All varieties

Barranquilla elides intervocalic /d/ hard (pesao, cansao) and aspirates coda /s/ heavily, all riding on the sing-song lilt of the Colombian Caribbean coast.

Cartagena shares Barranquilla's aspiration and /d/-elision but leans even further into syllable dropping, giving it the loosest, most syncopated feel on the Caribbean coast.

Central Cuba is genuinely transitional: speakers here vary between Havana's gemination and Oriente's plain aspiration rather than committing fully to either strategy.

Oriente favors straightforward aspiration over Havana's gemination: coda /s/ softens to a breathy [h] rather than doubling the next consonant.

Havana's trademark is gemination: a lost coda /s/ or /r/ doesn't just vanish, it assimilates and doubles the next consonant, so carta becomes "catta".

Santo Domingo and the southern regions favor lambdacism: coda /ɾ/ and /l/ collapse into [l], so puerta comes out "puelta".

Cibao is famous for i-vocalism: coda /r/ and /l/ both turn into a glide, so puerta becomes "pueita" and mujer becomes "mujei".

The eastern DR bucks the regional trend of losing coda /ɾ/: instead it reinforces it, so porque can surface with a doubled, trilled r.

The educated, prestige-influenced register of the capital softens the countryside's strongest features: coda /s/ merely aspirates rather than vanishes, and coda /r/ gets only a light, optional softening instead of full lambdacism.

The conservative folkloric heartland of Panama keeps coda /s/ and the infinitival /r/ far more intact than the syncopated, syllable-dropping speech of Panama City.

Panamanian Spanish drops whole unstressed syllables in casual speech (syncopation) and softens rather than deletes the infinitival /r/.

The velar/guttural 'jueRRte' R is Puerto Rico's calling card, paired with coda /s/ aspiration and the lambdacism that turns puerto into "puelto".

The rural highland jibaro register pushes past the island-wide velar-R norm: lambdacism runs heavier, /f/ aspirates to [h] before /w/, and casual speech drops whole unstressed syllables.

Caraqueño Spanish aspirates coda /s/ and velarizes final /n/, and that velar nasal spreads right across word boundaries: 'un amigo' surfaces with a lingering [ŋ].

Zulia is the only voseo-using region of Venezuela, giving Maracaibo Spanish a distinctive melody that sets it apart from the tuteo norm everywhere else in the country.

Bogotano Spanish is famously conservative and clear: fully retained coda /s/, standard consonants across the board, and a strong preference for the polite usted even among close friends and family.

Paisa Spanish is one of the most recognizable accents in the Spanish-speaking world: a buzzy assibilated /r/, voseo, and a sing-song intonation all its own.

Colombia's southernmost highland variety is conservative and close to the Ecuadorian sierra just across the border: /ll/ survives as [ʎ] and /b d g/ resist lenition.

Valluno Spanish pairs voseo with coda /s/ that weakens all the way to deletion at Cali's fast, danceable tempo -- a Pacific-valley blend that differs sharply from Bogotá's ustedeo, retained-/s/ norm just over the mountains.

Ecuador's easternmost variety, shaped by Kichwa contact, shares the highlands' assibilated /r/ and retained /s/, but at a slower, more level pace than Quito's rising Sierra norm.

Guayaquil breaks sharply from the Ecuadorian highlands: coastal speech aspirates and deletes coda /s/ like a Caribbean variety, closer to Peru's coast than to Quito's mountains, and jota aspirates to [h] rather than the highlands' velar fricative.

Quiteño Spanish shares the Andean assibilated /r/ and adds a distinctive zh-like realization of intervocalic /ll,y/ that stands out from Rioplatense zheísmo in its buzzier, more fronted quality.

Peru's Amazonian lowlands sound distinct from both coastal Lima and Andean Cusco: /b d g/ resist lenition, /f/ is documented as bilabial rather than labiodental, and phrase-final intonation rises noticeably.

Quechua contact shapes Cusco Spanish: /r/ becomes assibilated and buzzy, unstressed vowels devoice or drop, and /b d g/ resist softening between vowels, staying occlusive.

Limeño Spanish is one of this schema's comparison points: crisp, fully retained coda /s/ and clean, open vowels, contrasting with the vowel/consonant reductions documented for highland varieties. 'Comparison point' is a modeling convenience, not a claim that this is more correct Spanish than the highlands -- every variety here differs from every other; this one just happens to anchor the engine's internal diff.

Peru's northern coast is most reliably distinguished by its intonation, not its consonants: a rising, song-like contour that Peruvians themselves routinely name as the clearest marker of 'hablar norteño.' Segmental behavior (coda /s/, vowels) stays close to the Lima coastal pattern; this preset deliberately does not claim specific consonantal differences from Lima that this author could not verify against a source (flag for native-speaker review). The region also carries a well-known distinct vocabulary this app's phonetic matrix does not represent -- lexicon sits outside what this model captures.

Conservative highland Venezuelan Spanish bordering Colombia keeps coda /s/ sharp and defaults to ustedeo, standing apart from the aspirating, tuteo-leaning Caribbean coast.

Porteño Spanish is unmistakable: sheísmo turns 'yo' and 'llamar' into 'sho' and 'shamar', voseo replaces tú, and the melody rises and falls in the famous Rioplatense bounce.

The famous 'tonada cordobesa' gives Cordoban Spanish a sliding, musical quality distinct from the Porteño bounce. Linguistically the tonada stretches the pretonic syllable rather than the phrase-final one; final_syllable_lengthening is used here as the engine's closest available proxy for that stretching effect. Like the rest of the Rioplatense dialect area, Córdoba also shares sheísmo/zheísmo -- the /ll,y/ rehilamiento is not a Buenos Aires-only feature.

Guaraní contact along the Argentine littoral gives Corrientes a distinctive melody and more conservative consonantism that sets it apart from the Porteño sheísmo core.

The Andean-influenced northwest breaks from the Rioplatense norm: assibilated /r/ and a slower, highland-adjacent lilt replace the sheísmo and bounce of Buenos Aires.

Montevideo shares Buenos Aires' voseo but keeps the older, voiced zheísmo [ʒ] rather than fully devoicing it to [ʃ], giving it a softer, buzzier palatal sound.

Costa Rican Spanish pairs an assibilated, buzzy /r/ (like 'pero' sounding almost like 'pedro') with the -ico diminutive and a strong ustedeo default among Ticos.

English-Creole contact on the Caribbean coast makes Limón Spanish sound utterly unlike the Central Valley's assibilated-r, ustedeo norm: coda /s/ weakens all the way to deletion (further along than the Central Valley's fully retained /s/), final /n/ velarizes, and the rhythm shifts toward stress-timing.

Salvadoran Spanish uses voseo consistently and carries the region's sing-song contour, with coda /s/ softening toward aspiration in casual speech and jota realized as [h] like the rest of Central America.

Guatemalan Spanish mixes tú and vos depending on formality and region, riding on the gentle sing-song intonation shared across Central America, and -- like the rest of the isthmus -- aspirates jota to [h] rather than the velar fricative of Spain or the Andes.

Maya-substrate Spanish in the western highlands fortifies /b d g/ and keeps coda /s/ sharp, standing apart from Guatemala City's mixed tú/vos, aspirated norm.

Honduran Spanish leans on voseo and the same sing-song Central American melody, with coda /s/ regularly aspirated in fast speech and jota realized as [h] like the rest of the isthmus.

An English-Creole-speaking Caribbean archipelago where Spanish is a second language for many islanders. When spoken, it carries a stress-timed rhythm, Caribbean-style vowel opening, and coda /s/ weakened all the way to deletion -- further along than mainland Tegucigalpa's aspirating-only norm.

Nicaraguan Spanish is Central America's strongest voseo hold-out, used almost universally, layered on the region's characteristic sing-song rhythm, aspirated /s/, and jota realized as [h] like the rest of the isthmus.

English-Creole and Miskito contact set Bluefields apart from Managua's Pacific norm -- voseo survives, but the rhythm leans stress-timed and coda /s/ weakens all the way to deletion, further than the Pacific side's aspiration-only norm.

The Andean-influenced far north softens Santiago's radical consonant loss: coda /s/ merely aspirates, and the assibilated /r/ of the neighboring Peruvian/Bolivian highlands creeps in.

Chilean Spanish is the fastest variety in the Spanish-speaking world: glottal stops replace dropped codas, /tʃ/ fronts to [ʃ], /tɾ/ frictionalizes, and both tú and vos appear depending on the speaker.

The most conservative Chilean variety: island isolation preserved lleísmo remnants and slowed the mainland's radical consonant loss and glottal-stop insertion.

The ceceo belt around Cádiz and rural western Andalusia does the opposite of seseo: s merges into [θ] as well, so 'sí' and 'gracias' both come out with that lisped, dental buzz.

Western Andalusia (Seville) is a seseo zone within Spain: c/z merge with s rather than staying distinct as [θ]. Coda /s/ weakening goes past aspiration all the way to outright deletion in casual Sevillian speech -- further along the same lenition path than the aspirating-only Caribbean varieties it helped seed centuries ago -- and intervocalic /d/ elides on top of that.

The Canaries sound more Caribbean than Peninsular: seseo, aspirated coda /s/, a lilting final-syllable stretch that Cuban and Puerto Rican Spanish inherited from centuries of Canarian emigration, and the same aspirated [h] realization of jota shared with the Caribbean and Andalusia rather than mainland Spain's velar fricative.

A transitional zone between Castilian distinción and Andalusian lenition: Extremaduran Spanish aspirates coda /s/ and elides intervocalic /d/ like its southern neighbor, while remaining a seseo area like the rest of non-Andalusian Spain.

The reference Peninsular variety: distinción keeps c/z as [θ] distinct from a fully retained apical /s/, and jota is a hard uvular [χ]. This is the mainstream 'Castilian standard' description found across the literature, though plenty of madrileños show casual /d/ and /s/ weakening in fast speech.

Heavy vowel opening is the hallmark here: coda /s/ drops almost entirely, but the preceding vowel opens so dramatically that the lost consonant is still audible as a vowel-quality (and length) shift -- the classic 'eastern Andalusian' vowel system.

Central Mexican Spanish's defining trait is unstressed-vowel devocalization: weak vowels next to /s/ get muted or dropped entirely, while that same /s/ stays sharply retained and consonants come out crisp and precise.

The Gulf coast breaks from the rest of Mexico with a Caribbean-type consonant system: coda /s/ aspirates, final /n/ velarizes, and intervocalic /d/ elides, all riding on a bright, fast coastal rhythm.

Norteño Spanish fronts /tʃ/ to [ʃ] and moves noticeably faster than the central highland norm, while keeping coda /s/ just as sharp and retained.

Maya-substrate contact gives Yucatecan Spanish glottal stops in place of dropped codas, occlusive /b d g/ that resist softening, and a lengthened, distinctively rising phrase-final intonation. (Yucatecan Spanish is also documented realizing word-final /n/ closer to [m] in some words, a lexical-phonetic detail this matrix does not model directly.)

Quechua/Aymara contact on the high altiplano produces a profile close to Cusco Peru: lleísmo, assibilated /r/, devoiced unstressed vowels, and occlusive /b d g/ that resist softening.

Bolivia's eastern lowlands sound almost like a different country from the Andean altiplano: coda /s/ weakens all the way to deletion, voseo is standard, and the pace picks up considerably, giving Camba Spanish a distinctly Caribbean-adjacent feel -- the same lowland-tropical /s/-lenition pattern found in Caribbean coastal Spanish, well past highland Bolivia's fully retained /s/.

Widespread Guaraní-Spanish bilingualism shapes Asunción Spanish: /ll/ survives distinct as [ʎ], /r/ is assibilated, voseo is universal, and /b d g/ resist the lenition common elsewhere in Latin America.

The only Spanish-official variety native to Sub-Saharan Africa: /b d g/ resist lenition, coda /s/ shows none of the aspiration common across Latin America, and a Bantu-substrate prosody gives it a distinct, level, syllable-timed cadence unlike any American or Peninsular variety.

A Mexican base reshaped by generations of English contact in the American Southwest: retained crisp /s/ and the fronted CH of northern Mexico persist, riding on an English-influenced stress-timed rhythm.

A Cuban base preserved remarkably intact across generations in Miami's dense exile community: gemination and glottal jota persist alongside a distinctly English-contact stress-timed rhythm.

A Puerto Rican base reshaped by generations of English contact in New York: the velar R and lambdacism survive, but the rhythm drifts toward the stress-timing of surrounding English.

Belém shares Rio's coda-/s/ chiado -- an isolated, non-adjacent match that sets it apart from the rest of the North region -- layered on top of its own distinct, indigenous-contact-influenced prosody.

Baiano speech carries a distinctive melodic lilt widely attributed to Salvador's deep African (chiefly Yoruba/Bantu) substrate contact, with open, unreduced vowels and a slower, singsong pace compared to the Southeast.

Brasília is a planned capital built by internal migrants from across Brazil, so its Portuguese is a genuine koiné: it keeps the pan-Brazilian core (palatalization, final raising, nasalization) but without a single strong regional marker like Rio's chiado or Minas's syncopation.

Caipira's calling card is the retroflex 'R caipira' [ɻ], heard in porta and mulher -- the schema has no dedicated retroflex-R value, so this preset approximates it with r_initial='assibilated' (the closest available IPA-adjacent option, [ɹ̝~ʐ] rather than the true [ɻ]) and r_guttural_strength='none' to keep it firmly non-guttural.

Carioca speech is Brazilian Portuguese's most recognizable export: the 'chiado' turns every coda /s/ into [ʃ] (mas -> [maʃ]), the R is a strong guttural fricative, and /t d/ affricate hard before /i/.

Florianópolis was settled largely by Azorean colonists, and it shows: it has its own chiado (coda /s/ -> [ʃ]), inherited from Azorean Portuguese rather than from Rio, which sets it apart from its Gaúcho neighbor to the south.

Border contact with Rioplatense Spanish shapes Gaúcho Portuguese: an alveolar-tap (not guttural) R, weaker or absent /t d/-palatalization in much of the state, and a melodic contour close enough to Argentine/Uruguayan Spanish that this preset borrows the engine's 'rioplatense_bounce' intonation label as the closest available proxy.

Mineiro speech is famous for swallowing whole unstressed syllables (uai virou 'uai' but 'está' becomes 'tá', 'para' becomes 'pra') and for heavy unstressed-vowel reduction, riding on the drawled 'uai' prosody.

Recife Portuguese keeps its open/close mid vowels crisp rather than reducing them, and several Nordestino subvarieties documentedly lack (or strongly weaken) the /t d/-palatalization that defines Rio/São Paulo speech -- a real point of internal Nordestino variation, not a uniform rule across the whole Northeast.

Paulistano speech shares Rio's /t d/-palatalization and final-vowel raising, but firmly rejects the chiado: coda /s/ stays a plain sibilant, and the R is a light tap rather than Rio's guttural rasp.

Azorean Portuguese is documented as having a notably different vowel quality from mainland Lisbon -- less of Lisbon's extreme schwa-like reduction -- which is also why the Azorean settlers of Florianópolis (see the Brazilian preset library) left their chiado behind but not Lisbon's full reduction system.

Alentejo's rural southern speech is slow and deliberate, with noticeably lengthened final syllables and vowels -- a much gentler delivery than Lisbon's clipped, reduced norm, even though the same underlying reduction and uvular R are present.

Algarve Portuguese follows the mainstream southern-Portuguese profile (Lisbon-style reduction, uvular R, chiado); its most-cited distinguishing traits are lexical/prosodic rather than segmental, so this preset is deliberately close to -- but not identical to -- the Lisbon standard.

Standard European Portuguese's defining trait is brutal unstressed-vowel reduction -- casa is barely two full vowels, closer to ['kaz.ɨ] -- paired with a uvular R and, unlike Brazilian Portuguese, no /t d/-palatalization before /i/ at all.

Madeiran Portuguese is often described impressionistically by mainland listeners as having even more pronounced vowel-raising/closing than Lisbon, alongside a strong, consistently uvular R -- modeled here as Lisbon's full reduction system plus the strongest guttural-R setting.

Northern Portuguese is best known for betacism, the historic b/v merger (a lexical/segmental trait the current schema has no dedicated field for, so it isn't phonetically rendered here -- noted for the record, not modeled). What *is* modeled: Porto keeps the alveolar trill rather than adopting Lisbon's uvular R, and resists Lisbon's heaviest unstressed-vowel reduction.

Benguela shares Luanda's core Bantu-substrate profile (fortified stops, unreduced vowels) but carries a noticeably slower, more rural-inflected cadence than the capital's urban norm.

Luanda Portuguese is shaped by deep Kimbundu (Bantu) contact: unlike European Portuguese, unstressed vowels resist reduction and stay full, intervocalic /b d g/ fortify rather than lenite, and final vowels are retained rather than raised the Brazilian way.

This preset models the *Portuguese* spoken in Cape Verde (not Cape Verdean Creole/Kriolu, which belongs to the Creole preset agent's file) -- a variety in close, ongoing contact with both European Portuguese norms and the Creole that is most Cape Verdeans' first language, which shows up here as informal-register syllable-dropping.

Guinea-Bissau Portuguese sits alongside Guinea-Bissau Creole (Kriol) as a minority first language for most speakers; its West African Niger-Congo substrate produces the same broad profile as Angola/Mozambique -- unreduced vowels, fortified stops -- without Angola's Kimbundu-specific details.

Maputo Portuguese shares Angola's Bantu-substrate resistance to vowel reduction, but its consonant system is documented as distinct: Mozambican Portuguese generally does not show the same degree of /b d g/-fortition found in Angolan Portuguese.

São Tomé e Príncipe's Portuguese exists alongside several distinct local Creoles (Forro, Angolar, Lung'Ie); its Portuguese carries the same broad Bantu/Niger-Congo-substrate profile as its African Lusophone neighbors.

Portuguese in Macau survives today as a co-official but minority language spoken fluently by a small community, historically alongside the (now nearly extinct) Patuá creole. As with Timorese Portuguese, rigorous instrumented phonetic studies of a distinct 'Macanese Portuguese' accent are scarce, so this preset is a cautious, documented approximation rather than a well-attested standard.

Portuguese is a co-official language in Timor-Leste alongside Tetum, spoken mostly as a learned second/administrative language. Instrumented phonetic documentation of a distinct 'Timorese Portuguese' accent is sparse compared to Brazil/Portugal/Angola, so this preset is a lower-confidence approximation built from the general pattern documented for Portuguese spoken with a Tetum-substrate accent (full, unreduced vowels; syllable-timed rhythm), not a fully attested regional standard.

Aruban Papiamento shares Curaçao's Iberian-lexified, tonal creole grammar but leans more Hispanicized in spelling and vocabulary -- the island's closer historical and economic ties to Venezuela show up as the '-o' ending (Papiamento) rather than Curaçao and Bonaire's Dutch-influenced '-u' (Papiamentu), and in a heavier layer of Spanish-derived vocabulary. Known gap, shared with every Papiamento/u preset: the documented lexical tone contrast has no representation anywhere in this schema -- not scored, not synthesized -- so treat 'tonal' as a real fact about the language you won't be able to hear or be graded on here.

Bonaire's Papiamentu sits closest to Curaçao's norm and shares its '-u' spelling convention, but its smaller, more isolated speech community is documented as preserving older lexical items and a slower cadence than the more urban Willemstad.

Palenquero is the only Spanish-lexified creole language in the Americas, spoken in San Basilio de Palenque, the first free African town in the New World, founded by escaped enslaved people (cimarrones) in the early 1600s. Its grammar shows a clear Kikongo (Bantu) substrate underneath a mostly Spanish-derived vocabulary: preverbal TMA particles, postverbal negation, and a prosody markedly different from the surrounding Caribbean Colombian Spanish. UNESCO recognized the Palenque cultural space as a Masterpiece of the Oral and Intangible Heritage of Humanity in 2005.

Papiamentu is an Iberian-lexified creole -- Spanish and Portuguese word-stock filtered through Dutch colonial administration and West African (Kikongo/Kimbundu) substrate grammar -- documented (Kouwenberg & Murray 1994) as having genuine lexical and phrasal tone, a pitch contrast neither Iberian source language has. Curaçao (with Bonaire) uses the Dutch-influenced '-u' spelling convention, Papiamentu, distinct from Aruba's more Hispanicizing 'Papiamento'. Known gap: there is no tone-bearing-unit representation anywhere in the current schema, so this preset cannot express lexical tone at all -- not in the scoring matrix and not in the synthesized audio (espeak's flat pitch can't approximate it either). Read tone as a real, cited fact about the language, not something this tool can currently teach you to hear or produce.

The Kreyòl of Cap-Haïtien and the Plaine du Nord is well documented for pronoun forms and agreement patterns that diverge from the Port-au-Prince standard, riding on the same nasal-vowel, TMA-particle grammar shared island-wide, but with a noticeably heavier, more pervasive nasal vowel quality.

Kreyòl ayisyen is a full, autonomous language with its own 1979 IPN orthography and its own grammar -- not a French dialect and not 'broken French'. The Port-au-Prince/standard variety anchors dictionaries and schooling: phonemic nasal vowels, a seven-vowel oral system that spells the open/close mid vowels separately (e vs è, o vs ò), and preverbal tense-mood-aspect particles (te, ap, pral) rather than Spanish/French-style verb conjugation.

The southern Kreyòl of Les Cayes and the Grand'Anse shares the island-wide TMA-particle grammar and nasal-vowel system but is documented as leaning toward phrase-final syllable lengthening, giving southern speech a distinctive drawn-out cadence.

Jamaican Patois (Patwa) is a full, systematic creole language, not 'broken English' -- most of its vocabulary is English-derived, but its grammar is not: preverbal tense-mood-aspect particles ('a' for the progressive, 'did'/'en' for the past, 'a go'/'wi' for the future) replace English-style verb conjugation, the same structural pattern this app's Haitian Kreyòl and Cape Verdean Kriolu presets already model for their own lexifier languages (Cassidy & Le Page 1980, Dictionary of Jamaican English). This preset models Kingston urban speech toward the basilectal (most creole-grammar-heavy) end of Jamaica's creole continuum, in deliberate contrast with this app's existing 'jm-jamaican' preset, which models the acrolectal English end -- the two are related but genuinely different presets, not duplicates. TH becomes a stop, shared with Jamaican English (thing -> ting, this -> dis), and coda consonant clusters simplify, a well-documented Caribbean creole pattern (best -> bes'). This app's engine cannot yet mechanically render two of Patois's most audible features -- TH-stopping and the documented FACE->[ie]/GOAT->[uo] diphthongs (bakes -> 'bieks', boat -> 'buot') -- because the transformer rules for both are gated to `language='en'` presets specifically, and this preset is correctly tagged `language='mixed'` instead (Patois is an autonomous creole, not an English dialect). Both are documented here in the blurb rather than silently omitted; the app's rendered audio for this preset will not yet capture them.

The triple-frontier zone around Ciudad del Este, Foz do Iguaçu, and Puerto Iguazú layers Portuguese-Spanish contact on top of Paraguay's existing Guaraní-Spanish bilingualism, producing a three-way contact variety that is less codified than the more stabilized Rivera-Livramento DPU further south.

Along the Uruguay-Brazil border, generations of bilingual contact produced DPU (Dialectos Portugueses del Uruguay), popularly called 'Portuñol Riverense' after the twin cities of Rivera and Santana do Livramento. It is not code-switching between two separate languages but a genuinely mixed, rule-governed variety with its own stable phonology -- Portuguese-style nasal vowels and final-vowel raising riding on a largely Spanish-derived lexicon, a real showcase for Familect's mix-and-match premise.

Chabacano (Zamboangueño) is a Spanish-lexified creole that developed in the Philippines under Austronesian (Visayan) grammatical influence, making it geographically the furthest-flung Spanish-lexified creole in the world. Verb aspect-marking and word order follow the Philippine-Austronesian substrate even though most content words trace back to Spanish.

São Vicente's Barlavento Kriolu -- the variety of Mindelo -- diverged from Sotavento Santiago Kriolu partly through later settlement and heavier 19th-20th century Portuguese contact via the island's coaling-port economy, giving it a phonology closer to European Portuguese in places even though its grammar remains fully Kriolu, distinct from Portuguese.

Kriolu (Kabuverdianu) is a Portuguese-lexified creole, not Portuguese itself: its grammar (preverbal TMA particles, its own pronoun and negation system) diverges sharply from European or Brazilian Portuguese even though most vocabulary traces to 15th-16th century Portuguese. Santiago's Sotavento Kriolu is the older, more Africanized branch (Mandinka/Wolof substrate), phonologically and lexically more conservative than the northern Barlavento varieties.

Dominican-York Spanish is the second-generation NYC Dominican profile -- distinct from the existing Nuyorican (Puerto Rican) preset in this library. It carries a Cibao-style i-vocalized coda liquid, reshaped by the same English-contact stress-timing found across the city's Spanish-speaking diaspora, plus frequent code-switching with NYC English.

The heritage-speaker profile describes a widespread, well-studied pattern among second-generation US Spanish speakers: strong receptive competence and largely intact phonemic perception, alongside production that shows systematic English-influenced simplification -- English-timed rhythm and occasional final-consonant lenition. This is a documented bilingual-acquisition pattern, not a deficient or 'incomplete' Spanish; heritage speakers are full bilinguals with their own describable system.

Isleño Spanish descends from Canary Islanders resettled in Spanish Louisiana in the 1770s-80s and survived in St. Bernard Parish's fishing communities for two centuries in relative isolation. It is now critically endangered: fieldwork and documentation projects describe only a handful of elderly speakers left after Hurricane Katrina scattered the community. This preset represents a documented, historically important variety, not a thriving one, and should be understood that way.

Descended from colonial-era Spanish largely isolated from later immigration waves after 1848, Traditional New Mexican Spanish preserves archaisms lost almost everywhere else: forms like 'naiden' for nadie and 'ansina' for así, an aspirated jota, and a conservative, fully retained coda /s/. Linguists treat it as a genuine relic dialect -- arguably the most historically significant Spanish variety spoken on US soil.

Salvadoran-American Spanish, concentrated in Los Angeles' Pico-Union/Westlake corridor and the DC-Maryland-Virginia area, retains El Salvador's voseo across second- and third-generation US-born speakers even as English becomes the dominant home language -- a striking case of a morphosyntactic feature outlasting heavy language shift.

Tejano Spanish anchors South Texas's Rio Grande Valley: a northern-Mexican consonant base reshaped by generations of stable English-Spanish bilingualism, with a stress-timed rhythm and frequent Tex-Mex code-switching that sets it apart from both interior Mexican Spanish and the more recent-immigrant Chicano Southwest profile already in this library.

African American English is a systematic variety with its own consistent phonology and grammar, spoken across the whole United States rather than in one region, and it varies internally: a Detroit speaker, a Houston speaker and a rural Mississippi speaker differ from each other in rhoticity, in vowel quality and in how much of the Southern Vowel Shift they share. This preset models the widely-described common core -- variable non-rhoticity, PRICE monophthongized before voiced sounds and finally, TH realized as a stop word-initially and as a labiodental fricative elsewhere, and coda cluster reduction -- and it is a profile, not a definition. Its rules are rules, not errors, and it is modelled here alongside General American, not against it. Two well-described features are again out of the matrix's reach because they are environmentally conditioned rather than lexical: the pin-pen merger and the voicing-conditioned split in PRICE.

The accent US network broadcasting settled on, and the one most non-American learners are taught. Rhotic; TRAP and BATH are the same vowel, so 'bath' rhymes with 'math'; LOT and PALM are the same vowel, so 'father' and 'bother' rhyme; NORTH and FORCE have fallen together, so 'horse' and 'hoarse' are homophones for the large majority of speakers. Intervocalic /t/ and /d/ are both a quick tap, which is why 'latter' and 'ladder' sound alike. LOT and THOUGHT are kept apart here, but that is the contrast currently disappearing fastest in North America -- for many younger and western speakers 'cot' and 'caught' are already identical, and this preset models the conservative side of a change still in progress rather than pretending it has finished.

Traditionally non-rhotic, which is what gives 'car' and 'four' their vowel-final shape, though rhoticity has been returning steadily since the 1960s and younger speakers are often fully rhotic. The defining feature is the short-a split: TRAP raises to a long in-gliding [ɛə] in some words and stays [æ] in others, and which is which is lexically and phonologically conditioned rather than predictable -- this preset models the raised outcome on the BATH set, which is where it most reliably falls. THOUGHT is high and in-gliding, and LOT stays clearly apart from it: New York has no low back merger. TH-stopping ('thirty-third' with [t] and [d]) is variable and register-sensitive; it is on here because it is what makes the accent recognisable, not because every speaker does it every time.

'Southern US English' is an umbrella over a genuinely varied family -- Appalachian, Texan, Coastal Southern and Lowland Southern differ from each other in rhoticity, in the low back merger, and in how far the Southern Vowel Shift has gone. This preset models the Inland South profile: rhotic, with PRICE flattened to a long monophthong so 'ride' has no glide in it, GOOSE and GOAT fronted, and FACE lowered to a broad [æɪ]. Two features that define the accent for most listeners are deliberately NOT modelled here, because they are conditioned by their environment rather than by lexical set and the matrix has no place to put them: the pin-pen merger (KIT and DRESS fall together only before a nasal) and Southern breaking (a short vowel gaining an extra schwa in a stressed open syllable).

Rhotic. Its most audible feature is what happens to TH: it becomes a stop, but a DENTAL one, with the tongue against the teeth rather than the ridge -- which is why 'thin' and 'tin' stay different words here, unlike in the varieties where TH-stopping produces a plain [t]. TRAP and BATH are the same vowel; FACE and GOAT are long monophthongs; MOUTH is fronted to [æʊ] and PRICE backed. Intervocalic /t/ is often a slit fricative rather than either a tap or a stop, which this matrix approximates with flapping rather than inventing a field for a sound only one variety uses.

Cockney and Estuary English are the two ends of one London continuum, and this preset sits toward the broader end. The whole diphthong system is rotated: FACE opens to [æɪ], PRICE backs to [ɑɪ], MOUTH fronts to [æʊ] and GOAT starts from an open back position. The consonants carry the rest: /t/ is a glottal stop almost everywhere including between vowels, TH fronts to [f] and [v] (with word-initial 'that' keeping its [ð]), coda /l/ vocalizes to a [w], word-initial /h/ drops, and yod coalesces so 'tune' begins like 'chew'. Non-rhotic.

The two splits that define southern English never reached the north, and their absence is what defines this accent. FOOT and STRUT are the same vowel, so 'put' and 'putt' are homophones and 'blood' has the vowel of 'good'. TRAP and BATH are the same vowel, so 'bath', 'grass' and 'laugh' have the short vowel of 'cat'. FACE and GOAT are long monophthongs rather than diphthongs, so 'face' is [feːs] and 'goat' is [ɡoːt]. Non-rhotic. happY is lax [ɪ] here, which is the traditional northern value and the opposite of the southern and southern-hemisphere direction of change.

The accent English-as-a-foreign-language teaching outside the Americas has historically been built on. Non-rhotic, so 'car' ends on a long vowel and 'here' on a centring diphthong. It keeps the two southern-English splits: TRAP and BATH are different vowels, so 'bath' does not rhyme with 'math'; FOOT and STRUT are different vowels, so 'put' and 'putt' are different words. It merges NORTH with FORCE, so 'horse' and 'hoarse' are the same, and CLOTH with LOT. happY is given as tense [i] here, which is the modern description; conservative RP has lax [ɪ], and this is one of the changes that separates recordings from the 1950s from speakers now.

Rhotic, and with a vowel system organised on different principles from any English accent south of the border. It has no length contrast of its own: vowel length is predictable from what follows (the Scottish Vowel Length Rule), which is why nothing here is marked long. FOOT and GOOSE are the same vowel, so 'pull' and 'pool' are homophones. LOT and THOUGHT are the same vowel, so 'cot' and 'caught' are homophones -- Scotland has had the low back merger for far longer than the parts of North America now acquiring it. TRAP, BATH and PALM are all one vowel. FACE and GOAT are pure long monophthongs. The /r/ is usually a tap rather than the English approximant.

A northern vowel system with a distinctly non-northern set of diphthongs. Like Lancashire it has neither the FOOT-STRUT nor the TRAP-BATH split. Unlike Lancashire, PRICE backs and rounds toward [ɑɪ], GOAT starts open at [ʌʊ], and NG-coalescence never happened -- so 'singer' really does rhyme with 'finger', which is the single most reliable tell of the region. H-dropping is general. Non-rhotic.

Jamaican English is a systematic variety with its own vowel inventory, not a modified version of anyone else's. TRAP, BATH and PALM collapse into a single low vowel, so 'cat', 'bath' and 'calm' share it. It is non-rhotic, and where a rhotic accent has an r-coloured vowel this one has a centring diphthong: 'start' is [stɑət], 'north' is [nɔət], 'force' is [foəs]. TH is a stop, so 'thing' begins with [t] and 'this' with [d]. The weak vowels are full: commA and lettER are [a], not schwa, which is a large part of why the rhythm is heard as more even than in southern British or American English. Note that Jamaica has a continuum from Jamaican Patois (a creole with its own grammar) through mesolectal speech to acrolectal Jamaican English; this preset models the English end, and the Patois FACE and GOAT diphthongs ([ie] in 'bieks', [uo] in 'buot') belong to the other end and are not applied here.

Trinidadian English is a system in its own right, and it shares the Caribbean core with Jamaican English -- non-rhotic, TH realized as a stop, TRAP and BATH merged, full unstressed vowels -- while differing from it in the back vowels and the diphthongs. LOT and THOUGHT fall together at [ɑ], so 'cot' and 'caught' are the same word, which Jamaican English does not do. FACE and GOAT are long monophthongs rather than diphthongs. The two are grouped under one macro-region because they share a history, not because they are interchangeable.

Non-rhotic, and the whole diphthong system has rotated one step round from its southern-English starting point: FACE opens toward [æɪ], PRICE backs toward [ɑɪ], MOUTH fronts toward [æʊ] and GOAT starts central. FLEECE and GOOSE are noticeably fronted, GOOSE almost to a central [ʉː]. BATH is variable across the country and even within cities -- Adelaide favours the long vowel, Melbourne is split word by word -- and this preset takes the long-vowel outcome. happY is strongly tense, which is one of the most reliable southern-hemisphere tells. Yod coalesces, so 'Tuesday' begins like 'choose'.

Nigerian English is modelled here as a single supraregional preset over what is, in reality, substantial internal variation shaped by which of Nigeria's major indigenous languages (Yoruba, Igbo, Hausa, and many others) most influences a given speaker's English -- not represented separately here. The single most-cited defining feature across the literature is syllable-timed rhythm: unlike the stress-timed rhythm of RP or General American, Nigerian English gives each syllable roughly equal weight, a direct carry-over from the syllable-timed indigenous languages most speakers also speak (Gut 2004, in *A Handbook of Varieties of English*). FACE and GOAT are realized as monophthongs rather than diphthongs, another widely documented feature. Non-rhotic, following the British-colonial-education norm most Nigerian English descends from.

This preset models General/Cultivated South African English, one point on a real sociolectal continuum that also includes Broad South African English, Cape Flats English, and Black South African English -- each shaped by different substrate contact (Afrikaans, and indigenous languages including isiXhosa and isiZulu) and not modelled separately here; that gap is stated explicitly rather than letting this one preset stand in for all of South African English. Non-rhotic, like RP. The single most recognizable South African feature is KIT-centralization: KIT raises and centralizes toward a schwa-like vowel, which is why 'fish and chips' is the classic outsider's parody of the accent (Lass 2002; Wells 1982 vol. 3). TRAP is raised toward [ɛ], closer to where DRESS sits in RP.

Indian English is, by population, the most-spoken English variety in the world -- and modelling it as one preset necessarily flattens enormous real regional variation, shaped by which of India's many first languages (Hindi, Tamil, Bengali, Punjabi, and dozens more) most influences a given speaker; that gap is stated explicitly rather than implied away. Like Nigerian English, the rhythm is syllable-timed rather than stress-timed, and FACE/GOAT are realized as monophthongs rather than diphthongs -- both widely documented (Wells 1982 vol. 3; Sailaja 2009, *Indian English*). This preset is modelled as rhotic, the commonly cited general pattern, though rhoticity genuinely varies by region and substrate. This app's schema has no field for the single most iconic Indian English consonantal feature -- retroflex realization of /t d/ (a genuine place-of-articulation difference from the alveolar /t d/ every other English preset in this app uses, not a th_realization-type change) -- so it is documented here as a real, cited feature this engine cannot yet mechanically render, rather than approximated into a field that doesn't actually describe it.

Urban Quebec French as spoken in Montreal, the variety most sharply diverged from European French of any major French dialect -- centuries of separate development since New France's 17th-18th century settlement, not a 'French accent' in the sense Metropolitan French is usually taught as. This preset renders Quebec's four nasal vowels, its consistently uvular R, its documented faster casual-register tempo, liaison, long-vowel diphthongization (père, fête), the alveolar affrication of /t d/ before /i y/ (tu, dire), and checked-syllable vowel laxing -- the full set of features this file's earlier drafts had to leave unrendered until the engine grew the fields for them (see this file's module docstring for what's genuinely still simplified: French schwa-deletion is not modeled, and liaison approximates which contexts are near-obligatory by word length rather than by grammar).

Quebec City French: the same Laurentian innovations Montreal has -- diphthongization, alveolar /t d/ affrication, uvular R, nasal vowels -- carried at a more conservative register (Ostiguy & Tousignant 1993). This preset exists specifically so Montreal isn't the only Quebec accent in the library, the same principle this app already applies to Sao Paulo vs. Rio and Madrid vs. Seville: two presets from one region, differing in register/rate rather than in which processes exist at all. Modeled here as a somewhat more careful liaison rate (`conservative_extended` rather than Montreal's `standard`) and an unhurried tempo (`moderate` rather than Montreal's `fast`) -- Quebec City's long-standing role as the seat of government and its own 'standard Quebec French' associations plausibly correlate with a less clipped, less joual-leaning casual register than Montreal (Dumas 1987); this is a defensible modeling choice given the literature's own framing of Quebec City as comparatively conservative, not a claim that Quebec City speakers use less affrication or diphthongization -- both fire exactly as strongly here as in Montreal, because both are documented as pan-Quebec, not Montreal-specific.

Saguenay-Lac-Saint-Jean French (colloquially 'Jean-nois' or 'Bleuet' French, after the region's blueberry industry): widely cited in Quebec dialectology as one of the province's most phonologically marked regional varieties (Dumas 1987) -- not a different set of processes from Montreal or Quebec City, but the SAME Laurentian processes realized at their most pronounced. This preset models that as higher feature weights on `long_vowel_diphthongization`, `t_d_affrication` and `checked_vowel_laxing` than either other Quebec preset, rather than inventing a fourth set of matrix values -- the region's sounds are more of the same thing, not a different thing. One honestly unmodeled trait: SLSJ French is also widely described as having its own distinctive melodic/'singing' intonation contour, but `IntonationContour`'s Literal values are all named after specific Spanish-speaking regions (rioplatense_bounce, chilean_fast_fall, ...); forcing SLSJ's genuinely different melody into one of those labels would misrepresent it more than leaving `intonation_contour` at its neutral default.

Acadian French, spoken across New Brunswick, Nova Scotia and Prince Edward Island -- descended from a SEPARATE 17th-century Acadie settlement, not from the New France settlement Quebec French descends from, and phonologically distinct from it in ways this preset takes seriously rather than treating Acadian as 'Quebec French with a different accent label'. Most importantly: Acadian French does NOT have the Laurentian /t d/ affrication or long-vowel diphthongization that are hallmarks of every Quebec preset in this library (Flikeid 1997; King 2013) -- both are set off here, not reused from Montreal. Acadian French's rhotic is also historically documented as more apical/coronal than Quebec's consistently uvular [ʁ], though younger/urban speakers increasingly show uvular influence under Quebec/media contact (King 2013) -- this preset models the more conservative, historically-documented apical baseline (`r_guttural_strength="none"`). One honestly unmodeled trait, left off rather than guessed at: whether Acadian French's own high-vowel system shows a laxing pattern comparable to Quebec's `checked_vowel_laxing` is less settled in the literature available for this build than the Quebec facts are, so this field is left at its off/false value here rather than asserting a specific answer.

Chiac, the heavily English-contact variety of Acadian French spoken in and around Moncton, New Brunswick -- a genuine, productive per-word blend of French and English, not code-switching between two separate grammars (Perrot 1995; King 2013). This preset is `language="mixed"` with `mixed_languages=("fr", "en")`, the same architectural shape this app already uses for Portunol/Portuñol -- a real per-word orthographic classifier (`app.services.phonology_pal._classify_word_fr_en`) routes each word to the French or English g2p engine rather than forcing the whole phrase through one language's rules, which would have either mangled every English loanword through French orthography or mangled every French word through English orthography. Segmentally, Chiac's French base is Acadian French, not Quebec French: this preset reuses `fr-ca-acadian`'s French-side matrix values (no Laurentian /t d/ affrication or long-vowel diphthongization, a historically apical rhotic, phonemic nasal vowels) rather than inventing new ones, since Chiac's documented distinctiveness is the English-contact lexicon and code-mixing pattern, not a different Acadian French phonology underneath it. One real, named limitation: the English portion renders through `phonology_en`'s neutral reference realization only. Every accent-specific English rule in `app.services.transformer` (rhoticity, /t/ flapping/glottalization, TH-fronting, and the rest) is gated strictly on `matrix.language == "en"`, so none of them fire for English-origin words inside a `"mixed"` matrix -- this is a genuine engine gap, not something this preset's field values paper over, and it applies equally to every existing Portunol/Chabacano preset's Portuguese or Spanish side when the other language's rules would have applied. Borrowed English words keep their English stress and vowel quality rather than being fully nativized to French phonology, which is itself a documented, real property of Chiac code-mixing (Perrot 1995) -- not purely an engine artifact, even though the specific accent-flavored realization is missing.