10 Common Transliteration Mistakes and How to Avoid Them
The most frequent errors people make when typing Indian languages phonetically — aspirated consonants, vowel length, schwa deletion, retroflexes, and more — with clear examples and fixes.
Transliteration is incredibly useful for typing Indian languages, but it's not foolproof. The conversion from Roman phonetic spelling to Indian scripts depends on you giving the tool enough information to make the right choice. Feed it ambiguous input and you get ambiguous output.
Most mistakes fall into predictable categories. Once you know what to watch for, your accuracy improves dramatically with very little extra effort.
1. Ignoring Aspirated vs. Unaspirated Consonants
This is the single most common source of incorrect output. Indian scripts (and Indian languages) distinguish between aspirated consonants (produced with a puff of air) and unaspirated ones. English doesn't mark this distinction consistently, which is why English speakers routinely miss it.
Consider the pairs:
- क / ক / க (ka, unaspirated) vs. ख / খ / க² (kha, aspirated)
- ग / গ (ga, unaspirated) vs. घ / ঘ (gha, aspirated)
- प / ப (pa) vs. फ / ফ (pha)
- ब / ব (ba) vs. भ / ভ (bha)
In transliteration tools, the convention is usually: single letter = unaspirated (k, g, p, b, t, d, c/ch), letter + 'h' = aspirated (kh, gh, ph, bh, th, dh, chh). The mistake: Typing "bharat" for भारत is correct. Typing "phal" for फल is correct. But people often type "pal" intending फल, or "ghar" intending घर when they should type "ghar" — which, in this case, is actually correct. The issue is with less obvious words. Example of where it goes wrong: "Priya" (प्रिया) is correct. But "Priti" typed as "Priti" often produces प्रीती correctly... however "Preeti" typed as "Priti" without the double-e might produce something unexpected. The fix: be deliberate with your 'h' after stop consonants. Fix: When a word contains kh, gh, th, dh, ph, bh — type those explicitly. "Pankh" not "Pank". "Badhai" not "Badai". Most good transliteration tools are forgiving if you add the 'h', less so if you omit it.
2. Getting Vowel Length Wrong
Indian languages distinguish between short and long vowels — a short "i" (इ) and a long "ii" or "ee" (ई) are different sounds that produce different words. Same for "u" (उ) vs "uu/oo" (ऊ).
Examples where getting this wrong changes the word:
- दिन (din — day) vs. दीन (deen — poor/humble)
- पुल (pul — bridge) vs. पूल (pool — pool)
- मिल (mil — mill/met) vs. मील (meel — mile)
The mistake: Typing single vowels when a long vowel is needed. Writing "din" for दीन or "mil" for मील. Fix: Use doubled vowels for long sounds: 'aa' or 'a' depending on context (many tools accept both), 'ii' or 'ee' for ई, 'uu' or 'oo' for ऊ. Different tools use slightly different conventions — check TranslitHub's guide for its specific mapping. The rule of thumb: if the vowel is held for longer when pronounced, double it in your input.
3. The Retroflexes Problem
Indian scripts have two distinct sets of sounds that both map to English 't' and 'd':
- Dental त, द — tongue at teeth
- Retroflex ट, ड — tongue curled back
English uses only dental sounds, so English speakers don't naturally distinguish these in writing. But in Hindi and other Indian languages, the wrong one produces a different word or an unnatural spelling.
Examples:
- पटना (Patna, city) uses retroflex ट
- पतन (patan — fall/decline) uses dental त
- दिल्ली (Delhi) — the ल्ल here matters, and the 'D' is retroflex ड
The mistake: Typing "patna" and getting पतना instead of पटना, or vice versa. Fix: Most transliteration tools use capitalization to distinguish these: capital T/D for retroflex (ट/ड), lowercase t/d for dental (त/द). So "paTna" = पटना (Patna). "patan" = पतन. Learn this convention — it covers a significant source of errors.
Similarly: capital N vs. lowercase n can distinguish ण (retroflex nasal) from न (dental nasal). "raN" vs. "ran" produces different outputs.
4. Schwa Deletion
This is a phonological rule of Hindi (and several other North Indian languages) that doesn't exist in Sanskrit. The schwa is the inherent 'a' vowel present in every Devanagari consonant. In natural Hindi pronunciation, schwas at the end of words or in certain internal positions are deleted — not written, but not spoken either.
The classic example: राम is romanized as "Raam" or "Ram" but pronounced "Raam", not "Raama". कमल (Kamal) is pronounced "Kamal", not "Kamala". But कमला (Kamala) — with an explicit long ā matra — is pronounced "Kamala".
The mistake: Typing "Raama" and getting राम (correct) with no problem, OR typing "Rama" thinking it means राम but getting रम (a different word meaning "pleasing" or a proper noun). Fix: Understand schwa deletion for Hindi specifically. The written form doesn't always match what you'd naturally romanize from pronunciation. For proper nouns especially, look up the standard spelling — "Laxman" vs "Lakshman" vs "Lakshmana" all map to the same person but transliterate differently.5. Confusing Short 'a' and Missing 'a' Entirely
In transliteration input, there's a difference between:
- "ka" → क (consonant with inherent 'a')
- "k" before another consonant → क् (consonant with virama, no vowel)
- "kaa" or "ka" with context → का (consonant with explicit long ā)
The mistake: Typing "akbar" intending to get अकबर (Akbar) but getting unexpected output because the tool interprets the initial 'a' differently. Fix: For words starting with a vowel, most tools handle initial vowels correctly. The issue arises in consonant clusters. If you get an unexpected conjunct or halant, check whether you've implied an 'a' vowel in the wrong place. Read your output carefully and correct cluster formations manually.
6. The Nukta Problem
Several Hindi characters use a nukta — a dot below the character — to represent sounds borrowed from Persian and Arabic: ज़ (z), ख़ (kh as in loch), ग़ (gh), फ़ (f), क़ (q).
The mistake: Typing "zaroor" and getting ज़रूर (correct, with nukta) OR getting जरूर (without nukta, technically incorrect but common). Many texts and typists drop the nukta.This matters most for: proper names (ग़ालिब, not गालिब), loanwords like ज़िंदगी (life), ख़ुशी (happiness), क़िताब (book).
Fix: Check your transliteration tool's nukta handling. TranslitHub and most modern tools produce nukta characters when appropriate. But when accuracy matters (poetry, formal documents, personal names), verify the output visually.7. Anusvara vs. Full Nasal Consonant
The anusvara (ं) is a dot above the character representing a nasal. It can replace a full nasal consonant before another consonant in certain positions. Both are technically valid but not always interchangeable.
Compare:
- हिंदी or हिन्दी — both spellings of "Hindi" are used, but हिंदी (with anusvara) is more common in contemporary usage
- संस्कृत or सन्स्कृत — similarly
The mistake: Accepting the anusvara form when your context requires the full consonant form (or vice versa). Some formal and educational contexts prefer the full nasal consonant spelling. Fix: For personal and informal use, either is generally fine. For formal documents or academic writing, follow the stylistic convention of your institution or publication. In most practical transliteration for everyday use, the anusvara form is correct and widely accepted.
8. Half-Letter vs. Full Conjunct Formation
In Devanagari, conjuncts can form in different ways. Some consonants take a "half-letter" form (the right side is removed), others use a subscript form below the following consonant, and others use a fully merged ligature.
The mistake: Not understanding why the output looks different from what you expected. Type "karma" and you might expect क + र to form a visible conjunct, but Devanagari often writes this as क्र (k + ra, using a subscript ra). Type "dharma" and ध + र forms ध्र.This isn't really an error in the output — it's the correct form. The issue is when learners think the output is wrong because it looks different from what they visualized.
Fix: Trust the Unicode rendering for common conjuncts. The display depends on which font is being used. If you see boxes or unexpected characters, the font doesn't support the conjunct — switch to a Unicode-complete font like Noto Sans Devanagari.9. Wrong Handling of 'ri' / 'ri' vs. 'ri'
The vocalic 'r' (ऋ) is a specific Sanskrit vowel that appears in common Hindi/Sanskrit words: ऋषि (rishi, sage), कृपा (kripa, grace), कृष्ण (Krishna), संस्कृत (Sanskrit).
The mistake: Typing "rishab" and getting रिषभ instead of ऋषभ, or typing "kripa" and getting क्रिपा instead of कृपा. Fix: For the vocalic ṛ sound, use "ri" followed by a consonant in most tools (the context distinguishes it from consonant-r combinations). Some tools use "Ri" (capital R) or "R" as a specific code for ऋ. Check your tool's documentation. For common Sanskrit-derived proper nouns, know the correct form and verify: कृष्ण for Krishna, कृपा for kripa, ऋषि for rishi.10. Treating All Indian Languages Like Hindi
Different Indian languages have different transliteration conventions and different sound inventories. A transliteration that works perfectly for Hindi may produce wrong output for Tamil, Telugu, or Kannada.
The mistake: Using a Hindi-mode transliteration for Tamil text. Tamil doesn't have aspirated consonants — so 'kh' in Tamil context doesn't produce ஃ or a special character; it may produce incorrect output. Tamil also doesn't have the retroflex/dental distinction in the same way, and has unique sounds like the retroflex lateral (ழ, the "zh" in "Tamil" itself, in "Tamizh").Similarly, Malayalam has a rich set of conjuncts and aspirated sounds that differ from Hindi's phonological system.
Fix: Always select the correct target language in your transliteration tool before typing. TranslitHub has language-specific modes — use the right one. Don't assume Hindi phonetics apply universally.Quick Reference: Common Mappings Across Tools
| Sound | Typical Input | Devanagari | Notes |
|---|---|---|---|
| Long aa | aa or A | ā (मात्रा ा) | "baad" = बाद |
| Long ii | ii or ee or I | ī (मात्रा ी) | "beet" = बीत |
| Long uu | uu or oo | ū (मात्रा ू) | "phool" = फूल |
| Aspirated k | kh | ख | "khet" = खेत |
| Aspirated g | gh | घ | "ghar" = घर |
| Retroflex t | T (capital) | ट | "paTna" = पटना |
| Retroflex d | D (capital) | ड | "DaakGhar" = डाकघर |
| Retroflex n | N (capital) | ण | "raN" = रण |
| z sound | z or Z | ज़ | "zaroor" = ज़रूर |
| Vocalic r | ri or R | ऋ | "Rshi" = ऋषि |