There are four things we need to know about the perfect, one concerning its meaning, and three concerning its formation. The meaning of the perfect in Classical Sanskrit is identical to the imperfect: both refer to the past. There was once a semantic difference between them (which is why there are two different sets of forms), but this difference has been lost.
Perfect forms consist of three elements:
a) a reduplicative syllable
b) the verbal root in strong or weak form
c) a special perfect ending
Combining a) and b), we arrive at the perfect stem. Note that this stem is unrelated to the present tense stem, and thus there are not ten different classes of perfect formations, but just the one.
Compare the perfect of √विश् ‘to enter’ (given in transliteration to facilitate recognising the individual elements):
Systematically put, the perfect is formed in the following way:
a) The reduplicative syllable comes first and consists of the first consonant and the vowel of the zero-grade root: the reduplicative syllable of √विश् is thus वि-. (See below for more details.) b) The root is strong (i.e. its vowel stands in guṇa) in the active singular (√विश् –› वेश्-), and weak (vowel in zero grade) in all other forms (√विश् –› विश्-), thus displaying the same pattern of stem gradation as found in athematic verbs in the present tense (–› Chapter 18).
c) The endings are unique to the perfect (but note the usual -v- in the 1st Du, -m- in the 1st PL, and similarity between 2nd and 3rd DU endings). Roots ending in a consonant insert an -i- before endings that begin with a consonant.
To analyse two forms explicitly:
Notes on Reduplication
The example above features the simplest kind of reduplication, in which the consonant reduplicates in identical form. Yet note the various possible changes, which are identical to those in Class III verbs (–› Chapter 18).