When you dig into the Nepali Script (and the romanized version) more and more, you will start to notice disparities in transliterations. This time, I want to set things straight in Transliteration and this will be the first step to understanding how people write phrases in English and the vernacular uses of the language. This lesson is aimed to clear any misunderstandings in transliterations. 

Transliteration comes from the word ’trans-’ which means ’Across’ and the Latin ’littera’ which means ’Letter’. You can guess that transliteration means to convert ‘letters’. Hence, Transliteration is the process of representing the words or letters of one language into another language using the nearest corresponding letters.

First thing here, I will consider Transliteration and Transcription to be the same (I know the are not!). This should now keep things simple and ‘Transcription’ will be ‘Transliteration’ unless otherwise mentioned! The other thing is, transliteration will be strictly into the Roman Script ! *unless otherwise mentioned*

Transliteration is concerned with spellings, representing foreign letters as they appear. Transcription is concerned more with the sound than spelling.


Nepali uses the Devanagari Script for writing. Attempts to transliterate Devanagari (which is used by languages like Hindi, Sanskrit etc.) resulted in a quite a few ways to represent Devanagari in the Roman Script. One or the other system is used in transliterations of Scholarly articles. Some of them are IAST, Hunterian etc. I won’t be listing them because you can see them in this article anyway.


Transliterations are important for a learner; they help a learner to understand words and letters until they can no longer depend on it. However, Nepali people and those who know Nepali in general usually type transliterations than Nepali text. So, a question arises, why do we write like ’ke gardai’ instead of the more natural ‘के गर्दै’ while talking in the cyberspace? 

An answer to that is, first…nepali has too many sounds. Also, everything around us is in English. We also simply lack Nepali Keyboards. There isn’t a ‘default’ installation of the Nepali Language Pack either. So, that leaves us with English keyboards with English Software. Another reason is, Nepali is way too complicated to type and even if we had a keyboard, it would take quite some time to type like da da da daaa...


What are informal transliterations? These are translations which an individual things is an appropriate substitution for the Nepali equivalent. However, there are various disparities which I will now talk about.

First exercise, spell these vowels: A E I O U

Did you? Now take this word, for example: किरण

It spells kiran, right? Well, that is what I have told you if you have been reading from this site for some time. Well, what if I told you that people type alternative spellings as well?

That i in ‘ki’ sounds like the english alphabet ‘E’ doesn’t it? So, some people type ‘kiran’ as ‘keran’. I have one friend who according to me is chronically affected by this. He types every word where there is supposed to be an ‘i’ with an ‘e’. So, words like ‘kira’ and ‘pidit’ become ‘kera’ and ‘pedet’. But the question is, is he incorrect? 

I would say no, but as the reader I would find it pretty difficult to read it. He is not incorrect because for him, these are as normal as words were to Shakespeare (what a poor analogy). For him, these letters appear normal.

There are also times when you see that आ and अ are often represented with a single ‘a’. For example, you might see ‘khola’ and wonder whether it is ‘open’ or ‘river’. For this, we must now depend upon intuition and context. Are they talking about things or sceneries? If they say ‘khola badhyo’, they obviously mean ‘the river grew’. However, if they say ‘gift khola’, then they mean ‘open the gift’. We must also note that river is खोला (more correct kholaa) while ‘open!’ is खोल (correctly, khola).

Of course, sometimes context can fail. *profanity ahead* There is a joke in Hindi (translated into english except the part which is important) which says:

A: You’re always online. Chutiya hae kya?
B: Excuse me? You must be the chutiya here.
A: What? I meant holidays!

Transliteration and context fails above. The bold part ‘chutiya hae kya’ can be read as ‘चुतिया hae kya’ which means more or less ‘Are (you) an idiot’. However, what ‘A’ wanted to say was ‘छुट्टियां’ which means ‘holidays’. Both can be transliterated into ‘chutiya’ hence the miscommunication. Unfortunately, context twisted the very intended meaning. 

Therefore, we must never be certain on what the transliteration is trying to say. A common sentence can begin like this:

खाना भो (khānā bho)

Which could be transliterated into:

  • khana vo
  • khaanaa bho
  • khana bho
  • khaana bho

and so on. There are limitless possibilities. If you notice, the ‘v’ sound mimics the ‘bh’ sound closely, so some people substitute ‘bh’ for ‘v’ (like vaivarkhar etc.) Similarly, both the त and ट series are represented by ‘t/d’ series so transliterations never really say what is retroflex and what is dental (unless you use IAST or Devanagari). For example, ‘taal’ could be the dental ताल (= lake) or the retroflex टाल्  (= stick it).

All in all, there is no one standard that everyone adheres to. You stick to one that you follow and hope others understand you as well. With experience, you must build the ability to read transliterations appreciably. Here are some observations I have noted down (but exceptions exist):

  • The trailing ‘a’ is usually an ‘आ’ sound (like father). Examples: asina (असिना = hail), gharra (घर्रा = drawer) etc.
  • Some people write ऐ (ai) and आइ/आई (aai) as ‘ae’. So words like gardai (गर्दै) and malai (मलाई) become gardae and malae respectively.
  • Some people (read: teenage wannabes) think substituting ‘l’ (l for Love) for ‘r’ is cute and trendy. So don’t panic when you sometimes see ‘melo’ and ‘timlo’ and don’t find in the dictionary. It simply means ‘mero’ and ‘timro’ (which means ‘mine’ and ‘yours’ respectively) [mero = मेरो | timro = तिम्रो]
  • The greatest discrepancy is between च and छ. IAST lists them as ca and cha respectively. Some people write cha and chha but most people make no distinction. This is quite tricky. Words like chucho doesn’t help either. Does it mean ‘sharp’ (चुचो) or ‘an ill-mannered person’ (छुचो)? When this happens, always look at the context. I am certain people are not asking for ill-mannered knives or a sharp person!
  • Some people write फ as ‘fa’. While not incorrect, ‘pha’ is more correct.
  • The three s’ श  ष  स are simply represented by ‘sa’ while some people may use ‘sha’ for the first two.
  • The letter व can be read as ‘va’ or ‘wa’. It is more common for people to use ‘ba’ instead of ‘va’. 
  • The nasal markings ं and ँ are often written using ‘n’ but in case of ँ, it is often omitted altogether. For example, बाँस (baas) means ‘bamboo’ and is pronounced with a nasal on ‘ba’. However, people write ‘bans’ as well to differentiate it from बास (baas) which means ‘place of stay’.


Nepali Word (meaning)
Common transliterations [Formal IAST in the leftmost]

छ (is)
cha/ chha /xa /6 [6 because it is prononunced ‘chha’ in Nepal]

के (what)
ke/ k

Any syllable that ends with ‘a’ sound (like ago) could be represented as x-a/ah/oh/uh. For example, ta/tah/toh/tuh …

केटी (girl)
keṭī/ keti/ kti/ kt

केटा (boy)
keṭā/ keta/ kta

हिमाल (mountain)
himāl/ himal/ himaal



1) घारी (grove)

2) तीर (arrow)

3) भाडा (rent)

4) भाँडा (utensil)

ANSWERS (in IAST/ some alternatives) (illustrative only)

1) ghārī/ ghari/ ghaari/ ghaaree

2) tīr/ teer/ tir

3) bhāḍā/ bhaadaa/ bhaada/ bhada

4) bhām̐ḍā/ bhada/ bhanda/ bhaadaa

Schwa Syncope


Over the past few lessons you may have noticed that some sounds went silent, especially those around the end. Although Devanagari, the script Nepali uses, is a phonetic script with no silent letters (unlike English), the sound in the end vanishes away like it never existed, although it seems that the sound should really exist. This sound is known as ‘schwa’ and the phenomenon of it being deleted from pronunciation is called ‘Schwa Syncope’. It is also known as Schwa Deletion Phenomenon.

First, let’s see what a schwa really is. Take an example word like ghām, which means ‘Sun’ in Nepali. It is written as घाम in Devanagari, which ideally should be pronounced as ‘ghāma’ (according to the spelling) but somehow, the last ‘a’ simply vanished. This unstressed vowel is called a schwa and is usually represented as ‘ə’ in the IPA (International Phonetic Alphabet). Technically, a schwa in English phonology is an unstressed vowel in the mid-central region, but we will nonetheless still call the end unstressed vowel “schwa”, mostly because we are focused on Nepali phonology rather than English phonology. Specifically, schwa here refers to the end vowel sound made by अ (a), because it is the only one that is affected by this phenomenon. Other vowel sounds are not deleted.

Syncope /ˈsɪŋkəpi/ means deletion, or rather omission of certain sounds from a word. An example of this would be the omission of ‘a’ from the pronunciation of library /ˈlʌɪbri/.


Now, you may say “Why not use a halanta instead to delete the schwa?” This is an interesting question, but the reason why we don’t is because of history. Nepali came from Prakrit, which itself came from Sanskrit. Sanskrit used Devanagari to write, much like Nepali, but it did not delete the schwa from the words. Thus, घाम would be read with the schwa i.e. ghāma. It is only when things went awry in Prakrit, when the final ‘a’ (or the schwa of the word) was deleted for unknown reasons (perhaps, for convenience). We’ll never know. What we do know is that the change in pronunciation did not reflect upon the word, thus we still retain the spelling for words that originally had a schwa. 

Another reason is that it makes separation of words easier. Morphemes, or unit of words, usually retain their schwa-less pronunciation while building up bigger words. Thus, one can easily glance at the large word and guess what it means by looking at the morphemes used. Another reason may be because using a lot of halanta looks really crowded, thus for aesthetics this is not done. 

As an evidence, many languages that use Devanagari as the script delete the schwa from their words. This is more or less pronounced, with Nepali being a bit towards the lesser side. Nepali is way more phonetic than say Hindi, because the original sound of Sanskrit is somewhat more preserved than in other languages.


It is very important to be able to delete schwa appropriately if you want to read properly. As said before, only the vowel sound of अ (a) is affected by this phenomenon. Do not delete the other vowel sounds! 

I will assume you have encountered terms like ligatures and consonant clusters in the previous lessons. If you have not, click here. In brief, ligatures are characters that represent two or more sounds, much like the German ‘ß’ (Eszett). 

For this lesson, I will separate each syllable with an interpunct (·). I have separated the consonant with the schwa as well, but take care not to pronounce that ligature’s schwa.


If a consonant (or a consonant cluster) is modified by a vowel sound in a word (other than the default ‘a’ sound) to form a syllable, you do pronounce that syllable. Example:
मुला /radish/ = mu · lā

Consonants marked with a halanta are not pronounced with any vowel sound. Example:
सन् /CE/ = san (not sa · na)

Generally retain the schwa of all characters but the final one. Example:
कलम /pen/ = ka · la · m (not ka· la· ma or any other variation like kla · ma

Now that we have these two rules in mind, let’s learn how to delete that pesky ‘a’ sound efficiently. Note that the below guidelines apply only to words that have the schwa in the end. Also, I have included the deleted schwa [ə] in square brackets. Do not include [ə] in your pronunciation! I have only included it for reference. For example, read rā ·m[ə] as Rām and not Rāma.


Schwa of independent characters are retained

A standalone character’s schwa is always spelt out, unless dictated otherwise by a halanta. Examples:
म /I/ = ma
ल /okay/ = la

Schwa of nouns and adjectives are usually deleted

The schwa of nouns are usually deleted, but exceptions can arise if the word was imported fairly recently. Examples:
नेपाल /Nepal/ = ne · pā · l[ə]
हिमाल /mountain/ = hi · mā · l[ə]
नरम /soft/ = na · ra · m[ə] 

Schwa of verbs are retained

Generally, every syllable in a verb (and its conjugates) is pronounced. This is because the verb relies on absence or presence of the schwa to indicate different moods. Examples:
गर /do/ = ga · ra
गर् /do/ = ga · r (notice the lack of schwa as it is removed by the halanta)

Schwa of ligatures are usually retained

When a schwa is present in a ligature, the schwa is usually spelt. Exceptions are usually imported words; native words almost always pronounce the schwa of ligatures. Examples:
वाक्य /sentence/ = wā · kya
पूर्व /East/ =  pū · rwa 
साहित्य /literature/ = sā · hi · tya 
पत्र /newspaper/ = pa · tra 

रङ्ग /colour/ = ra · ṅg[ə] [Some people don’t omit the schwa, so this is in grey area]

Schwa of grammatical functions are retained

Grammatical markers and functions like case markers retain their schwa. Examples:
पर /across/ = pa · ra
बाट /from/ = bā · ṭa 

State of schwa of words persist even if the word gets modified/ concatenated

Addition of case markers, adverbs, adjectives, postpositions etc. do not affect the retention (or lack thereof) of the schwa of the previous word. They function independently and do not affect each other. Examples:
नेपालबाट /from Nepal/ =  ne · pā · l[ə] – bā · ṭa
घरलाई /to house/ = gha · r[ə]  – lā · ī 

This also means that, when words are joined together to make larger words, each component retains their original pronunciation. Examples:
एकदिन /one day/ = e · k[ə] · di · n[ə]
आजभोलि /nowadays/ =  · ā · ja · bho · li 

These rules do not apply for imported words

The schwa syncope rules do not really apply for imported words, whose lack of schwa (or presence thereof) depends a lot on the original pronunciation of the word as it was imported. Examples:
घटना /event/ [Hindi] = gha · ṭ[ə] · nā 
जानकारी /information/ [Hindi] = jā · n[ə] · kā · rī


With the above rules, you should be able to make educated guesses on the schwa status. While I will not guarantee you will get it right all the time, with this you should be able to pronounce most of the words correctly. 

Schwas are usually not deleted if they are part of a song or a poem. This is done for a poetic effect and should not be taken as the Gospel truth for pronunciation. 

Finally, if any word confuses you, it is best to consult that word with a native speaker. They can tell the correct pronunciation so you can be sure of your guess. 



1. पारस /lamp/
2. हिड्ँछ /walks/
3. गरीब /poor/
4. तँ /informal you/
5. निर /at/


1. इष्ट /relation/
2. लेख्य /written/
3. तर /but/
4. सिमसिम /light pitter-patter of the rain/
5. चीन /China/


1. धाक /bluff/ = dhāka
2. व्युत्पन्न /etymology/ =  vyutpann 
3. नाम /name/ =  nāma
4. पाटनबाट /from Patan/ =  pāṭna bāṭ 
5. सरल /simple/ = srala


A.1. Yes; Noun
A.2. No; Verb conjugate
A.3. Yes; Adjective
A.4. No; Single syllable word
A.5. No; Postposition/ Grammatical function
B.1. iṣṭa
B.2. lekhya
B.3. tara
B.4. simsim
B.5. cīn
C.1. dhāk
C.2. vyutpanna
C.3. nām 
C.4. pāṭan-bāṭa   
C.5. saral

Reading The Devanagari Script


Over the previous few lessons, we have gone through the script in thorough detail. Now that you know how to write the script, it is now important how to read it as well. After all, people do not write individual letters but rather words, so it is vital learning how to read the script fluently. 

Fortunately, Nepali is a very phonetic script. There are some exceptions pertaining to vowel sounds, but overall, you read what you see. One infamous rule is the Schwa deletion, where you randomly delete the last ‘a’ sound from the word, but we’ll discuss this topic here. For now, just omit the final a-sound unless specified.


कलम (pen)
It compromises of three characters, namely: क (ka), ल (la) and म (ma). Omitting the schwa in the end, we read this as kalam.

काम (work)
We now introduce diacritics. This is the ā diacritic attracted to क (ka), thus we read it as . Overall, the word is read as kām.

दूध (milk)
Similar to above, it uses the ū diacritic. It reads as dūdh.


हुन्छ (okay) 
Now we introduce ligatures and conjuncts. This is the ncha conjunct. It should be noted that the schwa of conjuncts are usually retained, thus this word reads as huncha.

गर्छ (does)
This is the rcha conjunct. We read it as garcha.

प्रेस (press)
This is the pre conjunct, made up of pr e. The word reads as pres.

मृग (deer)
The diacritic used belongs to the semi-vowel ‘ṛ(i)’, thus the word is read as ‘mṛ(i)ga’. The schwa here is not deleted for some reason.


संसार (world)
We now introduce nasalisers. This particular dot adds an ‘ṃ’ sound after the syllable, thus we read it as saṃsār. Note that IAST denotes this sound as ‘m’, although a closer real approximation is actually ‘n’.

बाँस (bamboo)
The second type of nasaliser, this one adds a nasal sound to the syllable itself than add it after the word ends. Thus, this reads as bām̐s. Note that the actual pronunciation is closer to ‘bās’ and the ‘’ only tells you to read ‘’ with a nasal voice.

श्रेष्ठ (good)
It contains the ligature śre and ṣṭha. The word overall reads as śreṣṭha.

श्रुतिसम्भिन्नार्थक (homonym)
Full of ligatures, it contains the following conjuncts: śrumbhinnārtha. With proper care, we can read this word as śrutisambhinnārthak.

ओर्ह्लिनु (to descend)
A rather complicated ligature. We can easily see that the main character here is ह्ल (hla), which is modified into ह्लि (hli). Since the sickle used above it is the r-diacritic (which precedes the word), the character is read as rhli. Overall, the word is orhlinu.



1. घरमा पानी पर्‍यो 
2. हामीलाई सञ्चो थिएन 
3. अहिले काठमाडौंमा शित्तल छ
4. अस्ति बाटोमा ओर्ह्लिरहेको मृगलाई देखेँ
5. आकाशको रङ्गसँगै तिमी उडिगयौ स्वप्नसरि, नहेरी नफर्की बाटै बाटो लाग्यौ 


ANSWERS (translations are approximate

A.1. ghar-mā pānī par‍yo
A.2. hāmī-lāī sañco thiena 
A.3. ahile kāṭhmāḍauṃ-mā śittal cha
A.4. asti bāṭo-mā orhliraheko mṛ(i)ga-lāī dekhem̐
A.5. ākāś-ko raṅg-sam̐gai timī uḍigayau swapna-sari, naherī napharkī bāṭai bāṭo lāgyau
B.(A.1.). Rain fell on (the) house
B.(A.2.). We were not (feeling) fine
B.(A.3.). (It) is cool in Kathmandu now
B.(A.4.). Day before yesterday on the road (I) saw a deer going downhill
B.(A.5.). You flew away with (the) colours of (the) sky like (a) dream, without looking (or) turning back (you) went (for the) road

Writing The Devanagari Script


Many younger people in Japan find it difficult to write their complex Kanji script due to the advent of digitalisation. Since there is no incentive in remembering how to write them, many people just opt reading the letter. 

Of course, if you are only interested in reading Nepali and not writing it, there is no incentive to learn how to write the script! However, this does not mean you should not learn how to write it down, because you many come across many situations which obligates you to write the script down. We may succeed in writing the word down, but it will usually be inconvenient. 

This is because we have been seeing the script as glyphs, rather than building blocks. Thus, we might be tempted to write each and every letter as they appear, making the text blocky and choppy. This is horribly inefficient.

So, how do people write the Devanagari script?


The primary thing to understand here is that, unlike the Japanese or the Chinese script, the order of strokes (which build up your character) does not matter. However, in order to increase the efficiency of writing, we employ the following guidelines:

  • Write the characters without the horizontal line above it
  • Attach any diacritics (other than dots) after you are done writing that particular character (don’t make the horizontal line yet)
  • After you are done, carefully make the horizontal line above the word
  • Finally, jot down the remaining dots or strokes after you are done

Caveats: Of course, I’d hate to be dogmatic about your writing style. Since everyone writes differently, you can try and explore your own writing style. For example, I know people who write their diacritics as they come. Some people may omit the horizontal line for speed, as seen in pharmacies. I have attached an animated GIF below to show you how I would write the sentence below. Note that this is my writing style and serves as a reference only. For example, I draw my horizontal lines from right to left, since I am a left handed person. Also, you may wish not to emulate my handwriting as it is not exactly the most calligraphic. 

Sentence: I like dogs.
Nepali: मलाई कुकुर मनपर्छ (ma-lāī kukur manparcha)


  1. ma
  2. l(a)
  3. Diacritic of ā
  4. ī 
  5. Horizontal line
  6. k(a)
  7. Diacritic of u
  8. Repeat steps 6-7
  9. r(a)
  10. Horizontal line
  11. ma
  12. n(a)
  13. pa
  14. cha
  15. r as part of the ligature rch-a
  16. Horizontal line
  17. Full stop 


When using lined paper, it is better if we use the line itself as the horizontal line guide. In other words, you write the letters on the top rather than the bottom.



1. घरमा पानी छ
2. यो मन त मेरो नेपाली हो 
3. किकि, तिमीले मलाई माया गर्छौ ? 
4. संसारको काम गर्ने कालुको बाङ्गो बञ्चरोले बाँस काट्यो 

Handwritten Numbers And The Number System


Numbers are written slightly differently than in print form, thus is an important thing to learn when you are practising Nepali. As an added bonus, I have included the number system as well, because numbers are counted a bit differently in Nepali.





Unlike the International System, Nepali has its own number system, with intervals in two instead of three. For example, writing 12 Million will be “12,000,000” but in Nepali, it will be “1,20,00,000”. See the following diagram:

The first three numbers will be in the interval of three, while any number after that will be grouped in intervals of two.



1. 145652
2. 46474790
3. 6245338


1. 1456524773473
2. 146474712135
3. 6245244


B.1. 14,56,52,47,73,473
B.2. 1,46,47,47,12,135
B.3. 62,45,244