[fr. Skt. संिध saMdhi2, `putting together' < sam- `together' (cf. English `same') + dhaa- `put, set' (cf. English `do')]

An interesting aspect of sandhi in Sanskrit is that writing as well as speech takes the transformations into account. For example, some external sandhi (i.e., sandhi occurring between words or at the end of words) rules:

And some rules of internal sandhi:

  • Between two vowels, the first of which is short, nasals are doubled.
  • a + i becomes e, a + e becomes ai, a + u becomes o, a + o becomes au; and for any vowel V, i + V becomes yV, u + V becomes vV, R + V becomes rV, and L + V becomes lV.
  • When it follows any vowel other than long or short a and precedes a voiced sound (a vowel or voiced consonant) -H becomes -r.
  • Bartholome's law: a voiced aspirated stop followed by t or th becomes a voiced unaspirated stop followed by dh. Often, but not always, -h- is treated as -gh- for the purposes of this rule---it depends on the etymological origin of that -h-.
  • Ruki: s becomes S after r, u (long or short), k, or i (long or short) when not followed immediately by r or R. It does not have to immediately follow the r, u, k, or i.

It requires some experience to break sandhis when translating a Sanskrit text. In material for beginners, texts are often presented in pada form, where sandhi is ignored. PaNini was one of the first linguists to fully analyse Sanskrit sandhi.

1: This node uses the Harvard-Kyoto system of transliteration from devanagari to the latin alphabet.