AI Sundanese Speaking Practice: Lemes/Kasar Register, Aksara Sunda, and Bandung Fluency

Sundanese Is Not Javanese — A Critical Distinction

The most common misunderstanding among non-Indonesian learners: Sundanese and Javanese are often conflated because both originate on the island of Java and share Indonesian national language as an umbrella. In reality, they are separate languages — mutually unintelligible, with distinct vocabularies, phonologies, writing systems, and cultural traditions.

Both belong to the Austronesian language family (specifically the Malayo-Polynesian branch), but so do Malay, Tagalog, Malagasy, and Hawaiian. Being Austronesian does not make them close relatives in the practical sense any more than English and Persian being Indo-European makes them mutually intelligible. Key structural contrasts:

• Vocabulary — The core lexicon of Sundanese and Javanese overlap only modestly; a Sundanese speaker cannot understand Javanese conversation and vice versa without dedicated study
• Speech levels — Javanese has an elaborate three- (or five-) tier system (ngoko, madya, krama); Sundanese has a two-tier system (kasar/informal and lemes/formal) that is genuinely distinct in its vocabulary sets and social logic
• Script — Traditional Sundanese uses Aksara Sunda Baku, a separate script from Javanese Hanacaraka (both are descended from the Brahmic Kawi script but diverged centuries ago)
• Phonology — Sundanese has sounds not present in standard Javanese, including a distinctions between short and long vowels in certain registers and specific phonemes influenced by the Sunda Wiwitan ritual tradition
• Cultural frame — Sundanese identity is tied to West Java (Sunda region); the Sundanese regard themselves as a distinct ethnic group (suku Sunda), not a regional subgroup of Javanese culture

For Indonesian speakers learning Sundanese, this is actually a helpful reframe: you are not learning a dialect of something you already know — you are learning a second language, even if you grew up in Jakarta surrounded by Sundanese neighbors.

The Lemes/Kasar Register System

Sundanese has a speech level system that is one of its most distinctive features for learners — and one of the most practically important to master. The two primary levels are:

• Kasar (rough/informal) — The everyday register used among friends, family members of similar age, and in casual contexts. Despite the name meaning "rough," kasar is not rude when used appropriately; it is simply the unmarked register among intimates. Indonesian speakers new to Sundanese often default to kasar equivalents because they parallel the relaxed register of colloquial Indonesian.
• Lemes (refined/polite) — The register used when speaking to elders, people of higher social status, strangers in formal contexts, or in any situation where respect is required. Lemes is not merely about word choice — the entire vocabulary set shifts. Many everyday verbs and nouns have completely different lemes equivalents that share no phonological resemblance with the kasar form.

Unlike Javanese, which layers multiple intermediate levels between the informal and the most refined forms, Sundanese primarily operates on this two-level distinction in modern urban speech, though older or more traditional varieties recognize additional nuances. The Priangan highlands (Bandung, Garut, Tasikmalaya, Cianjur) are considered the heartland of refined lemes usage.

The practical challenge: using aing (kasar first person) when speaking to an elder or someone you've just met is a significant social error, while using abdi (lemes first person) with a close childhood friend sounds oddly stiff and formal. Register selection is a social skill, not just a linguistic one — it requires reading the relationship, the setting, and the other speaker's expectations simultaneously.

Sundanese Grammar: SVO, Active/Passive Voice, and Prefixation

Unlike Javanese (which tends toward a verb-medial order in certain constructions) and unlike Japanese or Korean (which are strictly verb-final), Sundanese follows a relatively clear Subject-Verb-Object word order. This makes the basic clause structure more immediately accessible to Indonesian and Malay speakers, since Indonesian follows the same SVO pattern.

Where Sundanese grammar becomes complex is in its voice system — the distinction between active and passive voice expressed through verb prefixation:

• Active voice — formed with the prefix N- (a nasal prefix that assimilates to the following consonant): maca (to read, from baca), nulis (to write, from tulis), nyieun (to make, from jieun)
• Passive voice (actor focus) — formed with the prefix di-: dibaca (read by someone), ditulis (written by someone), dijieun (made by someone)
• Causative — prefix mere- or nge- signals causing or enabling an action
• Reciprocal — prefix silih- (silih teuteup — to look at each other)

The nasal prefix assimilation rule is one of the most productive phonological processes in Sundanese and must become automatic for fluent speech. The prefix N- becomes m- before b, p; becomes n- before d, t; becomes ny- before c, j, s; and becomes ng- before g, k and vowels. Producing these automatically in real-time speech — rather than computing them consciously — is a hallmark of B1–B2 Sundanese.

Additionally, Sundanese makes extensive use of circumfix affixation — wrapping a root with both a prefix and a suffix simultaneously: ka-dahar-an (to be accidentally eaten), pa-nulis-an (a writing tool / place of writing). These circumfixes interact with the register system, producing a grid of forms that learners need to build up through extensive listening and speaking practice.

Aksara Sunda Baku: The Revived Traditional Script

Sundanese has a traditional script — Aksara Sunda Baku(Standardized Sundanese Script) — that was codified and revived in the 1990s and is now officially recognized and taught in West Java schools. The script is an abugida descended from the ancient Kawi script (the same Brahmic ancestor as Javanese Hanacaraka and Old Balinese script), but it diverged sufficiently over centuries to be considered a distinct writing system.

The Aksara Sunda revival was a deliberate cultural reclamation project. During the Dutch colonial period and the early decades of Indonesian independence, Roman script (using the same Latin alphabet as Indonesian) became dominant for everyday Sundanese writing. By the late twentieth century, Aksara Sunda had become a historical curiosity known mainly to scholars. The 1990s standardization effort — led by Sundanese language and cultural organizations — created a unified standard and pushed for its inclusion in Unicode (the script was included in Unicode 5.1 in 2008) and in the West Java provincial curriculum.

Today, Aksara Sunda appears on road signs in West Java, in cultural materials, in academic publications, and in digital contexts (the Unicode block includes both the original and a supplementary set for Sanskrit-origin words). However, for everyday writing — text messages, social media, informal correspondence — Roman script is used overwhelmingly. This means:

• Script literacy is optional for most learner goals. You can reach full oral fluency and handle everyday written Sundanese (Roman script) without knowing Aksara Sunda. Heritage learners reconnecting with family, researchers, and those with cultural preservation goals will benefit most from script study.
• AI voice practice is script-agnostic — speaking and listening fluency develop independently of reading the traditional script. This means the most time-sensitive oral fluency goals can be pursued in parallel with (or before) script study.
• Reading competency in Roman Sundanese is the practical baseline — understanding how Sundanese vowels are romanized (including the schwa eu and the open-mid é) matters for learners who study from textbooks or online resources.

Sundanese Phonology: Key Sounds for Learners

Sundanese phonology has several features that are challenging for Indonesian speakers (whose national language has a relatively shallow vowel inventory) and for learners from other backgrounds:

• The schwa vowel (eu) — Written eu in Roman Sundanese, this is a mid-central vowel similar to the schwa in English "the" but slightly more rounded and phonemically distinct. It does not exist in standard Indonesian. Minimal pairs like beu (full/satiated) versus bo (empty) test perception quickly. Consistent mispronunciation of eu as e or o is a reliable marker of non-native speech.
• The é vowel — A distinct mid-front vowel (like Frenché), separate from the schwa. Written é in careful romanization but often written plain e in casual contexts, causing learner confusion when the schwa eu and the é are both represented inconsistently.
• Glottal stop (hamzah) — Word-final consonants in Sundanese, particularly the glottal stop at the end of syllables, are phonemically contrastive. The difference between basa (language) and basa' (wet, with glottal stop) illustrates this contrast.
• Bilabial trills and uvular variants — Regional varieties show variation in the realization of the r phoneme; Priangan Sundanese tends toward a uvular or retroflex realization in certain environments, while coastal and Banten varieties may use a simpler flap.

For Indonesian speakers, the most important training targets are the eu schwa and the register vocabulary. For English-background learners, both the phonology and the register system require focused attention from the start.

Regional Varieties: Priangan, Bandung, and Banten

Sundanese is not uniform across its territory. The main regional varieties reflect geography, historical settlement patterns, and contact with neighboring languages:

• Priangan Sundanese — The variety of the highland interior (Bandung, Garut, Tasikmalaya, Sumedang, Ciamis, Cianjur). Considered the prestige variety, closest to the standard taught in schools, and the source of the refined lemes register associated with Sundanese cultural identity. The Priangan highlands were the heartland of the Dutch colonial Preanger system (forced coffee cultivation), which paradoxically preserved strong Sundanese local culture by creating a relatively stable landed class with high cultural investment.
• Bandung urban Sundanese — Bandung (population: 2.5M city, 8M+ metro) is the dominant cultural and economic center of West Java. Urban Bandung Sundanese shows strong Indonesian language influence, English borrowings from the tech and creative industries, and a younger speaker style that compresses lemes distinctions and favors a more mixed register. Bandung is also Indonesia's fashion and music capital — the city's distro (independent clothing label) culture and indie music scene produce extensive creative Sundanese language content.
• Banten Sundanese — Spoken in the western province of Banten (separated from West Java province in 2000) and in the western coastal areas. Shows heavier influence from Betawi (the Jakarta creole variety), older Malay contact, and the distinct historical experience of the Banten Sultanate. Phonologically and lexically diverges from Priangan standard in ways that can cause comprehension difficulty between Banten and Priangan speakers in rapid informal speech.
• Northern coast Sundanese — Areas along the Java Sea coast (Cirebon border zone, Karawang, Bekasi) show significant Javanese contact, as Cirebon historically belonged to a transitional cultural zone between Sundanese and Javanese kingdoms. Cirebonese (Basa Cerbon) is sometimes classified as a Sundanese-Javanese mixed variety.

The Baduy people — a traditional community in the Kendeng mountains of Banten who follow strict adherence to Sunda Wiwitan, the indigenous Sundanese belief system — speak a variety of Sundanese that is considered particularly archaic and conservative. The Baduy divide into Baduy Dalam (inner Baduy, who reject all modern technology) and Baduy Luar (outer Baduy); their Sundanese is studied by linguists as a window into historical Sundanese grammar before heavy Indonesian influence.

Cultural Context: Gamelan Degung, Pencak Silat, and Bandung Identity

Understanding Sundanese cultural context makes language practice richer and more memorable. Key touchpoints:

• Gamelan Degung — The Sundanese gamelan tradition uses a distinct pentatonic scale (pélog degung) that differs from Javanese gamelan tuning systems, producing a recognizably different sonic character — more lyrical and melancholic in aesthetic feel. Degung music and its associated vocal form tembang Sunda (Sundanese art song) use refined lemes language in their lyrics, making them an excellent source of formal vocabulary in emotional and poetic registers.
• Pencak Silat — The Sundanese martial art tradition is one of the most developed in Indonesia; many of the foundational pencak silat schools originated in West Java. The martial arts community maintains its own vocabulary domain with specific terminology for techniques, training relationships, and school hierarchies — vocabulary that is culturally significant and practically useful for any Sundanese cultural engagement.
• Wayang Golek — The Sundanese wooden puppet theatre (distinct from Javanese wayang kulit leather puppets) uses an elaborate formal literary Sundanese in the narration and dialogue of classical characters, while clown characters (punakawan) typically speak informal kasar Sundanese with comic effect. Wayang golek is the clearest illustration of register contrast in Sundanese performance tradition.
• Sunda Wiwitan — The indigenous Sundanese religious tradition, still practiced by the Baduy community and with survivals among other Sundanese groups. Distinct from Javanese abangan (syncretic Islam-Hindu) traditions; Sunda Wiwitan maintains a connection to the pre-Islamic Sunda Kingdom period. Its ritual language is an archaic Sundanese variety studied by scholars of Old Sundanese.
• Bandung as creative capital — Bandung hosts Indonesia's largest concentration of indie musicians, fashion designers, and creative entrepreneurs. The city's creative scene has generated a distinctly Bandung Sundanese aesthetic — youth language, brand names, and social media content in a mixed Sundanese-Indonesian-English register that is now exported culturally to the rest of Indonesia.
• Jawa Barat provincial identity — West Java is Indonesia's most populous province (50M+ people), and Sundanese identity within it is politically and culturally significant. The province has consistently pushed for greater recognition of Sundanese language in education and official contexts, a source of ongoing cultural politics between Jakarta-centric national identity and regional Sundanese distinctiveness.

Learner Profile: Who Studies Sundanese and Why

Sundanese has a distinctive learner profile that differs from most major language-learning communities:

• Indonesian speakers learning Sundanese (reverse immersion) — Many Indonesians — especially those from Jakarta and other non-Sundanese regions — have grown up surrounded by Sundanese cultural references and Sundanese-origin colleagues without learning the language itself. Indonesian (which has strong Malay and some Javanese features in its standard form) does not transfer well to Sundanese. These learners often underestimate the difficulty because they assume shared vocabulary, then are surprised by how different the core lexicon is.
• Heritage learners — Sundanese families in Jakarta, Surabaya, and the Netherlands (the Netherlands has a significant community of Indonesian-origin residents, including Sundanese) often maintain passive Sundanese from family use but lack productive fluency. Heritage learners may recognize lemes vocabulary from grandparents' speech but feel more comfortable in Indonesian or Dutch for active production.
• Indonesia researchers and anthropologists — Academic interest in Sundanese frequently comes from scholars of Indonesian religion (Sunda Wiwitan, Islamic practice in West Java), performing arts (wayang golek, tembang Sunda, pencak silat), colonial history (the Preanger system), and environmental anthropology (the Baduy community, highland agriculture, Ciliwung river watershed).
• People with West Java professional connections — Bandung is a growing tech hub (multiple tech unicorns have significant West Java operations), a manufacturing center, and the gateway to Priangan agricultural export regions. Business relationships in West Java are meaningfully enhanced by even basic Sundanese greetings and cultural vocabulary.

The Sundanization of Modern Vocabulary

One of the most interesting contemporary features of Sundanese is what linguists call Sundanization — the process by which modern Sundanese speakers adapt Indonesian and English loanwords into Sundanese phonological and morphological patterns, rather than simply borrowing them wholesale. This reflects a linguistic confidence and creative identity in the Sundanese speech community.

Examples of Sundanization in everyday Bandung speech:

• Indonesian sekolah (school) → Sundanese sakola (with vowel shift following Sundanese phonotactics)
• English smartphone → adapted with Sundanese verb prefix to produce make smartphone-na (using his/her smartphone) in casual Bandung speech
• Technical vocabulary from Indonesian national language news media is absorbed but often given Sundanese phonological shape or Sundanese affixation when used in Sundanese-frame conversation

This Sundanization process means that Sundanese is not simply being replaced by Indonesian — it is actively adapting and incorporating modern vocabulary into its own morphological system. Learners at B2+ benefit from understanding this process because it allows them to predict how new vocabulary will behave grammatically.

Practice Scenarios by Level

A1–A2: Greetings, Register Basics, and Core Vocabulary

B1–B2: Register Navigation and Bandung Life

C1+: Literary Sundanese, Wayang Golek, and Cultural Depth

Advanced Sundanese practice engages with the literary and performance traditions that represent the highest register of the language. Tembang Sunda (Sundanese art song) uses a specialized poetic language with archaic vocabulary, fixed melodic forms, and a level of lemes refinement that goes beyond everyday formal speech — the singer-poet tradition of Cianjuran tembang draws on Sanskrit-derived vocabulary that has been absorbed into Sundanese over centuries of contact with Hinduism and Buddhism before Islamization.

Wayang golek narration at C1+ level involves the dalang's (puppeteer's) formal narration style (carita), which uses a highly inflected, syntactically complex Sundanese with full deployment of the honorific vocabulary system. The clown scenes provide contrast — spontaneous kasar Sundanese that reflects contemporary slang and regional jokes. Exposure to both within a single performance session is one of the most intensive register workouts available in Sundanese.

Old Sundanese (Bahasa Sunda Kuno), attested in inscriptions and manuscripts from the Sunda Kingdom period (7th–16th centuries), is accessible to C1 learners with linguistic training and is the domain of academic researchers. The Carita Parahyangan (a 16th-century Sundanese historical text) and Sanghyang Siksa Kandang Karesian (1518, a moral treatise) are the canonical Old Sundanese texts. Sunda Wiwitan ritual poetry in archaic Sundanese represents a related but living tradition.

Why AI Multi-Persona Practice Works for Sundanese

The lemes/kasar register system — more than any other feature of Sundanese — is where a multi-persona AI voice room has structural advantages over single-tutor or textbook study:

• Register-specific AI personas — Configure one AI voice as a formal elder speaking full lemes (building your productive formal vocabulary and receptive recognition of polite register), and a second AI voice as a Bandung peer in kasar (building the informal register you need for social life and family interactions). Switching between conversations in the same session is the fastest route to register flexibility.
• Nasal prefix drilling in context — A patient grammar tutor persona can flag when you produce the wrong nasal assimilation in real time — immediately after the error, while the conversational context is still active. This is faster than post-session error review and produces stronger correction.
• Cultural scenario practice — Role-play a traditional Sundanese family gathering (formal lemes with grandparents, mixed with cousins), a Bandung angkot (minibus) ride (informal kasar with the driver), or a wayang golek post-show discussion with a dalang. The AI adapts vocabulary and register to match each scenario.
• Phonological feedback on eu — AI voice practice provides immediate listening exposure to the schwa eu in natural Sundanese speech; repeated exposure builds perceptual categorization of this phoneme faster than rule memorization alone.
• Heritage learner bridge — Heritage learners who understand lemes passively from grandparents but default to Indonesian in production can use AI practice to push that passive knowledge into active speech — without the social anxiety of speaking imperfect Sundanese to real family members.
• Flexible scheduling — 30-minute daily sessions fit diaspora learners in the Netherlands or Jakarta who are reclaiming heritage language alongside work and family obligations; no need to coordinate time zones with a tutor in Bandung.

Getting Started with Sundanese

Personaplex is free to try — 30 minutes of voice chat per day, no credit card required. For Sundanese learners, the recommended starting configuration is:

• AI Persona 1 (formal elder) — Set to full lemes register, speaking as a traditional Sundanese elder from the Priangan highlands. Focus your first three sessions on the pronoun system (abdi / Anjeun) and the core lemes verb set (tuang, kulem, sumping, angkat). Getting the formal register grounded first gives you the prestige baseline; informal register is easier to learn afterward.
• AI Persona 2 (Bandung peer) — Set to informal kasar register, speaking as a Bandung university student. Use this persona for the everyday conversational vocabulary, city navigation, food and music discussion. The contrast between the two personas in the same session makes register differences viscerally clear in a way that no table in a textbook can achieve.

Begin with the greeting set (Wilujeng enjing / siang / wengi) and the food vocabulary, then introduce the pronoun system in session two. By session five, attempt your first deliberate register switch mid-conversation — tell AI Persona 1 something in lemes that you just discussed with AI Persona 2 in kasar. The gap between the two vocabulary sets will become concrete rather than abstract, and that concreteness is what turns memorized word lists into functional register awareness.