Download and install the indispensable font(s) to view Sanskrit in its full glory
Read Transliterating (2) (English) to fully understand the transliteration system

Learning Sanskrit - Rules of Sandhi

Rules of Sandhi or Combination (compiled)

VOWEL SANDHI
Prim. Rules	Sec. Rules	VISARGA SANDHI		CONSONANT SANDHI
1^st 2^nd 3^rd 4^th 5^th 6^th 7^th	1^st 2^nd 3^rd 4^th 5^th 6^th 7^th	1^st 2^nd 3^rd 4^th 5^th 6^th 7^th	8^th 9^th 10^th	1^st 2^nd 3^rd 4^th 5^th 6^th 7^th	8^th 9^th 10^th 11^th 12^th 13^th 14^th	15^th 16^th 17^th 18^th 19^th 20^th 21^st	22^nd 23^rd 24^th
Further information				See Combination for more information

Vowel Sandhi: Primary Rules

1^st Primary Rule
Two Sanskrit vowels cannot be placed together (one following the other).

2^nd Primary Rule
Guṇa is formed from adding "a" or "ā" to the simple vowels. If the process is repeated, Vṛddhi is formed thereby.

3^rd Primary Rule
If a simple vowel (not a diphthong), short or long, be followed by a similar vowel, short or long, both of them will merge into the similar long vowel.

4^th Primary Rule
When "i-ī, u-ū, ṛ-ṝ and ḷ" are followed by a dissimilar vowel, then "y, v, r and l" are respectively substituted for them.

5^th Primary Rule
When "a" or "ā" are followed by a vṛddhi letter, both of them will be absorbed into the vṛddhi letter.

6^th Primary Rule
"e", "o", "ai" and "au", when followed by a vowel "within one word", are changed to "ay", "av", "āy" and "āv" respectively. However, when there are "two words", one ending in "e", "o", "ai" and "au", and the other beginning with any vowel, this rule is optionally used. If not used, you must use the 7^th primary rule.

7^th Primary Rule
a) "e" and "o" at the end of a word, when followed by any vowel (except "a") are firstly transformed into "ay" and "av" (6^th primary rule) and then "y" and "v" may be optionally dropped. No coalescence after that, of course. That is, the words remain separate.
b) "e" and "o" at the end of a word, when followed by "a", do not undergo any changes. However, "a" merges into "e" and "o" and the apostrophe (avagraha) is written in its place.
c) "ai" and "au" at the end of a word, when followed by any vowel are firstly transformed into "āy" and "āv" (6^th primary rule) and then "y" and "v" may be optionally dropped. No coalescence after that, of course. That is, the words remain separate.

top

Vowel Sandhi: Secondary Rules

1^st Secondary Rule
When a dual form (verb, pronoun or noun) ends in "ī", "ū" or "e" is not to be combined, that is, no Sandhi is possible here.

2^nd Secondary Rule
When "ṛ" or "ḷ" be followed by "ṛ" or "ḷ"; "ṛ" or "ḷ" may optionally be substituted for both.

3^rd Secondary Rule
When "a" is followed by the sacred word "Om" or the word "ā", is dropped.

4^th Secondary Rule
According to the 4^th Primary Rule: "When "i-ī, u-ū, ṛ-ṝ and ḷ" are followed by a dissimilar vowel, then "y, v, r and l" are respectively substituted for them". In turn, these "y, v, r and l" may be optionally doubled if they come after "h" or "r". In fact, this very rule may be applied to any consonant except "h".

5^th Secondary Rule
When "i-ī", "u-ū", "ṛ-ṝ" or "ḷ" at the end of a word are followed by a dissimilar vowel (except in a compound), are optionally not combined, and when so they are shortened if long.

6^th Secondary Rule
When "a-ā", "i-ī", "u-ū" and "ḷ-ḹ" at the end of a word are followed by "ṛ" (short form only), are optionally not combined, and when so they are shortened if long.

7^th Secondary Rule
Words consisting of just a single vowel are not combined at all. The word "ā" is obviously also included, except when it means "a little" or "a limit", or when used as a preposition and it is followed by a vowel.

top

VISARGA SANDHI

1^st Rule
The final letters "s" and "r" may be changed to Visarga (ḥ). There are words "originally" ending in "s" and words "originally" ending in "r", but Visarga may be substituted for the final "s" and "r" according to the following sub-rules: a) Final "s" of a word followed by any letter or by nothing. b) Final "r" of a word followed by a hard consonant or by nothing. Of course, once "ḥ" is substituted for "s" or "r", you might have to apply other rules to the word ending now in Visarga.

2^nd Rule
Visarga (substituted for "s" and not "r") preceded by "a" and followed by "a" or a soft consonant, is changed to "u". In turn, if there is an initial "a" in the second word, an apostrophe (') is to be substituted for it.

3^rd Rule
When Visarga is followed by "c", "ch", "ṭ", "ṭh", "t" and "th", themselves not followed by a sibilant (ś, ṣ or s), is changed to "ś" (before "c" and "ch"), "ṣ" (before "ṭ" and "ṭh") and "s" (before "t" and "th"). It is to be noted that the words "saḥ" and "eṣaḥ" does not follow this rule at all (See 10^th Rule).

4^th Rule
When Visarga is preceded by "i" or "u" and does not belong to any particular termination, is transformed into "ṣ" [except in the word "muhuḥ" (originally "muhur") --suddenly, at once, incessantly, etc.--] if followed by a hard consonant belonging to guttural or labial class.

5^th Rule
When "dviḥ", "triḥ" and "catuḥ" play the role of adverbs showing frequency, they change optionally their Visarga to "ṣ" if followed by a hard consonant belonging to guttural or labial class.

6^th Rule
A word ending in "iḥ" or "uḥ" changes optionally its Visarga to "ṣ" if followed by a hard consonant belonging to guttural or labial class, only when the following word is necessary in order to complete the sense.

7^th Rule
When Visarga is preceded by any vowel (except "a" or "ā") and followed by a vowel or a soft consonant is to be transformed into "r".

8^th Rule
When a word ending in "r" (either originally or after applying the previous 7^th Rule) is followed by "r" or "ḍh", the final "r" is dropped, and any preceding "a", "i" or "u" is made long.

9^th Rule
When Visarga (preceded by "ā") is followed by a soft consonant must be dropped. However, it is optionally dropped when: (1) followed by a vowel, (2) preceded by "a" and followed by a vowel --except "a"--. Finally, if one chooses not to drop it, Visarga must be changed to "y" in those two cases.

10^th Rule
The Visarga of the masculine pronouns "saḥ" (he, that) and "eṣaḥ" (he, this) is dropped before a consonant when they are not used in a negative Tatpuruṣa compound. Besides, sometimes in poetry, the final Visarga is dropped before a vowel (except "a") to meet the requirements of the meter. Obviously, "a" in the resulting "sa" is furtherly combined according to the well-known rules of Vowel Sandhi.

top

Consonant Sandhi

1^st Rule
No Sanskrit word can end in two or more consonants. The only exceptions are the endings "rk", "rṭ", "rt" and "rp". (Roots are not to be included in this rule)

2^nd Rule
(1^st sub-rule) Any final "hard" consonant (except Sibilants) of a word placed at the end of a sentence remains unchanged.
(2^nd sub-rule) Any final "hard" consonant (except Sibilants) placed before another "hard" consonant generally remains unchanged, except when you are forced to use some other rule of Sandhi.
(3^rd sub-rule) Any final "hard" consonant (except Sibilants) placed before a "soft" letter (any soft consonant --except "nasals", but this exception is optionally overlooked sometimes; See 6^th rule--, semivowel or vowel --except "Visarga"--) is changed to the respective "soft" consonant.
(When a "nasal" consonant is following, you will have to use the 6^th Rule of Consonant Sandhi).

3^rd Rule
Either at the end of the final word of a sentence or at the end of a single word which is ready to be inserted in a sentence, there can be only (1) Vowels (except ṛ, ṝ and ḷ), (2) Unaspirate hard Consonants (except "c"), (3) Nasals (except "ñ"), (4) Visarga (ḥ) and (5) The letter "l", which is a semivowel. Any other letter being situated in the aforesaid position, must be replaced according to the following sub-rules:
(1^st sub-rule) The final unaspirate and aspirate soft letters (except the palatals "j" and "jh") are transformed into the corresponding "unaspirate" hard letter.
(2^nd sub-rule) The final Palatals (except "ñ") are frequently changed to "k". However, sometimes "j" is also transformed into "ṭ". The final "ñ" is to be changed to "ṅ".
(3^rd sub-rule) The final "ś" is changed to either "k" or "ṭ". In turn, the final "ṣ" and "h" are generally transformed into "ṭ", and rarely into "k".

4^th Rule
(1^st sub-rule) If a final "s" (dental) comes in contact with "ś" or any other Palatal (c, ch, j, jh or ñ), "ś" is substituted for "s" (See 1^st Rule of Visarga Sandhi). In turn, regarding the combination between some of the remaining Dentals and Palatals, it may be stated the following: (a) "t" changes to "c" or "j" before "c", "ch", "j" or "jh" respectively; (b) "n" changes to "ñ" before "c", "ch", "j", or "jh". (Exceptions: "t" or "n" coming after "ś" are not affected by this rule, that is, they are not transformed into the respective palatals).
(2^nd sub-rule) If a final "s" (dental) comes in contact with "ṣ" or any other Cerebral (ṭ, ṭh, ḍ, ḍh or ṇ), "ṣ" is substituted for "s" (See 1^st and 3^rd Rules of Visarga Sandhi too). In turn, regarding the combination between some of the remaining Dentals and Cerebrals, it may be stated the following: (a) "t" changes to "ṭ" before "ṭ"; (b) "n" changes to "ṇ" before any cerebral.

5^th Rule
Any consonant --except Nasals, Semivowels and Sibilants-- is changed to the first of its class, when followed by a hard consonant (this is "mainly" a kind of complement to the 2^nd sub-rule of the second rule of Consonant Sandhi; go to the 4^th rule for more specific information about changes in Sibilants).

6^th Rule
If a final consonant, except "r" or "h", be followed by a Nasal, the Nasal of its class is optionally substituted for it.

7^th Rule
If a word ending in a Cerebral letter be followed by "s" or a letter of the Dental class (t, th, d, dh and n), both "s" and the aforesaid Dentals (except "n" of "nāma" --name--, "navati" --ninety-- and "nagarī" --city--) remain unaffected.

8^th Rule
When "h" comes after any of the first four letters of a class (Guttural, Palatal, Cerebral, Dental or Labial), is optionally transformed into the soft aspirate consonant of that class. That is, "h" may be optionally changed to the fourth letter of the respective class.

9^th Rule
When "ś" is both preceded by a word ending in any of the first four letters of a class and followed by a Vowel, a Semivowel, a Nasal or "h", is optionally transformed into "ch".

10^th Rule
The letter "m" situated at the end of a word is to be always transformed into Anusvāra
when followed by a Consonant.

11^th Rule
According to the tenth Rule, "m" is always transformed into "Anusvāra" when situated at the end of a word and followed by a Consonant. In turn, if the Consonant is not "ś, ṣ, s or h", Anusvāra (ṁ) may be optionally transformed into the Nasal of the class to which that consonant belongs. This would be a complement to the 10^th rule of Consonant Sandhi really. However, this transformation is necessary if Anusvāra is in the middle of a word.
Exception: The words "sam" and "pum" have their own way of combination, and therefore they do not come completely under the present rule. These are also ruled by the following rules:
a) The "m" of "sam" (a prefix indicating either "conjunction" or "union" or "thoroughness", "intensity" or "completeness") is transformed into "ṁs" (and not either into "ṁ" or into "ṅ" as stated by the present rule) when followed by any form of the root "kṛ" (to do, to make, to perform, etc.). Anusvāra may also be optionally changed to Anunāsika (m̐). Thus, "ṁs" is turned into "m̐s".
b) The "m" of "pum" (a form which is used in compounds for "puṁs" --man, a human being, spirit of man, etc.--) is optionally changed to "ṁs" (and not either to "ṁ" or to the corresponding Nasal as stated by the present rule) when followed by a hard consonant (except a Sibilant) preceding a Vowel, a Semivowel, a Nasal or "h" (Careful with this!). Anusvāra may also be optionally changed to Anunāsika (m̐). Thus, "ṁs" is turned into "m̐s".

12^th Rule
When "m" and "n" are not at the end of a "pada" (technical term for any "inflected or declined word"), they are changed into Anusvāra if followed by a consonant (except a Nasal, a Semivowel or "h"). (Note that even though "any word" is generally called "pada", in this case it refers to an inflected word, i.e. a word having undergone any type of declension or conjugation. Careful here!).
Exception: a) This change is optional if "m" is followed by the conjunct "hm". If it is not changed, "m" remains unchanged (i.e. it remains "m").
b) In turn, that very "m" may be optionally turned into "n" if followed by "hn". If not so, "m" is transformed into Anusvāra.
c) In the event of "m" when followed by "hy, hv, hl", then, "m̐y, m̐v, m̐l" (nasalized forms of those three Semivowels) are optionally substituted for "m". If not so, "m" is turned into Anusvāra.
The present Rule should be used together with the 10^th Rule, because it is a perfect complement, indeed. You should use the 10^th Rule freely, but making sure that you are not violating any premise stated by the 12^th Rule. Be careful, then.

13^th Rule
(1^st sub-rule) If a consonant is followed by another consonant of the same kind (except Nasals, Semivowels and "h"), it may be optionally dropped.
(2^nd sub-rule) If a consonant belonging to any of the first five classes --Guttural, Palatal, Cerebral, Dental and Labial-- (except "ñ") comes after a Semivowel, the former is optionally doubled.
(3^rd sub-rule) If a consonant (except "h") comes after a vowel and is not followed by one (i.e. by a vowel), it is optionally doubled.
(4^th sub-rule) If a consonant belonging to any of the first five classes (except Nasals and, of course, the soft unaspirate consonants --i.e. "g, j, ḍ, d and b"--) is followed by a soft unaspirate or aspirate consonant (i.e. g, gh, j, jh, ḍ, ḍh, d, dh, b or bh) is to be transformed into the soft unaspirate consonant of its class (i.e. g, j, ḍ, d or b, respectively). This sub-rule is a complement to the 5^th Rule of Consonant Sandhi.
(5^th sub-rule) A Semivowel is optionally doubled when preceded by a consonant belonging to any of the first five classes (i.e. Gutturals, Palatals, Cerebrals, Dentals and Labials), except "ñ".
(6^th sub-rule) If a consonant (except "h") comes after "r" or "h" preceded by a vowel, is optionally doubled. This is a complement to the previous second sub-rule.

14^th Rule
If "n" at the end of a word is followed by "ś", then, a "t" is optionally inserted between them. After this optional insertion, you may have to use other Rules in order to polish the Sandhi or combination.

15^th Rule
(1^st sub-rule) When "ḍ" or "n" at the end of word are followed by "s", "dh" (disguised as "t") may be optionally inserted in between. Note that, in the first case, you will have to turn the resulting conjunct "ḍts" into "ṭts" by the 5^th Rule of Consonant Sandhi. This sub-rule is a perfect complement to the 7^th Rule of Consonant Sandhi.
(2^nd sub-rule) When "ṅ", "ṇ" or "n" are at the end of a word and preceded by a short vowel, they must be doubled if followed by a vowel.

16^th Rule
The consonant "s" belonging to either a Substitute (i.e., when "s" is intended to be substituted in order to form inflected forms) or a Termination is obligatorily changed to "ṣ" when preceded by a Vowel (except "a" and "ā"), a Semivowel, a Guttural or "h". Besides, if "s" is followed by the Vowels "ṛ" or "ṝ", or the consonant "r", "s" remains unchanged (in short, "s" + "ṛ", "ṝ" or "r" = "sṛ", "sṝ" or "sr"), that is, the present Rule is not valid. Theorically, this Rule accepts any of the remaining consonants following "s", but as a matter of fact all soft consonants (the rest of Semivowels included) and Sibilants are generally excluded. If these consonants appear, the Rules of Visarga Sandhi along with other Rules of Consonant Sandhi are used instead. So, only the ten hard consonants (k, kh, c, ch, ṭ, ṭh, t, th, p and ph) might theorically follow "s", but you must be careful here when using this Rule in order not to conflict other Rules of Consonant or Visarga Sandhi. You will find that the current Rule works complementarily with some Rules governing the combination of a final "s" or Visarga. Here you are the list of Rules mainly working as a complement to the present Rule:
1) Consonant Sandhi: a) The third sub-rule of the 2^nd Rule; and b) The 4^th Rule;
2) Visarga Sandhi: The 3^rd, 4^th, 5^th, 6^th and 7^th Rules. This Rule is valid even if an Anusvāra substituted for "n", a Visarga or a Sibilant (ś, ṣ, s) intervene [i.e. even if they are situated between "s" and the preceding Vowel (except "a" and "ā"), Semivowel, Guttural or "h"]. The present Rule is commonly used only within a single word. (Note that an affix or a termination added to a word are not considered a second "separate" word really, because the final result after adding them to it is again a single word).

17^th Rule
When "n", at the end of a word (except "praśān" --a good and peaceful man--), is followed by "c", "ch", "ṭ", "ṭh", "t" or "th" (which are in turn followed by a Vowel, a Semivowel, a Nasal or "h"), "s" is inserted in between and "n" itself is changed to "ṁ" or "m̐". Sometimes, you will also have to use the 4^th Rule of Consonant Sandhi to polish the combination. This Rule is the perfect complement to the 11^th and 12^th Rules of Consonant Sandhi.

18^th Rule
When "n", comes after "r", "ṣ", "ṛ" or "ṝ", within the same word, is to be obligatorily changed to "ṇ" although a Vowel, a Semivowel (except "l"), a Nasal, a consonant belonging to the Guttural or Labial classes, or "h" comes between the abovementioned letters (i.e. "r", "ṣ", "ṛ" and "ṝ") and "n". Note that this transformation does not occur when "n" is at the end of a word, that is, the present Rule of Sandhi cannot be used with a final "n".

19^th Rule
(1^st sub-rule) The consonant "c" is obligatorily inserted between "ch" and the preceding short vowel.
(2^nd sub-rule) The consonant "c" is obligatorily inserted between "ch" and the preceding long vowel not standing at the end of "real" word, that is, the aforesaid long vowel must belong to a reduplicative syllable, etc. which are mostly used in order to form different kinds of verbal conjugations (Perfect, Aorist and the like).
(3^rd sub-rule) The consonant "c" is optionally inserted between "ch" and a preceding long vowel standing at the end of a "real" word, except "ā" in "mā" and the preposition "ā".
(4^th sub-rule) From the last statement, it is obvious that "c" is obligatorily inserted between "ch" and a preceding "mā" or the preposition "ā".

20^th Rule
(1^st sub-rule) When "ṅ" and "ṇ" are followed by a Sibilant (ś, ṣ or s), "k" and "ṭ"
are respectively suffixed to the former optionally.
(2^nd sub-rule) Hard unaspirate consonants are optionally turned into hard aspirates
when followed by a Sibilant.

21^st Rule
(1^st sub-rule) When followed by "l", any Dental letter (except "n" and "s") is changed to "l".
(2^nd sub-rule) When followed by "l", "n" is to be turned into nasalized "l".

22^nd Rule
When a Dental letter (except "s") is followed by "ṣ", it does not substitute its corresponding Cerebral. In other words, Dental letters (except "s") remain unaffected before "ṣ".

23^rd Rule
When a Cerebral letter is followed by a Dental one within a word (careful!), the Dental consonant is changed to the corresponding Cerebral. This rule is a kind of complement to both the 7^th Rule of Consonant Sandhi and (a)/(b) of the second sub-rule of the 4^th Rule of Consonant Sandhi.

24^th Rule
(1^st sub-rule) When "ś" is followed by "t" or "th", is changed to "ṣ". Afterward, provided that the Sandhi occurs within a single word, by 23^rd Rule of Consonant Sandhi, "t" and "th" are changed to "ṭ" and "ṭh". If not, "t" and "th" remain unaffected according to the 7^th Rule of Consonant Sandhi.
(2^nd sub-rule) When a root or a noun, both of them ending in a soft aspirate consonant, are followed by a suffix or termination beginning with "t" or "th", the latter (i.e. "t" and "th") must be changed to "d" and "dh" respectively, which attract the aspiration. In other words, the soft aspirate consonant in the root or noun is transformed into the respective soft unaspirate one, while "d" is turned into "dh". Obviously, "dh" coming from "th" remains the same despite stealing the aspiration from the previous consonant, that is, there is no "dhh". This is a complement of the fourth sub-rule of the 13^th Rule of Consonant Sandhi.
(3^rd sub-rule) When a root ending in "h" is followed by a suffix or termination beginning with "t" or "th", the latter (i.e. "t" and "th") must be changed to "d" and "dh", respectively. In turn, "h" itself is transformed into "gh". And "d" or "dh" ultimately attracts the aspiration in "gh", that is, "d" changes to "dh" and "gh" to "g" (obviously, "dh" coming from "th" remains the same despite stealing the aspiration from "gh", that is, there is no "dhh")

top

Further Information

This document was conceived by Gabriel Pradīpaka, one of the two founders of this site, and spiritual guru conversant with Sanskrit language and Trika philosophy.

For further information about Sanskrit, Yoga and Indian Philosophy; or if you simply want to comment, ask a question or correct a mistake, feel free to contact us: This is our e-mail address.