*deseti* and *desayati?* (both active voice) and the passive

This question and answer remain there, but the discussion that follows has been moved from the Lesson 4 thread.


Aren’t these the passive forms (-ya affix)?

1 Like

No, passive verb forms are a different animal. See Gair and Karunatillake, Ch.IX, section 4 (on p.127).
You’re probably mixing it up with the fact that often passive verbs are formed by adding -iya- or -īya- to the present stem of a transitive verb, e.g., detidīyati, pūjetipūjiyati, karotikariyati or karīyati. Then there are passives like hanati → haññati, bhindati → bhijjati. One really just has to learn the pairings as they come up.

1 Like

That’s interesting, because the Combined Pali English dictionary says it’s passive:

pass. pr. 3 sg. | desiiyati
of deseti
See also: deseti

Also, Kaccāyana gives this form as the “reversal of Attanopada vibhattis into Parassapada vibhattis” (sutta 518). In other words, they “look like” active verbs but are actually passive because of the ya affix. Strictly speaking, though, they originated as kamma specific attanopada terminations but have been retransformed into parassapada (active).

518, 446. Attanopadāni parassapadattaṁ.
[V] Attanopadāni kvaci parassapadatta’māpajjante.
• vuccati. • labbhati. • paccati.
Kvacī’ti kimatthaṁ? * karīyate. * labbhate. * paccate.

Edit: I realised the original word quoted by @acala was desayati and not desīyati. Not too sure about this one.

1 Like

My friends, this discussion is getting way beyond the scope of the class, please continue in a new thread. :pray:


I think I’ve found the answer!

According to Duroiselle:

  1. The Verbs of the Seventh Conjugation form their Special bases by adding to the root aya, which by contraction may be replaced by e. The forms in e are more commonly met than those in aya.

So yes there can be two versions of the verb with an e and an aya.


Yes, this is exactly what I was saying to you, @christie and what I originally answered @acala - verbs in this conjugation can have two forms , ending in both -eti and -ayati, and the meaning is identical. You confused the passive ending -iyati and īyati and took this discussion off on a different tangent.

Also, as @sujato wisely counsels us, introducing discussions based on Kacccayana with terms like attanopada and parassapada are both way off topic, and potentially offputting, for beginning Pāli students. So, please post those elsewhere. :pray:


No problems @johnk, I wasn’t suggesting you were wrong, but was seeking clarification, hence the question. It wasn’t clear to me that the parassapada passive forms had to start with ī or i as Kaccāyana only mentions the ya affix (the examples given in sutta 518 do not include iya) but thanks for clarifying that.

You have to remember I am trying to learn Pali along with everyone else. I apologise if my questions may put off other people, but I seek guidance as much as anyone else. In the end we all learn from each other. I thought @acala 's question was very relevant because I had the exact thought myself.


Thanks @gillian for splitting the posts into a new thread.

So, hopefully without the fear of “putting off” anyone else on the Pali course, let me (re)state why I am asking the question.

Based on my (arguably, very possibly faulty) understanding of Kaccāyana, sutta 518 does not require the passive form to have an i or ī in front of the ya affix.

This is further confirmed by Māgadhabhāsā:

Passive voice affix ya is used to express the stative passive, with or without i-interfix

Based on my extremely error prone reading of the Kaccāyana production rule, it seems entirely possible that desayati could be a passive form. Even if we started from desaya, we would have an intermediate form of desaya+ya+te and then eventually the yate will get elided leaving us back to the original desayati. Of course the more obvious production will be dese+ya+te leading into a possible desayati

Just like we have vuccati which is a passive verb according to Kaccāyana, even it contains neither ya or iya.

Of course, I am not saying I am correct, and it is almost certain I am wrong, hence the question.

Of course, the simple answer is that we already have desīyati as a passive form, and it is possible that the infix i has been added to disambiguate the passive from the active form. However, I can’t find iya as a production rule in Kaccāyana, which again leads to my question in the first place.

My apologies if I am somehow offending anyone in a genuine quest for knowledge.


Thank you John Kelly and Christie for helping to answer my question. :pray:t5:

I feel like I’m learning a lot from everyone else’s questions as well, so thanks to all those keen questioners! :grinning:


Going further down the rabbit hole, as I research more into the passive form …

In particular, I am trying to understand why @johnk says the passive form involves adding -iya - or -īya when there are clearly many examples of passive words which do not fall into this pattern.

This post is probably more to remind myself of what I’ve learnt rather than an attempt to answer the original question, so feel free to ignore it.

So it seems Kaccāyana has no less than 5 rules (or sutta) relating to the construction of passive verbs.

The first rule (440) is a “generic” one saying the ya affix is to be applied in conjunction with the Attanopada Vattamānā vibhattis to contruct a “middle” voice verb with a sense of the passive.

This results in a variety of words, not just words with iya or īya in them:

• ṭhīyate. • bujjhate. • paccate. • labbhate. • karīyate. • yujjate. • uccate.

Note that as per rule 518, these verbs are then translated to the parassapada endings such as ṭhīyati, bujjhati etc.

Kaccāyana then gives three additional rules for specific instances:

Rule 441 says sometimes the last component consonant of the root together with the ya affix changes into the ca, va, ya group consonants. This is the case for words like vuccate, uccate, paccate etc. where it appears the ya affix has been elided.

Rule 442 is the only rule where the infix i or ī is inserted before the ya affix, resulting in words such as karīyate, gacchīyate etc. This is the one that @johnk is referring to.

Rule 443 states that sometimes an additional morphological change called pubbarūpa causes the result to have double consonants, resulting in words like vuḍḍhate, phallate, dammate

So out of these 4 rules, only 1 require the use of the i infix.

I note that this is what G&K says regarding the passive form on page 127:

There are are some verbs in Pali that have a passive sense. That is, the subject of the verb is that which is effected or brought about. Thus vuccati ‘is said’, dassiyati ‘is seen’ etc. Such passive verbs are commonly related to transitive verbs. Often, the passive verb will be formed by adding -iya- or -iya- to the present stem of the transitive verb, which may undergo further changes of form. Hence deti ‘gives’, dīyati ‘is given’, pūjeti 'worships’, pujiyati ‘is worshipped’, karoti ‘does’, kariyati or karīyati ‘is done’ etc. Sometimes the passive verb has a double consonant while the related transitive verb has a single one or a consonant cluster: thus hanati ‘kills’, hannati 'is killed’, bhindati ‘breaks’, bhijjati ‘is broken’, pacati ‘cooks’, paccati ‘is cooked’ , etc. (these double consonants occurred because some passives were formed earlier by adding -y-, which doubled the preceding consonant, sometimes changing it, and disappeared.) In any case, in Pali these pairings must be learned as they occur, since the relationship may be more or less transparent due to the changes that have taken place.

So it would seem even G&K agree not all passive verbs have to include iya in them, they only assert it is “often” the case (which I question, given Kaccāyana’s rules).

Note also G&K’s speculation about the origins of the consonant doubling does not agree with Kaccāyana. I am not going to comment on who is more likely to be correct (but of course I have a view).

As an aside, others also seem to think the rule is to insert iya rather than the simpler ya, so I can understand why @johnk is quoting it. The translator of Kaccāyana (A. Thitzana) also made this mistake by stating that the infix i must always be inserted before the ya affix (on page 608):

The formula of passive voice form is: (Try to remember)
Root+i or ī +ya+relevant attanopada vibhatti. (4 parts)

and then (somewhat humorously) constructs the passive form of pacati (to cook) as pacīyati which is incorrect (this is a completely different word meaning “is accumulated; is increased; is heaped up” according to DPD). Kaccāyana correctly gives the word as paccati.

I guess the moral of the story is beware of what you read in Pali textbooks, they may not always give the correct information.

Now, going back to desayati - although this would appear to conform with rule 440 it doesn’t appear to follow any of the subsequent production rules (441-443). On a hypothesis (unproven) that 440 needs to be applied in conjunction with 441-443 (Kaccāyana doesn’t actually say this, so I am just speculating) then it would seem unlikely this is a passive form. Of course, we already know it isn’t from Duroiselle, but I wanted to see if we can eliminate the possibility just from the production rules - and the answer is “maybe”.

Wow, this is some advanced Pali indeed.

Aleix Ruiz-Falques had an online class on the Kaccayana Grammar, which can be found on youtube.

He also ran an online class on Saddanīti, which I participated in a bit. Entering into the world of the ancient grammars is difficult, at least for me. There is a whole method of presentation that needs to be learned, a very different way of thinking about grammar from the west.

For instance, what Ruiz-Falques calls ‘metarules and megarules. ‘
But surely very rewarding.

PS. I would definitely have a look at the Ole Holten Pind edition of the Kaccayana and Kaccayanavutti (PTS) along side the Pariyatti edition done by U Thitzana.


The aya forms are sanskrit(ic) forms. The norm in Pāli is to contract Sanskrit ai, ay & aya to e (and similarly au, av & ava to o), but this doesnt happen across the board so there are some leftover forms with aya and ava in them.

desayati in Pali compares to the sanskrit causative form deśayati. The causative is not the usual passive, so look at causatives in detail for greater clarity.


You are right - it does look like causative forms based on the causative affix ṇaya. Although (based on my extremely limited grasp of the transformation rules) adding ṇaya to either the root dis in either dese or desaya stems will not result in desayati (some vuddhi “voodoo” needs to happen when the is elided in accordance to rule 523 - my best guess is the word becomes desāyati based on my reading of rule 483).

One day I hope to master sandhi and the vuddhi rules, but today won’t be the day.

One could entertain the notion of constructing a causative passive verb using causative affix ṇe plus the passive affix ya but I think that will result in something like desīyati which is already the passive form, so I am obviously doing it wrong, again. That’s why my original thinking was desīyati was the causative passive form, and desayati was just the normal passive form.

dis + (ṇ)aya + ti = d(e)s + aya + ti = desayati
The anubandha ṇ is elided (so ṇaya becomes aya), and the vṛddhi of the ‘i’ would make it ‘e’ (so dis becomes des). The present tense (third person, singular) termination ti stays as is, so we get desayati.

Your hypothetical form desāyati would require the verb to end in ‘a’ before the causative ṇaya is applied, so that the sandhi of a & a would make it ā, but here that doesnt happen as the verb ends in ‘s’ so dis + aya = desaya (after vṛddhi changes the i to e), not desāya.

On another point, you speak of the causative passive, which I am not aware of the existence of. Normally causatives only take active (parasmaipada or ātmanepada) endings.

Thanks I agree with you. After I posted, I realised I lengthened the wrong vowel and should have applied vuddhi on the first vowel. So you are correct.

Edit: although thinking about it perhaps the production should be on the disaya stem, so
disaya + ṇaya + ti = desayāyati???

I think in an instance like that, it’s probably more prudent to choose a different causative affix.

1 Like

No, it is normally to the root’s vowel (and not to a vowel of the causative affix) that the vṛddhi is applied. This is also seen in taddhita derivations where it is the first vowel that takes the vṛddhi.


Yes, but in my rather foolish attempt the elongated ā is my attempt at sandhi from joining the two aya together, so it’s sandhi not vuddhi, if that makes any sense.

1 Like

On this point, the word actually had a ‘ya’ in it in its pre-Pali (i.e. Sanskrit) stage - so the passive form in Sanskrit is ucyate (and it has an ātmanepada ending as you would observe, which passive forms have by default in Sanskrit). The cya in ucyate has gotten phonetically altered in Pali to ‘cca’ thus the Pali form as we have it now (vuccati) violates the general rule for passives.

Things like these led me to suppose that Pali (of the Pali EBTs) is probably not the phonetic language it is widely assumed to be, but rather a non-phonetical orthographic representation of the spoken language that was originally pronounced differently.


Sorry for my lack of understanding, does this mean you’re thinking that Pali (as we have it today) was written down after the ‘original’ pronunciation had shifted over time?