What is the minimum number of forms that one needs to memorize in order to be able to fully conjugate a Sanskrit verb? I've found differing answers: some say six, others ten, and in fact, I personally had thought that it was sufficient to know the class number and the voice (parasmaipada or ātmanepada) (and whether the verb was seṭ, aniṭ, or veṭ, but I also thought that could be derived based on the ending letter and class number of the root, with a number of exceptions to the rules of course).

Part of what has made it such a difficult question to answer for me (besides the fact that the idea of principal parts never really belonged to traditional pedagogy, but rather to that of Latin and Greek) is that sometimes the so-called "principal parts" can actually be derived from the root through rules, however complex they may be. For example, the perfect tense form is often given as a principal part, yet I thought it could be derived through the rules of reduplication (e.g., √arh [1P] has initial a followed by a conjunct consonant, so when it comes to reduplication, √arhān-arh-a ["he deserved"]). So, exceptions to the rules of reduplication, seṭ/aniṭ/veṭ identification, etc. aside, how many principal parts do Sanskrit verbs have? And how reliable are these rules I mentioned above?

  • Many of the Latin and Greek principal parts can be derived from the root through rules, too. The problem is knowing which rules to use, since verbs vary considerably. That's what principal parts are for -- they show the results of the relevant rules so you can identify the ones involved. – jlawler Jul 20 '17 at 14:57
  • @jlawler For Latin, almost all the 1st conj. verbs are "regular" in their principal parts, but the other conjugations (3rd in particular) are not so regular. As you said yourself, you need the principal parts for such verbs, because there are too many "irregularities" for there to be a rule (exact, as opposed to a "guideline"). The difference with Sanskrit is that, to my knowledge, there are "exact" rules, not generic guidelines (e.g., for 2nd conj. verbs, the first-person singular perfect active form will often end in -ui), to derive all the different forms from the root. (1/2) – user67444 Jul 20 '17 at 16:35
  • @jlawler Yet some sources still identify "principal parts." I'm just not sure whether these are actually necessary for conjugation, or they simply are given for convenience. And if they are necessary, how many are there? For example, although reduplication occurs in the Sanskrit perfect/future―again, as far as I know―in a fixed, predictable manner, the rules have been the source of many nightmares (reduplication occurs with 2nd and 3rd conj. Latin verbs in the perfect system, but I don't know whether one can always derive the exact perfect stem with the same regularity). (2/2) – user67444 Jul 20 '17 at 16:44
  • There's four basic varieties of Latin perfect stem -- reduplicated (dedī), sigmatic (dīxī), lengthened (iēcī), and digamma (thematic amāvī; athematic habuī). Call them 4 perfect inflection classes, like the 5 present inflection classes we call the first thru fifth declensions; these classes interconnect in all kinds of ways. – jlawler Jul 20 '17 at 18:41

As you may already know, there are 10 tenses/moods in Sanskrit (11 if you include the tense that is used only in the veda). Of these, you will see a similarity in the forms of 4 tenses/moods since they are all derived in a specific manner. Likewise forms of the remaining 6 tenses/moods are all derivable similarly.

If you take up verbs of the first group (called bhvaadi-gana भ्वादिगणः) and examine their conjugations in the 4 tenses/moods लट् (present tense), लोट् (imperative mood), लङ् (past tense) and विधिलिङ् (imperative/potential mood) (excuse me I don't have the best translations for these tenses/moods), they all follow a specific pattern. These are called saarva-dhatuka-lakaraaha सार्वधातुकलकाराः Knowing the 9 forms (singular, dual and plural for each of the first, second and third persons) will enable you to guess forms for other verbal roots in these 4 tenses/moods. As long as you are aware of certain rules that apply modifications to the verbal base, you should be able conjugate all forms with ease. The common factor for conjugating verbs in these 4 tenses/moods is that there's a suffix that interposes itself between the verbal root and the tense. For e.g

भू + लट् - भू is the verbal root and लट् is the lakara for present tense
भू + तिप् - लट् is replaced with तिप् for first person singular
भू + शप् + तिप् - शप् is this new suffix
भू + अ + ति - The indicatory letters are dropped
भो + अ + ति - A modification to the verbal base
भवति - When two vowels are next to each other, they are modified accordingly.

If you pick verbs from a different group (say the second one called adaadi-gana अदादिगणः), their conjugations will differ only in that the intervening suffix between the verbal base and the person/number suffix changes.

Of course, there are always exceptions to these generic statements and only a study of Panini's grammar will equip one with the tools to conjugate any verbal root for any tense/mood.

The second category is the remaining 6 tenses/moods are called aardha-dhatuka-lakaraaha आर्धधातुकलकाराः They are

  • लिट् (some kind of past tense)
  • लुट् (some kind of future tense)
  • लृट् (some kind of future tense)
  • आशीर्लिङ् (benediction)
  • लुङ् (some kind of past tense, probably aorist if my memory serves me right)
  • लृङ्(impossibility of action)

In conjugating verbs for these tenses/moods, it should be noted that a very specific suffix falls in place between the verbal root and the suffix that indicates person/number. For e.g

भू + लुट् - भू is the verbal root and लुट् is the lakara for future tense.    
भू + तिप् - लुट् is replaced with तिप् for first person singular
भू + तास् + तिप् - A new suffix तास् appears between भू and तिप् This suffix is the same, regardless of which verbal root you pick (from any of the 10 groups/गणाः)
भू + तास् + डा - तिप् is replaced with डा 
भू + इतास् + डा - The verb being सेट्, the augment इट् is added to तास् 
भू + इत् + आ - A portion of तास् is elided
भो + इता -  A modification to the verbal base
भविता - When two vowels are next to each other, they are modified accordingly.

In these 6 tenses/moods, the knowledge of whether a verb is सेट्/अनिट्/वेट् is very important. In लिट् the verb is always duplicated and undergoes many modifications. While it is easy to spot the duplication when you come across a conjugated form in reading, deriving the same (or sometimes identifying the verbal root itself) from a verb is non-trivial. In लुङ् too, there are many rules that govern the modifications to verbal bases and sometimes to suffixes.

Memorizing the conjugations of all verbs in all tenses/moods is a daunting task. One of the main benefits of studying Panini's grammar is that conjugating a verb in any tense becomes trivial with sufficient practice and knowledge of the rules. Someone who hasn't undergone such training may be able conjugate verbs to some extent but will always fall short when special knowledge is required in some places.

A sanskrit verb has just one part. The conjugated part is derived by adding one or more suffixes to the verb.

| improve this answer | |
  • 1
    This doesn't directly answer my question. So what you're saying is that knowing the dhātu, class number, voice, and iṭ-ness is enough? Everything else can be derived from this with perfect regularity (true irregulars aside)? And I thought that iṭ-ness can also be determined with certain rules, although there would be numerous exceptions? – user67444 Jul 20 '17 at 17:06
  • You have the list narrowed down correctly - one needs the dhatu, its class (gana), whether it takes atmanepadam or parasmaipadam and whether it is seṭ, aniṭ, or veṭ. The iṭ-ness is indicated by swara marking on a vowel (that isn't elided) in the dhatu, so one will need to know the swara to determine iṭ-ness – linuxfan Jul 20 '17 at 17:16
  • Just a little confused―what does "swara marking on a vowel" mean? – user67444 Jul 20 '17 at 17:40
  • A swara simply means a special intonation on a vowel. Roughly speaking, it can be normal (udatta), low (anudatta) or a mix of low and normal (swarita). They are somewhat akin to musical notes. In dhatus ending with consonants, if the swara marking on a vowel is anudatta it will be called anit. For dhatus that end in a vowel, there is a 2-line verse that lists them out and labels them seṭ. This way the iṭ-ness of all dhatus is known. – linuxfan Jul 20 '17 at 22:06
  • Got it, thanks. Do you know what the verse is? And in Antoine's A Sanskrit Manual, he gives rules for determining whether the verb was seṭ, aniṭ, or veṭ that don't have to do with pitch accent. For example, his first rule is: "All the roots of the tenth conjugation and of the derivative conjugations, all the roots ending with consonants, and all the roots ending with long vocalic 'r' and long 'u' insert 'i' before the terminations of the periphrastic future." He then gives dozens of exceptions. Is this rule valid? What is its relationship to the rules relating to accent that you gave? – user67444 Jul 20 '17 at 22:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.