I think that we all have had experiences where we pick up an old text for the umpteenth time, and looking over an old passage, see something new or interesting within it. Recently, I noticed something interesting while reading though Vātsyāyana’s Bhāṣya, in the section where he defends the Veda as a special case of authoritative testimony (Nbh 2.1.57-68). There’s a lot here of interest, including an expansion of his account of what makes a speaker authoritative (āpta), and the development of a strategy that articulates the status of non-ordinary modes of knowing as special cases of our familiar, time-tested pramāṇas. Common phrases in the section include variants of yathā loke. . . (“just as it is in this world”).
What just caught my eye is his response to the objection that anuvāda passages of the Veda are defective owing to the fault of repetition (NBh 2.1.66-7). I leave it to Andrew Ollett and Elisa to provide more context for the literary fault of repetition (if they so choose!). Suffice to say, it is minimally taken to be bad form to repeat something already said. From the perspective of pramāṇa-śāstra, we know that on one fairly common—if problematic—construal, genuine knowledge sources must give us something “new”. Whatever his motivations, Vātsyāyana’s interlocutor tells us that the fault with repetitive passages in the Veda is that they repeat words whose objects are “already known”, and are thus as useless (anarthaka) as ordinary repetition.
In his response, Vātsyāyana distinguishes between useless repetition and purposive repetition. Simply repeating something is [sometimes] useless, but there are clearly cases when repetition has meaning (arthavat).
Some examples:
“Hurry! Hurry!”, conveys the notion that one must increase one’s effort and pick up the pace.
“He’s cooking, he’s cooking”, conveys the fact that the activity of cooking continues uninterrupted.
“The god’s rain avoids Trigarta, it avoids Trigarta” conveys complete lack of rain on the region.
While this is not developed into an account of implicature (or, for that matter, speech acts), it is a pleasant, early recognition of the wider ranging power of language to produce contentful awareness beyond mere denotation.
Hi Matthew,
I think that pragmatics is very much a part of Indian linguistical investigations since at least Pāṇini (I am not claiming that you are saying the opposite, just elaborating on your post). His is the only grammar I have ever come across which discusses topics such as the prolongation of the last vowel while shouting at one in order to call him (a phenomenon we are probably all familiar with, but which is usually not “grammaticalised”), or the way one takes leave. Further interesting hints at this different awareness of what belongs to the realm of grammar can be found in Mīmāṃsaka discussions about language (about which see also this post by Andrew, which elaborates also on anuvāda ‘repetition’). In some of them, one finds, e.g., discussions about what is really meant by sentences such as “Your wife just gave birth to a son” and the role of one’s smiling face in conveying the news.
Very interesting post, Matthew. I had formerly subsumed this discussion within the phenomenon of perlocutionary effect of a repetition speech-act, but your second and third examples are much better described as conversational implicature. Two further questions then arise: Which conversational maxims are flouted when such implicatures are created? And how would one account for such cases within Nyāya theory of meaning: lakṣaṇā or vyañjana? Or should we rather say that here is one of those cases where an insight in the early texts is somewhat lost sight of in later, more systematic, theory?
Thank you for your kind replies, Elisa and Jonardon.
Elisa thank you for your reference to those fascinating portions of Pāṇini, and Andrew’s excellent post. And yes, I meant early *Nyāya*. Speaking also to Jonardon’s final reflections, what struck me is that here is an interesting passage in early Nyāya that could have perhaps been the basis for earlier, more developed systematic work in early Nyāya philosophy of language. Perhaps it is Stephen Phillips’s influence speaking here, but as a Nyāya sympathizer to some degree, it seems that one place that is underdeveloped (in the earlier school at least) is in the account of indirect meaning, which Nyāya tends to limply frame as merely a kind of inference.
Also, as a quick reply to the question of which Maxims are being flouted, perhaps third submaxim of Manner, to avoid prolixity. The second submaxim of Quantity may be a candidate, but one can perhaps doubt whether saying a word again is adding more information.
I am going to look up some of the later commentaries on this passage to see if anything interesting comes up. Will post here if they do.
In the meantime, relevantly (and completely coincidentally) my friend in the UTexas-sampradāya, Malcolm Keating will be posting a guest post on his work on Mukulabhaṭṭa and pragmatics in Indian philosophy. It will be posted later tonight or tomorrow.
A few more notes.
Uddyotakara says that this sort of repetition it faultless because it is the “cause of a distinct cognition” (pratyaya-viśeṣa-hetutvāt). What is this distinction exactly? He notes that there are many ways to unpack it.
Hearing the imperative “cook” for the first time, one has the cognition “cooking must be done”.
Hearing it a second time, one has a different cognition: “it is definitively settled (avadhāraṇa) that cooking must be done.” Or, “I must do the cooking”. Or, “I must continue cooking.” Or, the cognition is recognition that one is being requested “I am the one being proposed to do this.” Or, “I had better hurry.”
In these cases, Uddyotakara says, the new cognition (pratyaya) generated in the hearer is the same as that in the mind of the speaker.
It would have been nice if he had a purely declarative example of this phenomenon. Will continue to go down the list of commentators to see. . .
Very interesting. Still another possibility, now, is that the second utterance has a different literal meaning, that the context-sensitive rules, including āsatti, sometimes require a repeating utterance to have a literal meaning distinct from the repeated utterance.
Jonardon, are you speaking of akanksa, yogyata and sannidhi (or asatti)? If so, my understanding is that they apply to sentences (and, thus, to each of the two sentences pacatu, pacatu, and not to their complex). What helps to get distinct meanings for each of them is textual linguistics as implemented by Mimamsa, in particular prakarana.
It’s interesting, Elisa, to think about why the concern with Asatti, and this sort of analysis regarding semantic and syntactic appropriateness is typically performed on the level of the sentence. If anything, these examples of repetition, as Jonardon suggests, do indicate that we must employ similar sorts of considerations when looking at groups of sentences, not just groups of words which form sentences. Is there a deep theoretical reason to simply concern ourselves with the latter while considering these criteria?
For example:
Give my son the baseball. He is great at cooking it. (failure of yogyata)
Give my son the baseball. They is great at pitching it. (failure of akanksa)
Give my son the baseball. (Five minute pause. . . ) He enjoys having a catch. (failure of sannidhi)
Here are the three rules violated when looking at how sentences relate to each other.
Matthew, textual linguistics is one of my favourite topics, so I am glad you raised the issue. Asatti etc. are defined as the characteristics used to identify what a sentence is. I think Mimamsakas did the right thing when they did not propose to extend them to the identification of a single textual unity. Instead, they defined it in relation with the principal sentence. And in fact if you take a portion of a textual passage, you might easily find two sentences which seem to lack yogyata, but are still to be deemed parts of the same text because they are both connected with the main sentence. E.g.:
“(1)This is my son. (2)He is a great cook. (3)My wife thinks (4)children should learn to swim properly before they turn six. (5)Hence, he also learnt to swim early in life.”
Here, (2) and (3) lack yogyata among each other, but they are both connected with (1).
Concerning asatti, for the same reasons a sentence which is far away within a certain treatise might have to be connected with one which had been uttered well before (think of Panini!) and we would not want this not to occur because of the application of asatti.
Re. akanksa, in your example akanksa is violated in the second sentence itself. “te kausalo ‘sti” would not work no matter what precedes it.
I’m sure this is well known, but such kinds of repititions are very common while speaking in a vernacular language (hindi, bengali, etc.)
^repetitions
Yes, and this might have been one of the reasons why Panini, Uddyotakara and so on noticed the phenomenon and looked for a rationale for it.
Thanks, Elisa. This helps. The idea is that by definition, these are the criteria for sentences. We have other ways of determining larger units of linguistic communication that may use some kind of analogous considerations and more.
Vacaspati doesn’t have much to add to Uddyotakara’s commentary. He just emphasizes that repetition involved in saying “go quickly, go quickly” functions in a way that is akin to saying “go quicker” and thus adds a distinct cognition–that one should put forth exceptional effort– over and above the cognition produced by merely saying “go quickly!” once.
He also elaborates on the need for the cognition produced in the hearer to match that which prompts the speaker of the repetitive utterance, adding the (perhaps obvious) point that otherwise, the hearer’s cognition would merely be error (bhrAnta) owing to incongruity.
He adds an example at the end of his commentary: “the village, the village is delightful” (grAmo grAmo ramaNIyaH).
Since Matthew mentioned it: ālaṅkārikas always distinguished “mere repetition” (which is a literary fault) from certain kinds of “value-added” repetition. One difference was whether the repeated element had the same meaning in each case; according to Daṇḍin (3.137ff.), repetition could be an adornment if it conveyed some additional meaning (atiśayaḥ, similar to the cases that Matthew mentioned: Daṇḍin’s example is a verse in which a woman is repeatedly said to be “overwhelmed”—hanyate, hanyate—by love). Regarding the unit and scope of repetition, what ālaṅkārikas had a problem with was repetition of meaning rather than sound, but later ālaṅkārikas (I am thinking of Ruyyaka) recognized that the reader could “intercept” the meaning at various stages of its operation, so that what seemed to be a “mere repetition” turns out not to be.
Thank you for this, Andrew!