View source for Sandhi

{{Short description|Type of sound change at morpheme or syllable boundaries}}
{{Sound change}}

'''Sandhi''' ({{IPAc-en|s|æ|n|ˈ|d|iː}} {{respell|san|DEE}}; {{langx|sa|{{wikt-lang|sa|सन्धि}}|lit=joining}}, {{IPA|sa|sɐnˈdʱi|pron}}) is any of a wide variety of [[phonology|sound]] changes that occur at [[morpheme]] or [[Word boundary (linguistics)|word boundaries]]. Examples include fusion of sounds across word boundaries and the alteration of one sound depending on nearby sounds or the grammatical function of the adjacent words. Sandhi belongs to [[morphophonology]].

Sandhi occurs in many languages, e.g. in the phonology of [[Languages of India|Indian languages]] (especially [[Sanskrit]], [[Tamil language|Tamil]],  [[Sinhala language|Sinhala]], [[Telugu language|Telugu]], [[Marathi language|Marathi]], [[Hindi]], [[Pali]], [[Kannada]], [[Bengali language|Bengali]], [[Assamese language|Assamese]] and [[Malayalam]]). Many dialects of [[British English]] show [[linking and intrusive R]].

[[Tone sandhi]] in particular defines tone changes affecting adjacent words and syllables. This is a common feature of many tonal languages such as [[Burmese language|Burmese]] and [[Chinese language|Chinese]].

==Types==

=== Internal and external sandhi ===
Sandhi can be either:
* '''internal''', at morpheme boundaries within words, such as ''syn- + pathy'': ''sympathy'', or
* '''external''', at word boundaries, such as the pronunciation "''tem books''" for ''ten books'' in some dialects of [[English language|English]]. The [[Linking and intrusive R|linking {{IPA|/r/|cat=no}}]] process of some dialects of English ("''I saw-r-a film''" in [[British English]]) is a kind of external sandhi, as are [[French language|French]] ''[[liaison (French)|liaison]]'' (pronunciation of usually silent final consonants of words before words beginning with vowels) and [[Italian language|Italian]] ''[[syntactic gemination|raddoppiamento fonosintattico]]'' (lengthening of initial consonants of words after certain words ending in vowels).

It may be extremely common in speech, but sandhi (especially external) is typically ignored in spelling, as is the case in English (exceptions: the distinction between ''a'' and ''an''; the prefixes ''con-'', ''en-'', ''in-'' and ''syn-'', whose ''n'' [[Assimilation (phonology)|assimilates]] to ''m'' before ''p'', ''m'' or ''b''). Sandhi is, however, reflected in the orthography of [[Sanskrit]], [[Sinhala language|Sinhala]], [[Telugu language|Telugu]], [[Marathi language|Marathi]], [[Pali]] and some other Indian languages, as with Italian in the case of compound words with [[Lexicalization|lexicalised]] [[syntactic gemination]].

External sandhi effects can sometimes become morphologised (apply only in certain [[morphology (linguistics)|morphological]] and [[syntax|syntactic]] environments) as in [[Tamil language|Tamil]]<ref name=":0">{{cite book|first=Harold F.|last=Schiffman|url=https://books.google.com/books?id=Oqe-QsaZnnQC&q=sandhi&pg=PA20|title=A Reference Grammar of Spoken Tamil|publisher=Cambridge University Press|date=1999|page=20|isbn=9780521640749}}</ref><ref name=":1">{{cite web |title=Gemination of stops in Tamil: implications for the phonology-syntax interface |url=https://www.researchgate.net/publication/237415274 |author=Hemalatha Nagarajan |format=PDF }}</ref> and, over time, turn into [[consonant mutation]]s.

=== Tone sandhi ===
Most [[tone (linguistics)|tonal language]]s have [[tone sandhi]] in which the tones of words alter according to certain rules. An example is the behavior of [[Mandarin Chinese]]; in isolation, tone 3 is often pronounced as a falling-rising tone. When a tone 3 occurs before another tone 3, however, it changes into tone 2 (a rising tone), and when it occurs before any of the other tones, it is pronounced as a low falling tone with no rise at the end.

An example occurs in the common greeting {{lang|zh|你好}} {{Transliteration|zh|nǐ hǎo}} (with two words containing underlying tone 3), which is in practice pronounced {{Transliteration|zh|ní hǎo}}. The first word is pronounced with tone 2, but the second is unaffected.

==Examples==

=== Celtic languages ===
{{unreferenced section|date=November 2024}}
{{Main|Breton mutations|Cornish grammar#Initial consonant mutation|Welsh morphology#Initial consonant mutation|Irish initial mutations|Manx language#Initial consonant mutations|Scottish Gaelic phonology#Lenition and spelling}}
In Celtic languages, the [[consonant mutation]] sees the initial consonant of a word change according to its morphological or syntactic environment. 
Following are some examples from Breton, Irish, Scottish Gaelic, and Welsh:

{| class="wikitable" style="margin:1em auto;"
! Breton !! Welsh !! Irish !! Scottish Gaelic !! Gloss
|-
| {{lang|br|gwreg}} || {{lang|cy|gwraig}} || {{lang|ga|bean}} || {{lang|gd|bean}}* || woman/wife
|-
| {{lang|br|bras}} <!--compare with mor in old breton --> || {{lang|cy|mawr}} || {{lang|ga|mór}} || {{lang|gd|mòr}} || big
|-
| {{lang|br|ar '''w'''reg '''v'''ras}} || {{lang|cy|y '''w'''raig '''f'''awr}} || {{lang|ga|an '''bh'''ean '''mh'''ór}} || {{lang|gd|a' '''bh'''ean '''mh'''òr}} || the big woman
|-
| {{lang|br|kazh}} || {{lang|cy|cath}} || {{lang|ga|cat}} || {{lang|gd|cat}} || cat
|-
| {{lang|br|e '''g'''azh}} || {{lang|cy|ei '''g'''ath}} || {{lang|ga|a '''ch'''at}} || {{lang|gd|a '''ch'''at}} || his cat
|-
| {{lang|br|he '''c'h'''azh}} || {{lang|cy|ei '''ch'''ath}} || {{lang|ga|a '''c'''at}} || {{lang|gd|a '''c'''at}} || her cat
|-
| {{lang|br|o '''c'h'''azh}} || {{lang|cy|eu '''c'''ath}} || {{lang|ga|a '''gc'''at}} || {{lang|gd|an '''c'''at}} || their cat
|}

=== Portuguese ===
{{unreferenced section|date=November 2024}}
When two words belonging to the same phrase are pronounced together, or two [[morpheme]]s are joined in a word, the last sound in the first may be affected by the first sound of the next (sandhi), either coalescing with it, or becoming shorter (a semivowel), or being deleted. This affects especially the sibilant consonants {{IPA|/s/, /z/, /ʃ/, /ʒ/}}, and the unstressed final vowels {{IPA|/ɐ/, /i, ɨ/, /u/}}.

==== Consonant sandhi ====
As was mentioned above, the dialects of Portuguese can be divided into two groups, according to whether syllable-final sibilants are pronounced as postalveolar consonants {{IPA|/ʃ/, /ʒ/}} or as alveolar {{IPA|/s/, /z/}}. At the end of words, the default pronunciation for a sibilant is voiceless, {{IPA|/ʃ, s/}}, but in connected speech the sibilant is treated as though it were within a word ([[Assimilation (linguistics)|assimilation]]):

* If the next word begins with a [[voiceless consonant]], the final sibilant remains voiceless {{IPA|/s, ʃ/}}; {{lang|pt|bons tempos}} {{IPA|[bõʃ ˈtẽpuʃ] or [bõs ˈtẽpus]}} ('good times').
* If the next word begins with a voiced consonant, the final sibilant becomes voiced as well /z, ʒ/; {{lang|pt|bons dias}} {{IPA|[bõʒ ˈdiɐʃ] or [bõz ˈdʒiɐs]}} ('good days').
* If the next word begins with a vowel, the final sibilant is treated as intervocalic, and pronounced [z]; {{lang|pt|bons amigos}} {{IPA|[bõz ɐˈmiɣuʃ] or [bõz aˈmiɡus]}} ('good friends').

When two identical sibilants appear in sequence within a word, they reduce to a single consonant. For example, {{lang|pt|na'''sc'''er, de'''sç'''o, e'''xc'''esso, e'''xs'''udar}} are pronounced with [s] by speakers who use alveolar sibilants at the end of syllables, and {{lang|pt|di'''sj'''untor}} is pronounced with {{IPA|[ʒ]}} by speakers who use postalveolars. But if the two sibilants are different they may be pronounced separately, depending on the dialect. Thus, the former speakers will pronounce the last example with {{IPA|[zʒ]}}, whereas the latter speakers will pronounce the first examples with [s] if they are from Brazil or {{IPA|[ʃs]}} if from Portugal (although in [[relaxed pronunciation]] one of the siblants may be dropped). This applies also to words that are pronounced together in connected speech:

* sibilant + /s/, e.g., {{lang|pt|a'''s s'''opas}}: either [s] (most of Brazil); [ʃs] (Portugal, standard)
* sibilant + /z/, e.g., {{lang|pt|a'''s''' '''z'''onas}}: either [z] (mostly in Brazil); [ʒz] (Portugal, standard)
* sibilant + /ʃ/, e.g., {{lang|pt|a'''s ch'''aves}}: always [ʃ];
* sibilant + /ʒ/, e.g., {{lang|pt|o'''s g'''enes}}: always [ʒ].

==== Vowel sandhi ====
Normally, only the three vowels /ɐ/, /i/ (in BP) or /ɨ/ (in EP), and /u/ occur in [[Stress (linguistics)|unstressed]] final position. If the next word begins with a similar vowel, they merge with it in connected speech, producing a single vowel, possibly [[Vowel length|long]] ([[crasis]]). Here, "similar" means that nasalization can be disregarded, and that the two central vowels /a, ɐ/ can be identified with each other. Thus,

* /aa, aɐ, ɐa, ɐɐ/ → [a(ː)] (henceforth transcribed [a (a)]); {{lang|pt|toda a noite}} {{IPA|[ˈtoða (a)ˈnojtɨ]}} or {{IPA|[ˈtoda (a) ˈnojtʃi]}} ('all night'), {{lang|pt|nessa altura}} {{IPA|[ˈnɛs awˈtuɾɐ]}} or {{IPA|[ˈnɛs alˈtuɾɐ]}} ('at that point').
* /aɐ̃, ɐɐ̃/ → [ã(ː)]) (henceforth transcribed [ã (ã)]); {{lang|pt|a antiga}} ('the ancient one') and {{lang|pt|à antiga}} ('in the ancient way'), both pronounced {{IPA|[ã (ã)ˈtʃiɡɐ]}} or {{IPA|[ã (ã)ˈtiɣɐ]}}. The open nasalized [ã] appears only in this environment.
* /ii, iĩ/ → [i(ː), ĩ(ː)] (henceforth transcribed [i (i), ĩ (ĩ)]); {{lang|pt|de idade}} {{IPA|[dʒi (i)ˈdadʒi] or [di (i)ˈðaðɨ]}} ('aged').
* /ɨɨ/ → [ɨ]; {{lang|pt|fila de espera}} {{IPA|[ˈfilɐ ðɨʃˈpɛɾɐ]}} ('waiting line') (EP only).
* /uu, uũ/ → [u(ː), ũ(ː)] (henceforth transcribed [u (u), ũ (ũ)]); {{lang|pt|todo o dia}} {{IPA|[ˈtodu (u)ˈdʒiɐ] or [ˈtoðu (u) ˈðiɐ]}} ('all day').

If the next word begins with a dissimilar vowel, then /i/ and /u/ become [[Approximant consonant|approximants]] in Brazilian Portuguese ([[synaeresis]]):

* /i/ + V → [jV]; {{lang|pt|durante o curso}} {{IPA|[duˈɾɐ̃tʃju ˈkuɾsu]}} ('during the course'), {{lang|pt|mais que um}} {{IPA|[majs kjũ]}} ('more than one').
* /u/ + V → [wV]; {{lang|pt|todo este tempo}} {{IPA|[ˈtoˈdwestʃi ˈtẽpu]}} ('all this time') {{lang|pt|do objeto}} {{IPA|[dwobiˈʒɛtu]}} ('of the object').

In careful speech and in with certain function words, or in some phrase stress conditions (see Mateus and d'Andrade, for details), European Portuguese has a similar process:

* {{IPA|/ɨ/ + V → [jV]}}; {{lang|pt|se a vires}} {{IPA|[sjɐ ˈviɾɨʃ]}} ('if you see her'), {{lang|pt|mais que um}} {{IPA|[majʃ kjũ]}} ('more than one').
* {{IPA|/u/ + V → [wV]}}; {{lang|pt|todo este tempo}} {{IPA|[ˈtoˈðweʃtɨ ˈtẽpu]}} ('all this time'), {{lang|pt|do objeto}} {{IPA|[dwɔbˈʒɛtu]}} ('of the object').

But in other [[Prosody (linguistics)|prosodic]] conditions, and in relaxed pronunciation, EP simply drops final unstressed {{IPA|/ɨ/}} and /u/ ([[elision]]), though this is subject to significant dialectal variation:

* {{lang|pt|durante o curso}} {{IPA|[duˈɾɐ̃tu ˈkuɾsu]}} ('during the course'), {{lang|pt|este inquilino}} {{IPA|[ˈeʃtĩkɨˈlinu]}} ('this tenant').
* {{lang|pt|todo este tempo}} {{IPA|[toˈðeʃtɨ ˈtẽpu]}} ('all this time'), {{lang|pt|disto há muito}} {{IPA|[diʃta ˈmũjtu]}} ('there's a lot of this').

Aside from historical set [[Contraction (grammar)|contractions]] formed by prepositions plus [[determiner]]s or pronouns, like {{lang|pt|à/dà, ao/do, nesse, dele}}, etc., on one hand and combined [[clitic]]<nowiki/>pronouns such as {{lang|pt|mo/ma/mos/mas}} (it/him/her/them to/for me), and so on, on the other, [[Portuguese spelling]] does not reflect vowel sandhi. In poetry, however, an apostrophe may be used to show [[elision]] such as in {{lang|pt|d'água}}.

===German dialects===
In various [[German dialects]] or the spoken [[Standard German]] one can find phonological processes that can be analysed as Sandhi. For example some varieties of [[Central Hessian]] show a vowel length alternation where, if the same long vowel were else to repeat in two consecutive syllables, the vowel is shortened/reduced in the first, but maintained in the second. Examples are [hɪɡiː] for HG ''hingehen'' ("go towards") (hin corresponds to [hiː] in Hessian) or [kən aːnʒ̊ə] for HG ''kein einziger'' ("no single [thing]").

===English===
In [[English phonology]], rhotic sandhi can be seen in [[Rhoticity in English|non-rhotic dialects]], when a word ends in a vowel followed by /r/, and the next word starts with a vowel as well, a [ɹ] ([[voiced alveolar approximant]]) sound will be inserted between the word, see for example, in Standard Southern British English "law and order" pronounced as [lo:rəno:də], "America and
China" pronounced as [əmɛrikəɹənʧɑjnə]<ref name=":9">{{cite book|first=Geoff|last=Lindsey|
url=https://books.google.com/books?id=6nSKDwAAQBAJ&pg=PA87|title=English after RP: Standard British pronunciation today|
publisher=Palgrave Macmillan|date=2019|page=87|isbn=9783030043575}}</ref> (see [[linking and intrusive R]])

=== French ===
French [[liaison (French)|''liaison'']] and ''[[enchaînement]]'' can be considered forms of external sandhi.<ref>{{Cite journal |last=Bennett |first=William |date=April 1991 |title=Liaison in French |url=http://www.tandfonline.com/doi/full/10.1080/00437956.1991.11435832 |journal=WORD |language=en |volume=42 |issue=1 |pages=57–88 |doi=10.1080/00437956.1991.11435832 |issn=0043-7956|url-access=subscription }}</ref>

In ''enchaînement'', a word-final consonant, when followed by a word that starts with a vowel, is articulated as though it is part of the following word. For example, {{lang|fr|sens}} ({{gloss|direction}}) is pronounced {{IPA|/sɑ̃s/|lang=fr}} and {{lang|fr|unique}} ({{gloss|unique}}) is pronounced {{Ipa|/ynik/|lang=fr}}; {{lang|fr|sens unique}} ({{gloss|one-way}}, as a street) is pronounced {{Ipa|/sɑ̃‿synik/|lang=fr}}.

''Liaison'' is a similar phenomenon, applicable to words ending in a consonant that was historically pronounced but that, in [[Modern French]], is normally silent when occurring at the end of a phrase or before another consonant. In some circumstances, when the following word starts with a vowel, the consonant may be pronounced, and in that case is articulated as if part of the next word. For example, {{lang|fr|deux frères}} ({{gloss|two brothers}}) is pronounced {{IPA|/dø fʁɛʁ/|}} with a silent {{angle brackets|x}}, and {{lang|fr|quatre hommes}} ({{gloss|four men}}) is pronounced {{IPA|/kat‿ʁɔm/}}, but {{lang|fr|deux hommes}} ({{gloss|two men}}) is pronounced {{IPA|/dø‿zɔm/}}.

===Japanese===
{{unreferenced section|date=November 2024}}
In [[Japanese phonology]], sandhi is primarily exhibited in [[rendaku]] ([[consonant mutation]] from unvoiced to voiced when not word-initial, in some contexts) and conversion of {{lang|ja|つ}} or {{lang|ja|く}} ({{Transliteration|ja|tsu}}, {{Transliteration|ja|ku}}) to a [[geminate consonant]] (orthographically, the [[sokuon]] {{lang|ja|っ}}), both of which are reflected in spelling – indeed, the {{lang|ja|っ}} symbol for gemination is morphosyntactically derived from {{lang|ja|つ}}, and voicing is indicated by adding two dots as in {{lang|ja|か／が}} {{Transliteration|ja|ka}}, {{Transliteration|ja|ga}}, making the relation clear. It also occurs much less often in {{Nihongo|''[[renjō]]''|連声}}, where, most commonly, a terminal {{IPA|/n/}} on one morpheme results in an {{IPA|/n/|}} (or {{IPA|/m/|}}) being added to the start of the next morpheme, as in {{lang|ja|天皇: てん ＋ おう → てんのう}} ({{Transliteration|ja|ten}} + {{Transliteration|ja|ō}} = {{Transliteration|ja|tennō}}), meaning "emperor"; that is also shown in the spelling (the kanji do not change, but the kana, which specify pronunciation, change).

=== Korean ===
[[Korean phonology|Korean]] has sandhi which occurs in the final consonant or consonant cluster, such that a morpheme can have two pronunciations depending on whether or not it is followed by a vowel. For example, the root {{lang|ko|읽}} {{IPA|/ik/}}, meaning ‘read’, is pronounced {{IPA|/ik/}} before a consonant, as in {{lang|ko|읽다}} {{IPA|/ik.ta/}}, but is pronounced like {{IPA|/il.g/}} before vowels, as in {{lang|ko|읽으세요}} {{IPA|/il.gɯ.se̞.jo/}}, meaning ‘please read’. Some roots can also aspirate following consonants, denoted by the letter {{lang|ko|[[ㅎ]]}} (hieut) in the final consonant. This causes  {{lang|ko|다}} {{IPA|/tɐ/}} to become {{IPA|/tʰɐ/}} in {{lang|ko|않다}} {{IPA|/ɐntʰɐ/}}, ‘to not be’.<ref>{{cite web |title=Korean Sound Change Rules |url=https://www.missellykorean.com/korean-sound-change-rules-pdf/ |access-date=22 October 2022 |website=Miss Elly Korean|date=6 September 2020 }}</ref>

=== Tamil ===
Tamil 'punarchi' (புனர்ச்சி) or sandhi has been rigorously and exhaustively documented in Tamil grammar texts since the early centuries AD. As modern Tamil <span data-darkreader-inline-bgcolor="">is strongly characterised by </span>[[diglossia]]<span data-darkreader-inline-bgcolor="">: there are two separate registers varying by speech context, a high register and a low one.</span><ref>Arokianathan, S. [http://www.ciil-ebooks.net/html/piil/acharya2.html Writing and Diglossic: A Case Study of Tamil Radio Plays] {{webarchive|url=https://web.archive.org/web/20070928092717/https://www.ciil-ebooks.net/html/piil/acharya2.html|date=28 September 2007}}. ciil-ebooks.net</ref><ref>{{Citation |last1=Steever |first1=S. B. |title=Diglossia: A Study of the Theory, with Application to Tamil |journal=Language |volume=64 |issue=1 |pages=152–155 |year=1988 |doi=10.2307/414796 |jstor=414796 |last2=Britto |first2=F.}}</ref> This in turn presents two corresponding domains for forming Sandhi.<ref name=":2">{{Cite book |last=Schiffman |first=Harold F. |url=https://theswissbay.ch/pdf/Books/Linguistics/Mega%20linguistics%20pack/Dravidian/Tamil%2C%20A%20Reference%20Grammar%20of%20Spoken%20%28Schiffman%29.pdf |title=A Reference Grammar of Spoken Tamil |publisher=[[Cambridge University Press]] |year=1998 |isbn=978-0-521-64074-9 |pages=20–21 |language=en |archive-url=https://web.archive.org/web/20240411095759/https://theswissbay.ch/pdf/Books/Linguistics/Mega%20linguistics%20pack/Dravidian/Tamil%2C%20A%20Reference%20Grammar%20of%20Spoken%20%28Schiffman%29.pdf |archive-date=11 April 2024 |url-status=live}}</ref> Tamil employs Sandhi for certain [[morphology (linguistics)|morphological]] and [[syntax|syntactic]] structures.<ref name=":0" /><ref name=":1" />

==== Vowel position ====
The vowel sandhi occurs when words or morphemes ending in certain vowels are followed by morphemes beginning with certain vowels. Consonant glides ({{Langx|ta|ய்|translit=Y}} and {{Langx|ta|வ்|translit=V}}) are then inserted between the vowels in order to 'smooth the transition' from one vowel to another.<ref name=":2" />

"The choice of whether the glide inserted will be ({{Langx|ta|ய்|translit=Y|label=none}} and {{Langx|ta|வ்|translit=V|label=none}}) in Tamil is determined by whether the vowel preceding the glide is a front vowel such as {{Langx|ta|இ, ஈ, எ, ஏ or ஐ|translit=i, ī, e, ē or ai}} or a back vowel, such as {{Langx|ta|உ, ஊ, ஒ, ஓ, அ or ஆ|translit=u, ū, o, ō, a or ā|label=}}."<ref name=":2" />

===== Examples in Spoken Tamil =====
<ref name=":2" />

====== Sandhi following front vowels ======
{| class="wikitable"
|+
!Vowel Ending
!Noun
!Grammatical Suffix
!Result
|-
|இ
|{{Langx|ta|நரி|lit=Fox|translit=Nari}}
|Interrogative, {{Langx|ta|ஆ|translit=Ā}}
|{{Langx|ta|நரியா|lit=A fox?|translit=Nariyā}}
|-
|ஈ
|{{Langx|ta|தீ|lit=Fire|translit=Tī}}
|Interrogative, {{Langx|ta|ஆ|translit=Ā}}
|{{Langx|ta|தீயா|lit=Fire?|translit=Tīyā}}
|-
|எ
|''No literary Tamil word ends in எ''
|{{Longdash}}
|{{Longdash}}
|-
|ஏ
|''No literary Tamil word ends in ஏ''
|{{Longdash}}
|{{Longdash}}
|-
|ஐ
|{{Langx|ta|யானை|lit=Elephant|translit=Yāṉai}}
|Interrogative, {{Langx|ta|ஆ|translit=Ā}}
|{{Langx|ta|யானையா|lit=An elephant?|translit=Yāṉaiyā}}
|}

====== Sandhi following back vowels ======
{| class="wikitable"
!Vowel Ending
!Noun
!Grammatical Suffix
!Result
|-
|உ
|{{Langx|ta|குரு|lit=Guru, teacher|translit=Kuru}}
|Interrogative, {{Langx|ta|ஆ|translit=Ā}}
|Usually deleted, or added later after sandhi rules have applied.
A few exceptions: {{Langx|ta|குருவா|lit=A guru?|translit=Kuruvā}}
|-
|ஊ
|{{Langx|ta|பூ|lit=Flower|translit=Pū}}
|Interrogative, {{Langx|ta|ஆ|translit=Ā}}
|{{Langx|ta|பூவா|lit=A flower?|translit=Pūvā}}
|-
|ஒ
|''Doesn't occur in Spoken Tamil, but might occur in loan word''
|{{Longdash}}
|{{Longdash}}
|-
|ஓ
|{{Langx|ta|இளங்கோ|lit=Ilango (a name)|translit=Iḷaṅkō}}
|Interrogative, {{Langx|ta|ஆ|translit=Ā}}
|{{Langx|ta|இளங்கோவா|lit=(Do you mean) Ilango?|translit=Iḷaṅkōvā}}
|-
|அ
|{{Langx|ta|இருக்க|lit=To be, to sit ([[Sri Lankan Tamil]] resp. [[Old Tamil|Old]]/[[Middle Tamil]])|translit=Irukka}}
|Affirmative, {{Langx|ta|ஏ|translit=Ē}}
|{{Langx|ta|இருக்கவே|lit=It's there, all right!|translit=Irukkavē}}
|-
|ஆ
|{{Langx|ta|விழா|lit=A festival|translit=Viḻā}}
|Interrogative, {{Langx|ta|ஆ|translit=Ā}}
|{{Langx|ta|விழாவா|lit=A festival?|translit=Viḻāvā}}
|-
|ஔ
|''Doesn't occur in Spoken Tamil''
|{{Longdash}}
|{{Longdash}}
|}
In rapid speech, especially in polysyllabic words: {{Langx|ta|இந்த்யாவுலேருந்து|lit=From India|translit=Intyāvulēruntu}} may become — {{Langx|ta|இந்த்யாலெருந்து|translit=Intyāleruntu|label=none}}, which may then be further simplified to {{Langx|ta|இந்த்யாலெந்து|translit=Intyālentu|label=none}}.<ref name=":2" />

====Consonant sandhi====
In lateral-stop clusters, the lateral assimilates to the stop's manner of articulation, before c, ṇ too becomes ṭ, eg. nal-mai, kal-kaḷ, vaṟaḷ-ci, kāṇ-ci, eḷ-ney > naṉmai, kaṟkaḷ, vaṟaṭci, kāṭci, eṇṇey (ṟ was historically a plosive).

==== Elision ====
In Spoken Tamil the final laterals, nasals or other sonorants may lose the final position. The final retroflex laterals for pronouns and their PNG markers for example {{Langx|ta|ள்|translit=ḷ}} of (female gender marker) are deleted: (To indicate the omitted stop-consonant is covered in parantheses): {{Langx|ta|அவ(ள்) போறா(ள்)|lit=She goes|translit=Ava(ḷ) pōṟā(ḷ)}}.<ref name=":2" />

==== Noun cases ====
In some nouns, sandhi is triggered by the addition of a [[Grammatical case|case]] ending to the stem.

===Sanskrit===
Sanskrit has formalized and systematized sandhi changes; like in all Indian languages, the sandhi changes are also recorded in the written language. There are two categories of sandhi in Sanskrit: internal and external sandhi. Internal sandhi takes place within words, at the junctures of [[morpheme]]s. External sandhi occurs at word boundaries and between members of compounds.<ref name="goldman">{{cite book |last1=Goldman |first1=Robert P. |last2=Sutherland Goldman |first2=Sally J. |title=Devavāṇīpraveśikā: an introduction to the Sanskrit language |date=2011 |publisher=Motilal Banarsidass Publ |location=Delhi |isbn=9788120833753 |edition=Corr.}}</ref>

The basic rule is to make it easier to pronounce words and sentences. Therefore, clashing consecutive sounds are avoided as much as possible. In the case of internal and external vowel sandhi, this means, generally speaking, that two vowels should not come into direct contact. This is avoided by the combination of the two consecutive vowels into a single sound. That can happen in three different ways: by coalescence of the two vowels, by changing the first vowel to a consonant, or by dropping one of the vowels. Similarly to vowels, clashing consonants are avoided by assimilation of either one or both of the juxtaposed sounds.<ref name="goldman" />

The number of sandhi changes in Sanskrit is extensive, these are described in various books on Sanskrit grammar<ref name="goldman" /> and most notably, in the [[Aṣṭādhyāyī]] grammar by [[Pāṇini]]. A couple of examples are given in the following sections to illustrate the kind of changes which occur.

====Examples of external vowel sandhi====
In compounding, if the first word ends with /i, u/ and the second word starts with  a vowel, the i, u become glides y, v, e.g. su-āgata > svāgata.  If a word ends with {{IPA|/a, aː/}} and the second word begins with /i, u/ they become {{IPA|/eː, oː/}}, eg. mahā-utsava > mahotsava; if the latter vowel is long, it becomes /ai, au/, eg. pra-ūḍha > prauḍha.

====Examples of external consonant sandhi====
The [[visarga]] ('{{lang|sa|ः}}' {{IPA|[h]}}) becomes a /r/ before voiced phones, eg. duḥ-labha > durlabha. [[Anusvara]] + plosive makes it a homorganic nasal, before a fricative or /r/ it nasalizes the previous vowel and before {{IPA|/j, ʋ/}} it nasalizes the {{IPA|/j, ʋ/}}.

In come compounds s follows the [[Ruki_sound_law|RUKI rule]], eg. vi-sama > viṣama, pitr-svaseya > pitrṣvaseya.

== See also ==
* [[Alternation (linguistics)]]
* [[Crasis]]
* [[Elision]]
* [[Liaison (French)]]
* [[Linking and intrusive R]]
* [[Movable nu]]

== References ==
<references/>

== External links ==
{{Wiktionary|sandhi}}
* [http://www.sanskritdictionary.com/sandhi/ Sandhi Calculator by Vedic Society]

[[Category:Morphophonology]]
[[Category:Sanskrit words and phrases]]
[[Category:Vyakarana]]
[[Category:Phonaesthetics]]