waves enter the auditory canal, causing vibrations that reach the eardrum.

eardrum (tympanic membrane)

Vibrates in response to sound and passes vibrations to the middle ear.

middle ear bones (ossicles)

Transfer vibrations from the eardrum to the inner ear (cochlea).

cochlea

The spiral, fluid-filled structure in the inner ear.

basilar membrane

Structure inside the cochlea that vibrates in response to sound (hair cells on it perform transduction).

basilar membrane: low vs. high frequencies

The base (near the middle ear) is narrow and stiff and responds to HIGH frequencies; the apex (far end) is wide and flexible and responds to LOW frequencies.

From sound to the ear (steps 1-5)

1) Sound waves enter the auditory canal of the OUTER EAR. 2) Vibrations reach the ear drum. 3) The ear drum vibrates, sending vibrations to the MIDDLE EAR. 4) Bones in the middle ear transfer them to the INNER EAR. 5) Vibrations in the liquid in the cochlea jiggle the basilar membrane.

auditory pathway to cortex (steps 6-8)

Information travels from the cochlea through the brainstem, midbrain, and thalamus (MGN) to auditory cortex (A1), passed both ipsilaterally (same side) and contralaterally (opposite side).

full pathway: speaker to A1

Speaker's articulators/vocal folds produce a sound wave -> travels through air -> auditory canal -> ear drum -> ossicles (middle ear) -> cochlea fluid moves the basilar membrane -> hair cells transduce vibration to neural signals (auditory nerve) -> brainstem -> midbrain -> thalamus (MGN) -> primary auditory cortex (A1).

tonotopic organization

Neighboring frequencies map to neighboring locations. The map starts on the basilar membrane (base = high, apex = low) and is preserved in A1, so A1 is 'tonotopically organized'.

% speech perception (definition, Poeppel & Monahan)

The subroutine of comprehension that transforms an auditory signal into mental representations that can contact stored linguistic information (words).

Poeppel & Monahan: is there a single brain region for speech perception?

No. Speech perception is carried out by a distributed network (bilateral temporal cortex plus parietal and frontal regions), not by one dedicated area.

superior temporal gyrus (STG) / superior temporal sulcus (STS)

Key structures in the dual-stream model implicated in phonological processing.

dual-stream model of speech (Hickok & Poeppel)

Speech processing splits into a ventral stream (sound to meaning) and a dorsal stream (sound to articulation/motor).

ventral stream

Maps sensory/phonological representations to lexical/conceptual representations (sound-to-meaning).

dorsal stream

Maps sensory/phonological representations to articulatory-motor representations (sound-to-motor).

⚑ ventral stream: location + what damage would affect

Runs along the temporal lobe (STG/STS toward middle/anterior temporal cortex); sound -> meaning. Damage (with dorsal spared): trouble understanding words/comprehension, but can still repeat speech.

⚑ dorsal stream: location + what damage would affect

Runs from posterior temporal cortex toward parietal and frontal/motor regions; sound -> articulation. Damage (with ventral spared): comprehension relatively OK, but repetition/production (esp. of nonwords) suffers.

% phonological network

Region implicating mid-posterior superior temporal sulcus (bilateral); sits between spectrotemporal analysis and the lexical/articulatory networks in the dual-stream model.

phoneme/segment

A technical term for an individual speech sound.

phoneme (definition)

A speech-sound category that distinguishes words in a language (swapping it changes the word).

phonemic contrast

Two sounds are phonemic in a language if swapping them changes word meaning (a minimal pair). E.g., /b/-/p/ in English (bat/pat);

dental /d̪/ vs. retroflex /ɖ/ would be phonemic pair in Hindi but NOT in English.

% distinctive features

The smallest hypothesized units of speech representation, stated as articulatory primitives (e.g., [coronal], [voiced]).

mismatch negativity (MMN/MMF)

An EEG/MEG response that is larger for a 'deviant' stimulus than for a 'standard' stimulus in a series.

Näätänen et al. (1997) finding

Finnish and Estonian speakers both showed an MMN for a vowel deviant present in both languages' inventories, but only Estonian speakers showed an MMN for a deviant present only in Estonian - showing native-language representations shape early auditory processing, not just acoustics.

Kazanina, Phillips, & Idsardi (2006) finding

Russian speakers (whose inventory contrasts /t/ and /d/) showed an MMF for that contrast, but Korean speakers (whose inventory has only one sound /T/) did not - again showing native-language representations constrain auditory processing.

neural evidence that speech-sound processing depends on your language (P&M 2008 example)

Either MMN study works: Näätänen et al. (1997) - Estonian but not Finnish speakers showed an MMN for an Estonian-only vowel; or Kazanina et al. (2006) - Russian but not Korean speakers showed an MMF for /t/ vs. /d/. Same acoustics, different brain responses depending on the native inventory.

% sound-motor mapping (motor theory vs. auditory theory)

Motor theory holds that accessing (aspects of) production/articulatory mechanisms is critical to perceiving speech.

% audiovisual speech finding (van Wassenhove et al., 2005)

The more transparent/unambiguous a visual speech signal (viseme), the faster the auditory signal is processed in auditory cortex - supports analysis-by-synthesis models where top-down predictions shape early processing.

% what is analysis-by-synthesis

A model of perception in which top-down hypotheses, from available information, modulate lower-level sensory analysis.

% granularity mismatch problem

The difficulty of explicitly linking cognitive-science representations (e.g., features, syllables) to neurobiological mechanisms (e.g., neurons, synapses, oscillations).

voice onset time (VOT)

time delay between releasing a stop consonant (like p, b, t, or d) and vocal cords vibration.

It acts as a timing cue that helps distinguish between voiced and voiceless sounds:

• 
  

categorical perception

hearing gradient (continuous) acoustic differences across a stimulus continuum as if there were an actual categorical differences(e.g., hearing a VOT continuum as either /ba/ or /pa/ with a sharp boundary, not a gradient).

(the measurement is continuous but your perception is not)

English speakers on a /b/-/p/ VOT continuum

Tokens are heard as two categories with a sharp boundary: short VOT = /b/, long VOT = /p/. Discrimination is good across the boundary and poor between tokens within the same category (categorical perception).

English vs. Hindi speakers on a /d̪/-/ɖ/ (dental-retroflex) continuum

Hindi speakers hear two categories with a sharp boundary (the contrast is phonemic in Hindi). English speakers hear all tokens as one /d/ category with no boundary, because the contrast isn't phonemic in English.

perceptual narrowing (Werker & Tees, 1984)

Infants around 6-8 months can discriminate non-native contrasts (e.g., Hindi dental vs. retroflex /d/), but by 10-12 months, infants raised without exposure to that contrast (e.g., English learners) lose the ability to discriminate it, while infants exposed to the contrast (e.g., Hindi learners) retain it.

⚑ Werker & Tees: findings by language group and age

English-learning infants: most discriminate the Hindi contrast at 6-8 months, but few by 10-12 months. Hindi-learning infants (10-12 months) still discriminate it. Adults: Hindi speakers discriminate it; English speakers mostly cannot.

⚑ Werker & Tees: why did they need both language groups?

Hindi speakers/infants show that the contrast is perceivable and discriminable with these stimuli, so the English groups' failure comes from lack of experience with the contrast (not from the stimuli being too hard, or age alone). Comparing age groups shows the change happens during the first year.

Werker & Tees: what do results suggest about the first year?

Speech perception is tuned to the native language during the first year (roughly 6-8 to 10-12 months):

early on infants are 'universal listeners,'

then perception narrows toward the contrasts for their language.

why perceptual narrowing/reorganization could be advantageous

Sharpens perception toward the contrasts that matter for distinguishing words in one's native language, though it comes at the cost of reduced sensitivity to non-native contrasts.

⚑ do adults really 'lose' non-native contrasts?

Not entirely. Perception is 'warped': non-native sounds get pulled toward the nearest native category,

so sensitivity drops, but the ability is reduced, not erased.

⚑ what does it mean for perception to be 'warped'?

Native-language categories act as a filter: non-native sounds are perceived as (assimilated to) the closest native category, so differences within a native category are hard to hear and contrasts that fall inside one L1 category get merged.

⚑ what is learning a new L2 contrast like?

Effortful and gradual; needs lots of exposure or training; can improve but is often not fully native-like, because the sounds are being filtered through L1 categories.

predicting perception of a new contrast (template)

1) Is the contrast phonemic in the listener's language?

2) Infants ~6-8 months: can discriminate either way.

3) Infants 10-12 months and adults: discriminate it only if it is phonemic in their L1; otherwise both sounds fall into one category.

4) Justify with perceptual narrowing/warping.

Orthography

The field/study of writing systems and spelling conventions used to represent language visually.

Writing vs. language

Writing is only a means of expressing language - it is not language itself. Language is the underlying system of speech/meaning.

what does 'writing is secondary to language' mean?

Writing is a way of recording/encoding spoken language, so it depends on language. Speech comes first: it is universal and acquired naturally, while writing is a recent invention that has to be taught.

Protowriting

Early symbol systems that record information but don't encode a specific spoken message - no fixed mapping to sounds/words (e.g., Jiahu symbols, Vinča symbols, proto-cuneiform).

Protowriting vs. true writing

Protowriting records general info without a fixed message or link to spoken language. True writing systematically maps symbols onto units of a spoken language (morpheme, syllable, or phoneme).

Pictogram

A direct representation of an object/event in the world (looks like what it stands for).

logogram vs. pictogram

A pictogram depicts its visuals (looks like the thing). A modern logogram is a conventional symbol for a specific word/morpheme of a language (it encodes sound + meaning) and may not resemble the referent at all.

% Naxi Dongba script

A surviving pictographic system (Yunnan, China) used for ritual/religious texts.