Epistemology of Textual Re-use in the Nyāyamañjarī

The epistemology of śabda is one of the main themes in Bhaṭṭa Jayanta’s Nyāyamañjarī, and, in the hypotheses explored in this paper, also the conceptual basis of Jayanta’s textual re-use. The sixth chapter of the Nyāyamañjarī contains a debate between Vaiyākaraṇas and Mīmāṃsakas who, respectively, advocated an holistic or atomistic theory of language. Selected Jayanta’s re-uses from Vyākaraṇa, Mīmāṃsā, and Nyāya sources are here surveyed and analyzed, with a focus on their meaning and on the context. The method of analysis is partially following Moravcsik’s scheme for a classification of citations, as well as Small’s classification by symbolic functions. By re-using texts Jayanta not only imparted authority to his own arguments, but also reassessed the relation of his tradition with other ones. Re-used ideas and words stand for symbols of those authors’ tenets, and those authors represent symbols of their respective traditions. Moreover, by quoting a certain author Jayanta often anointed him with a symbolic status of trustworthy authority, and his statement with a status of śabdapramāṇa.


Bhat˙ṫa Jayanta and the Nyāyamañjarī (NM)
The Nyāyamañjarī (henceforth NM) is a treatise on the tenets of Nyāya, the system of epistemology, dialectics and logic traditionally rooted in the Nyāyasūtra (NS). It was written by Bhaṫṫa Jayanta in the ninth century CE. 1 The NM is composed of 12 chapters. In NM 3 to 6, about one third of the whole work, Jayanta examines in depth the epistemology of śabda (on this term, see Sect. 2). The present paper is grounded on a survey of a selection of re-use of texts in the sphoṭa section of NM 6. By "re-use" I mean the adoption of previously existing textual passages and ideas, as a general and widely encompassing term. Verbatim re-use will be here called "quotation", in cases where I am confident that Jayanta is consciously quoting existing texts.
This particular section of the NM consists of an investigation on the aetiology of śabda-produced knowledge, in which Jayanta stages a debate among Vyākaraṅa and Mīmāṁ sā schools. There are three main reasons behind the choice of this particular section: first, my critical edition of this portion, based on all the available manuscripts, 2 is at an advanced stage, so its text is for the purposes of this paper more reliable than that of other parts of the NM; secondly, with a well structured sequence of arguments and counter-arguments, such as that of this section of the NM, it is easier to assess the function of textual re-use in Jayanta's system; thirdly, in the course of the editing process, I had already identified a significant amount of re-used passages and their sources, which allows me to invest time and energy in evaluating re-uses, rather than in tracing them.

Truth-Values, Epistemic Function, and Symbolic Effect of Re-used Texts
In the analysis and classification I will mostly focus on the quality and the context of re-uses. Following in part the scheme for a classification of citations devised by Moravcsik and Poovanalingam (1975) and the classification by symbolic functions, introduced by Small (1978), I will classify re-uses in the NM by asking the following questions about Jayanta's intentions: (1) Truth value: Is the re-use confirmatory or negational? In other words, did Jayanta consider the re-used passage true or false? (2) Symbolic value: Which tradition, author, or idea does the re-used text stand for? Moravcsik and Poovanalingam (1975, p. 88) envision four dichotomies in their scheme: (1) conceptual/operational, (2) organic/perfunctory, (3) evolutionary/ juxtapositional, (4) confirmatory/negational. Of these four, (1) is ignored here, since it is specific to the natural sciences. (2) and (3) would both be relevant, but for want of space they will need to be addressed elsewhere. (4) is here discussed and implemented in terms of truth-value.
As for the symbolic function, it was introduced by Small (1978, p. 328): [Previous] studies have missed the role citations play as symbols of concepts or methods. This cognitive function arises from the formal requirement imposed on the scientist-author of embedding his references to earlier literature in a written text. This leads to the citing of works which embody ideas the author is discussing. The cited documents become, then, in a more general sense, 'symbols' for these ideas.
[…] In the tradition of scholarship, the references are the 'sources' which the author draws upon to give further meaning to his text. Reversing this view, as I am suggesting here, the author is imparting meaning to his 'sources' by citing them.
[…] In general, [by 'symbolic'] I mean that an object 'stands for' an idea; for citations, the cited document is the 'object' and the 'idea' is expressed in the text which cites it.
When applied to the re-use of texts in the NM, the term "symbolic" can encompass at least three different levels, the first two intended by the author, the third probably accidental: 3 • By re-using texts, Jayanta not only imparted authority to his own arguments, but also reassessed the relation of his tradition with other ones. Re-used ideas and words stand for symbols of those authors' tenets, and those authors represent symbols of their respective traditions.
• By quoting a certain author Jayanta may intend to assign him a symbolic status of trustworthy authority (āpta), and to his statement a status of śabdapramāṇa.
• By arranging textual passages in a new order (see Sect. 4), Jayanta endowed them with new meanings.
While analyzing the context and the structure of re-uses, I will therefore try to identify the tradition, author, and concept symbolized by the cited text.

Contents of this Paper
The epistemology of śabda is not only the subject matter of NM 6, but also, in the hypotheses explored in this paper, the conceptual basis of Jayanta's re-use of sources. Therefore, in Sect. 2 a brief sketch of the epistemology of śabda is provided.
In the sphoṭa section of NM 6 Jayanta hardly ever re-uses material from his own tradition. The reasons behind this absence of Nyāya sources are explored in Sect. 3.
Further clues about Jayanta's attitude in relation to his own tradition can be derived from some introductory verses of the NM. The most relevant passages are examined in Sect. 4.
Section 5 is a synopsis of the structure of sphoṭa section of NM 6, needed to facilitate the comprehension of my analysis of the style and the epistemic value of re-uses. Arguments for and against the sphoṭa theory are staged as an articulate dialog between Vaiyākaraṅas and Mīmāṁ sakas. Accordingly, in Sects. 7 and 8, textual re-uses in the whole sphoṭa section are surveyed in two distinct parts.
For a more exhaustive assessment of Jayanta's modality of re-use, I decided to compare his treatment of Mīmāṁ sā and Vyākaraṅa sources with that of Nyāya ones, and since Nyāya sources are almost absent in NM 6, in Sect. 9 some NS and NBh reuses from other parts of the NM are also examined.

The Epistemology of śabda
In Nyāya an authoritative instance of śabda must come from a source that satisfies the criteria of trustworthiness. If textual re-use has epistemic purposes, therefore, clarity about the identification of the source and its authoritativeness becomes particularly relevant. The formal style of a quotation, moreover, may reflect an appreciation or criticism of the source, in conformity with the assessment of a given quotation as representing an accepted truth or a false tenet. To contextualize the survey of quotations in the NM, some basic notions related to śabda epistemology are discussed next.

Comprehension, Knowledge, and Their Object
In most cases I will leave untranslated the terms śabda and artha, in want of English equivalents that can give justice to their technical and polysemous usages. Also, the two terms are not always treated homogeneously in Nyāya, Mīmāṁ sā and Vyākaraṅa. As a general orientation, in the present context śabda is to be understood as "epistemic linguistic expression", i.e., "linguistic expression having the purpose of conveying knowledge", and its artha as "object of epistemic linguistic expression". Although śabda is often rendered with "word" or, at best, "linguistic expression", and artha with "meaning", such equivalents fail to convey the epistemic import of these terms.
In the use of Sanskrit expressions such as śabdabodha, śabdārthapratīti, śabdārthasampratyaya, padārthapratipatti, vākyārthapratipatti, etc., the distinction between understanding words and knowing from words is not explicit and is philosophically problematic. As noted by Matilal and Chakrabarti (1994, Introduction, pp. 9-11), in Nyāya literature the implied assumption is that verbal knowledge (śabdabodha) is "knowledge derived from the words of a truth-teller", rather than "understanding of words", although in English translations a tendency to render śabda and artha in the latter context seems prevalent.
The conflation of "comprehension of words" and "knowledge from words" that we witness in Sanskrit literature, however, may not necessarily be a philosophical flaw, if we consider how comprehension is a fundamental and inextricable part of word-derived knowledge. 4

Śabda in Nyāya
In NBh ad 1.1.7, śabda as an instrument of knowledge is defined as follows: śabda is the instruction (upadeśa) of a trustworthy instructor. It is of two kinds, since it can have either a perceivable or an imperceivable artha.
āptopadeśaḥ śabdaḥ // sa dvividho dṛṣṭādṛṣṭārthatvāt // The trustworthiness of the source concerns language in general, so that it is considered the ground for the validity of both common and Vedic language. Unlike in Mīmāṁ sā, in Nyāya the authority of the Veda is also founded on the trustworthiness of their source: And that [Veda] is a valid source of knowledge because its trustworthy instructor is a valid source.
[…] tatprāmāṇyam āptaprāmāṇyāt // A trustworthy instructor is defined as follows: Trustworthy is the instructor (upadeṣṭṛ) who has directly experienced the essential property (dharma) [of things], and is moved by the desire to describe [things] as they are or they are not. āptaḥ khalu sākṣātkṛtadharmā yathādṛṣṭasyādṛṣṭasya cikhyāpayiṣayā prayukta upadeṣṭā / Being grounded on the utterance of a trustworthy instructor, śabda can by no means be considered permanent (nitya) by the Naiyāyikas, since such an utterance must have occurred at some point in time. Moreover, the relation between śabda and artha is considered conventional, rather than natural, because it is based on stipulations (saṅketa) established by divine or human beings (NBh ad 2.1.55-56).

Differences in Mīmāṁ sā and Vyākaraṅa
In Mīmāṁ sā, śabda is accepted ipso facto as natural and unalterable (Mīmāṃsāsūtra 1.1.5, see Ś Bh, p. 28), so issues related to its development or aetiology are deliberately ignored as pointless. The possibility of an author is irrelevant for its 4 The Nyāya and Mīmāṁ sā traditions are counted by Julie Jack (1994) among exponents of what she labels "Uniqueness School", according to which there is a "uniqueness of knowing from words as a form of knowledge,-its irreducibility either to perception or to inference […]" (1994, p. 165). In her essay she also explores the problematic overlap of comprehension, judgment, and rational belief in verbal testimony.
Textual Re-use in the Nyāyamañjarī 141 validity as an instrument of knowledge (which in Mīmāṁ sā is, specifically, knowledge of dharma) and for its reality, i.e. its relation with a (mental or concrete) object. The fixedness (nityatā) of such a relation necessarily implies the permanence of śabda: the idea that śabda is ephemeral, common to most opponents of Mīmāṁ sā, is rejected in Mīmāṃsāsūtra 1.1.18: [śabda] must be permanent (nitya) [in relation to its artha], because [its] appearance is for the purpose of [communicating with] someone else. nityas tu syād darśanasya parārthatvāt // If śabda were ephemeral, how could there be a continuous and consistent intersubjectivity of speakers and hearers, who comprehend a same expression in the same way? To account for the sheer phenomenon of communication, the relation of a śabda with an artha must be understood as stable and permanent.
In Bhartṙhari's philosophy of grammar, the ultimate reality (paramārtha) is a non-dual śabdatattva, while all dualities and differences, including the distinction into śabda and artha, appear as such only on the relative level of worldly interactions (vyavahāra), and not in an absolute, real sense (see VP, 1.1).
Moreover, since the ultimate reality is śabda, true knowledge must be grounded in it (VP, 1.131): In this world there is no cognition without the comprehension of śabda. Every knowledge shines as if penetrated by śabda.
na so 'sti pratyayo loke yaḥ śabdānugamād ṛte / anuviddham iva jñānaṃ sarvaṃ śabdena bhāsate // Jayanta sums up Bhartṙhari's philosophy of śabda as follows (NM Va II, pp. 156,19-157,1): The non-dual Absolute, śabda, whose apparent differences are caused by traces of a beginning-less nescience, is deceptively perceived as if related to the artha. There is actually no signified thing separated from the signifier. Therefore, this signified-signifier (vācyavācaka) subdivision, which is merely theoretical, is nothing but nescience used as a tool to achieve awareness (vidyā).
The non-dual reality is śabda, so in Bhartṙhari's perspective Nyāya concepts such as "trustworthy instructor" and "conventional meaning" are fictional ones. Also the śabda-artha dichotomy is imaginary, since such duality does not exist in reality. The segmentation of language into sentences, words, and phonemes is an artificial operation, certainly useful for didactic purposes, but ultimately unreal.

The Oral and Written śabda
Traditionally, in India, śabda typically pertains to the realm of orality, while present analyses of textual re-use mostly concern written literature. 5 It is debatable to what extent Jayanta used written sources or, rather, had them committed to memory when he wrote the NM.
We also do not know how works such as the NS or the NBh were originally composed and, if orally composed, when they began to be preserved and transmitted in written form. However, judging from the epistemic importance of śabda and from teaching and learning habits in traditional circles witnessed in modern times, it is possible that the performative tradition of these works was mainly oral, and that they were often memorized and taught without much reliance on writing. 6 For easily memorizable works, such as those in aphorisms and verses, the written record was probably perceived as secondary. In other words, the authority of orally taught truths, passed on from teacher to pupil, was probably superior to that of written ones.
As for large and digressive works such as the NM, they were more likely composed in written form to begin with and were not commonly meant to be memorized in their entirety. If this is the case, then, also the purpose of their written transmission must have been different. In 1472 CE, Ś itikaṅt˙hācārya Svāmin, the learned copyist of the oldest extant NM manuscript, wrote in his colophon (P, fol. II 271r,1) that the NM had been copied by him for teaching purposes (śiṣyān adhyāpayitum).
In any case, what is the epistemic role of written words for Jayanta? During his refutation of the sphoṭa, he makes clear that written letters convey knowledge of the artha indirectly, through an inferential process: […] Therefore the knowledge of the artha caused by the ink traits is based on the inference of the phonemes (tasmād varṇānumānapurassaraiva rekhābhyo 'rthāvagatiḥ).
It thus seems theoretically acceptable to consider writing as leading to an instance of epistemically effective śabda. Just like the perception of phonemes assists the hearer in knowing from śabda, so can the perception of the ink traits assist an inference of those very phonemes, which again assists the hearer in knowing from śabda. One should also keep in mind that in Nyāya epistemology the possibility of knowing one and the same object through different instruments of knowledge (pramāṇasamplava) is acceptable (NM Va I, pp. 87-93). Furthermore, loud reading of a written source, either by the teacher to the pupils or on one's own, may also play a role in these considerations.
Before Jayanta, in Nyāya sources, there had been no focus on the sentence, with linguistic analyses mostly concerned with words and their link to external things. In NM 5 (NM Va II,pp. 135,10), after examining various theories on sentence Footnote 5 continued what is written survives the act of writing. So all written words exist side by side […]". Mohanty (1994, p. 31) suggests slight modifications in the utterer-conditions to make room for written testimony as śabda. 6 This, incidentally, would also explain the absence of a manuscript tradition of the NS independent of the NBh in the early stage of the transmission.  objection:] This is a good point. That effort, however, was done by the author of the [Nyāya]sūtra in order to establish the epistemic validity of śabda, and to calm down protests that there is no contact between śabda and artha.
[Objection:] Yet, if this is the case, without an external object as the artha of the sentence the epistemic foundation of the science would remain shaky, so an effort should be done also in this area. [Counter-objection:] True. The author of the [Nyāya]sūtra, by teaching only the artha of the word, meant to take care [also] of that [artha of the sentence], so he did not teach the artha of the sentence separately from that of the word. Therefore, his idea is that the very artha of the word is the artha of the sentence; […] not that the artha of a single word is the artha of the sentence; rather, the artha of several words is the artha of the sentence.
In this passage, Jayanta stressed the epistemic focus of the Nyāya discipline. Also, he apparently endorsed a realist epistemology in which the artha is external and real (bāhya and vāstava). Lastly, he hinted at the strategy that he was going to adopt to explain the transaction from word to sentence signification, on the one hand respecting the letter of the NS (2.2.66), were the artha of the word is defined, and on the other integrating a modified version of the Bhāṫt˙a Mīmāṁ sā theory of sentence signification in the Nyāya system. 7

Jayanta on Re-use and Originality
In the introductory verses of the NM (NM Va I, pp. 2-4), there are further indications about Jayanta's relation with his own tradition, a source of many of his re-uses. In verse 4 he traced back the Nyāya tradition to sage Akṡapāda, to whom the NS is attributed, and in verses 7-9 he portrayed the tradition of Nyāya as the source of his tenets, claiming for himself only the merit of having craftily re-ordered pre-existing notions: 8 This superior essence was collected in the forests of the herbs of Nyāya. It was extracted like butter from the milk of reasoning (ānvīkṣikī). 9 How could I be capable of even envisioning a new topic? Here my only concern is a variety (vaicitrya) in arrangement of the [traditional] statements. Flower chaplets crafted in the past can generate new interest if their very flowers are recomposed on a new string.
After comparing the Nyāya system to a mighty tree cared for by Akṡapāda, Jayanta depicted himself as merely capable of a partial view of its full richness: In fact, the tall tree of logic laid down by Akṡapāda bears a load of fruits oozing thick nectar. I will gently shake it, being uncapable of climbing on it.
[Thus] I cannot even see its full burden of riches.
To sum up, Jayanta planned his NM as a new structure built with tenets of a rich pre-existing tradition, without any claim for originality, except for the arrangement of the presentation.

The Argumentative Structure of the sphoṭa Section of NM 6
To examine the formal aspect of quotations and other re-uses, and to assess their epistemic value, the context of the argumentative structure is crucial. The sphoṭa section of NM 6 can be segmented in five argumentative steps: problem, thesis, first antithesis, second antithesis, synthesis (sketched in Fig. 1).
• The topic is introduced by the philosophical problem (saṃśaya) that is going to be dissected and settled: in the epistemic process of knowing something upon hearing a linguistic expression, what is precisely the cause of knowledge of an artha? Is it the ephemeral phonemes, as claimed in Nyāya (pratijñā, thesis)? Or is it the indivisible entity called sphoṭa, as claimed in Vyākaraṅa (pūrvapakṣa, first antithesis)? Or is it the permanent phonemes, as claimed in Mīmāṁ sā (uttarapakṣa, second antithesis)?
• The thesis is that phonemes are heard in sequence form words and sentences, and thus they collectively cause knowledge of an object, though phonemes are not permanent entities, but rather ephemeral ones.

•
The first antithesis (pūrvapakṣa) is the solution of the problem as proposed in Vyākaraṅa, based on Bhartṙhari's holistic theory of meaning.

•
The second antithesis (uttarapakṣa), which refutes the first one, is the solution of the problem as proposed in Mīmāṁ sā, largely based on Kumārila's atomistic arguments. 10 • The synthesis (siddhānta) consists of a minor modification of the Mīmāṁ sā arguments, to suit the Nyāya needs. 9 On the evolution from Ā nvīkṡikī to Nyāya, see Preisendanz (2009). 10 "Atomism" and "holism" are here used as relative terms: Kumārila has an atomistic perspective in relation to Bhartṙhari, but Nyāya views before Jayanta could be considered more atomist than Kumārila's. On the use of the terms "holism" and "atomism" in relation to the context and composition principles and to the Indian theories of meaning, see Matilal and Sen (1988, p. 84).
The thesis and the two antitheses are in turn structured in a dialogic form of progressive objections (pakṣa), counter-objections (pratipakṣa), and verdicts (nirṇaya), so that in the first antithesis the objections and the verdicts are Bhartṙhari's, while the counter-objections are (mainly) Kumārila's; in the second antithesis the roles are reversed. The problem, the thesis, and the synthesis are presented very concisely, while the two antitheses form the bulk of the section.

Re-uses in the NM
In the following, a selection of re-uses in defense and refutation of the sphoṭa are documented in tabular form. For pragmatic reasons, I limited myself to re-uses of the Śābarabhāṣya (Ś Bh), Ślokavārttika (Ś V), Ślokavārttikatātparyaṭīkā (ŚVTṪ ), 11 and Vākyapadīya (VP). I ignored other sources such as the Bṛhatī (Bṙh) and the The debt of Jayanta to Uṁ veka's commentary is clear from many passages (see Sects. 7.1, 7.3, 7.6, 8.1, and 8.2 below). In a number of occasions (GBh,pp. 24,14,149,4,182,17), Jayanta's commentator, Cakradhara, also noticed the connection and explicitly mentioned Jayanta's references to Uṁ veka.
Sphoṭasiddhi (SphS). 12 The list of re-uses presented here is a selection without claims of exhaustiveness, as there are less evident re-uses which were discarded, and other possible ones that may have escaped my attention.
Re-uses are sorted according to their sequence of occurrence in the NM and are grouped in tables on the basis of the argumentative structure of the NM. Each table is contextualized by a brief synopsis of the philosophical issues at stake, and followed by some comments on the interesting features of the re-uses.
The first column contains the origin of the text re-used by Jayanta, with the reading of the most reliable edition available to me (NS, NBh, Ś V Ra , Ś VTṪ , VP; for quotations from parts of the Ś V not covered by Ś V Ra , I used Ś V Dvā instead); manuscript research on these sources, although important for a conclusive assessment, was not feasible at this stage. The second indicates the truth value (True/False) from the source viewpoint. The third, fourth and fifth columns contain the re-use as found in the NM, subclassified into three segments: the re-use (sequentially numbered for cross-reference purpose in this paper) preceded or followed by pre-and post-quote markers, when present; re-used expressions, including also differently inflected stems, are shown in bold face. Lastly, the sixth and the seventh columns indicate the truth value from the point of view of the symbolic speaker, and the tradition symbolized by the statement (Vyākaraṅa/Mīmāṁ sā/Nyāya).
All the NM passages reproduced in this paper refer to the pagination of NM Va . Whenever other relevant witnesses 13 -namely NM Ga , the Nyāyamañjarīgranthibhaṅga (GBh) and the important manuscripts (P and C)-have substantive variants that differ from NM Va , such readings are either received in the text or shown in critical notes below the tables. Vaiyākaraṅas maintain that phonemes cannot convey meaning, neither independently nor collectively. Independent phonemes are just meaningless parts of words. And if a word is considered a collection of phonemes, one must explain if they function simultaneously or in sequence, and both options are not tenable.
The simultaneous existence of phonemes is ruled out in (1). The only possibility of a simultaneous utterance of the phonemes composing a word would be that each 12 Although in the NM there are no explicit references to Maṅḋana or to the SphS, there are some instances that suggest some relation. The most striking one is the argument in NM Va , 150,13-14, where eke, contrasted to apare, i.e. to Bhartṙhari, may refer to Maṅḋana Miśra. The SphS, 89,4-6 reads: yathā ratnaparīkṣiṇaḥ parīkṣamāṇasya prathamasamadhigamānupākhyātam anupākhyeyarūpapratyayopāhitasaṃskārarūpāhitaviśeṣāyāṃ buddhau krameṇa carame cetasi cakāsti ratnatattvam. The NM reads: yathā ratnaparīkṣakāṇāṃ prathamadarśane ratnarūpam amalam prakāśamānam api punaḥ punaḥ parīkṣamāṇānāṃ carame cetasi cakāsti niravadyaṃ ratnatattvam. The terminological similarity is noteworthy. I could not find this jem/jeweller comparison in early sources, except for Vācaspati Miśra's Tattvabindu (see TB, p. 70), for which there are arguments in favor of its posteriority in respect to the NM. 13 A detailed explanation of the criteria for selecting relevant NM witnesses is provided in Graheli 2012b.
Textual Re-use in the Nyāyamañjarī 147 of a group of speakers simultaneously utters one of the phonemes composing the word, which is absurd (2). So, since a plurality of phonemes is uttered by a single speaker, there must necessarily be a chronological sequence of phonemes in such an utterance (3) ( Table 7.1).
In (1), the link with Kumārila's statement is more evident in the variant of P, with the reading sattā, in place of satām of the NM editions. For both opponents, as well as for Jayanta, (1) denies the possibility of the simultaneous existence of the phonemes forming a word. Although because of different reasons, for all of them it is true that by itself a group of phonemes can not cause verbal knowledge. For Mīmāṁ sakas śabda is permanent and thus an ontological simultaneity of phonemes is acceptable, but phonemes need to be uttered and grasped in order to be effective, and they are uttered and grasped in sequence. Also for Vaiyākaraṅas śabda is permanent, but an ontological simultaneity of phonemes is impossible because phonemes are ultimately fictional entities. For Naiyāyikas a simultaneous existence of phonemes is simply not possible, due to the ephemeral nature of śabda.
In (2) the absurdity of many people simultaneously pronouncing different phonemes of a same word is stated, and this is obviously shared by all. (3) is also true for everyone involved here, because if a single person utters a series of phonemes there must necessarily be a sequence.

The Last Phoneme does not Cause Knowledge
Conceding that phonemes do not convey meaning, neither independently nor collectively, the Mīmāṁ sakas argue that the perception of the last phoneme triggers verbal knowledge, while assisted by the memory of previous phonemes (4). This, however, is unacceptable, because such memory would be caused by a mental disposition (saṃskāra), which would be in turn produced by the perception of each phoneme. But memory and mental dispositions are not accepted as instruments of valid knowledge, so this theory is not sound (Table 7.2). The tenet expressed here is true for Mīmāṁ sakas, but false for Vaiyākaraṅas.

A Cognition Made of Memory and Perception is Impossible
The passage quoted in (4) has been interpreted by Kumārila by postulating a single, variegated cognition embracing the memory of the past phonemes and the perception of the last one, to avoid the issue raised by the Vaiyākaraṅas, namely that memory or mental dispositions of past phonemes cannot cause valid knowledge. But, argue the Vaiyākaraṅas, such a variegated and unitary cognition is also impossible, because mental dispositions cannot possibly produce a single cognition together with sense organs: mental dispositions cause recollections, while sense organs cause perceptions (Table 7.3).  In (5) the two concepts of a variegated (citrarūpā) and cumulative (saṅkalanātmikā) cognition used by Kumārila and Uṁ veka as two alternative scenarios in the theory of phonemes, seem to be reduced to a single one by Jayanta. The notion, in any case, is false from the Vaiyākaraṅas's point of view.

Language Acquisition does not Justify the Existence of Phonemes
In defense of the theory of phonemes, Mīmāṁ sakas claim that language-produced knowledge requires linguistic competence, which is acquired through the repeated observation of a given pattern of phonemes in relation to their artha (6-7) ( Table 7.4).
From the Vaiyākaraṅas' point of view it is false that language acquisition occurs through phonemes.

The Theory of Phonemes is not Supported by Common Usage
People commonly say that "from śabda we know an artha", where the use of the singular, "from śabda", matches the unity of the sphoṭa, not a plurality of phonemes (Table 7.5).
From the Vaiyākaraṅa's point of view it is correct that common usage favors the sphoṭa, so (8) is true. The same argument was present in the source in the form of an objection, and thus considered false there. To argue that śabda must by definition be audible, and that phonemes are audible while the sphoṭa is not, does not make sense, because audibility is not an exclusive character of phonemes (9) ( Table 7.6). The Mīmāṁ saka may argue that although it is not an exclusive characteristic, it is still the main one, and that the phonemes are audible while the sphoṭa is not. Yet, the characteristic mark of śabda is not the mere audibility, but rather, its capacity to cause knowledge of the artha (10-11). And such character belongs to the sphoṭa, not to phonemes.
People are erroneously led to think that phonemes are the cause of verbal knowledge only because the sphoṭa manifests through articulated language, but actually this idea is the result of a false-cause fallacy (12) ( Table 7.6).
It is false, from the Vaiyākaraṅa's viewpoint, that the audibility of phonemes qualifies them as śabda (9).
(10) and (11) are false for the Mīmāṁ sakas and true for the Vaiyākaraṅas, who have the interest of underscoring causality of cognition over audibility. (12), which is true for the Mīmāṁ sakas, is rejected by Vaiyākaraṅas as false.

The Theory of the sphoṭa is not Anti-economic
The Mīmāṁ sakas object that if the sphoṭa is manifested by phonemes, as maintained by some Vaiyākaraṅas (see above, footnote 12 on the SphS), a criticism moved against the theory of phonemes stands also against the theory of the sphoṭa, which basically adopts the theory of phonemes with its implications and on top of it postulates a further entity (13-14) ( Table 7.7).
The economy of the theory of phonemes in (13) and (14) is true for the Mīmāṁ sakas, but false for the Vaiyākaraṅas.

The sphoṭa is Manifested by Articulated Sound
According to other Vaiyākaraṅas, however, this is not true: the sphoṭa is not manifested by phonemes, but rather by phonetic sounds (dhvani), i.e., the combined result of breath, articulation, etc., which in any case do not have ultimate ontological status (15) ( Table 7.8).
From here to the end of the first antithesis, all re-uses are of Vaiyākaraṅa texts, and thus true also in the sources.

The Sentence is the Linguistic Unity
Not only are phonemes unreal, even words are fictional abstractions (16). The sentence is not a whole made of parts; rather, it belongs to a class of its own (17). One may argue that there is a one-to-one correspondence between each phonemic string and its respective artha, but this is not true; for instance (18), in unrelated words like kūpa, yūpa, and sūpa, we see that some phonemes are identical, yet there is no similarity in meaning (Table 7.9).

Words are Fictional Abstractions
The signification of words is thus a fictional device, like that of roots and suffixes. Words are an abstraction, useful for didactic purposes, etc., but devoid of ontological status (19-20) (Table 7.10).  In (20), remarkably, while in C and GBh only the second hemistich of VP 3.1.1 is present, in P and in the vulgata also the first hemistich is present. C and GBh, when reading together, are evidence of an earlier stage of the NM transmission.  The absolute reality is an indivisible śabda, while fictional differences are nothing more than instruments to move towards an awareness of the śabda unity, or didactic means to describe language (21-23) ( Table 7.11).

The Threefold Manifestation of śabda
Even if in reality it is one and indivisible, śabda manifests in the world of phenomena in three aspects, vaikhārī, madhyamā, and paśyantī (24-27) ( Table 7.12).

Second Antithesis: Re-use in the Refutation of the sphoṭa
With the exception of (40) below, in this section both re-uses and sources are arguments presented from the Mīmāṁ saka viewpoint, so they express true concepts both in the source and in the re-use.

Sequential Phonemes Can Cause a Cumulative Cognition
In response to the Vaiyākaraṅas' objections (see Table 7.1), the Mīmāṁ sakas argue that the sequential utterance of phonemes is not an issue: although phonemes are uttered and grasped in sequence, they still are, collectively, the cause of verbal knowledge. Instances of collective and sequential causes that bring about a cumulative effect are well known, as in the case of drills to memorize verses (28) or of intermediate sacrifices (29)(30) in the economy of the main one (Table 8.1).

Dispositions and Memory in the Theory of Phonemes
The principle that the perception of the last phoneme triggers verbal knowledge, aided by the memory of the previous phonemes (31), which was quoted and refuted by the Vaiyākaraṅas (see Table 7.2), actually stands valid. Objections on the capacity of dispositions to produce verbal knowledge do not hold, if we understand "mental disposition" as vāsanā, a quality of the self. From observation we know that perceptions cause dispositions, and observation is a universal instrument of knowledge (33). And it would be silly to ask "From where does such a mental disposition arise?", because everyone knows that mental dispositions are caused by perception (34) ( Table 8.2).
(31) was also quoted above (see Table 7.2). In that occasion the line in favor of atomism was not credited, unlike here. It could be because of the obviousness of the source, but it is quite likely that the absence of credits had rhetoric implications: the honorific tatrabhāvatā stresses the authoritativeness of the statement, while in the previous case, where the argument is developed from the Vaiyākaraṅas' perspective, the authoritativeness of the source was meant to be undermined. In (32) Jayanta gives a Vaiśeṡika twist to the word saṃskāra used by Ś abara, interpreting it as the quality of the self called vāsanā. The latter term is also used by Uṁ veka in the immediate context.

The Theory of Phonemes is More Economical
Alternatively, the Mīmāṁ sakas may concede that mental dispositions do not cause knowledge of the artha directly. Yet, they can do it through memory: mental dispositions of the past individual phonemes in sequence, together with the perception of the last one, cause a new mental disposition that produces the memory of the full word, which in turn causes the knowledge of the artha. When the Vaiyākaraṅas argue that this theory is anti-economical, the Mīmāṁ sakas, with Ś abara, reply that they are postulating a single entity (the mental disposition), while the Vaiyākaraṅas necessarily must postulate two (the sphoṭa and the mental dispositions, through which the sphoṭa manifests) (35).
When the Vaiyākaraṅas claim that, at least, they did not violate the law that mental dispositions cause only memory, the Mīmāṁ sakas reply that they did indeed violate it, because when they deal with the cause of the manifestation of the sphoṭa (see Table 7.8), they have to explain it in a similar way. In addition, they postulate the sphoṭa (36).
And the explanation of the gradual manifestation of the sphoṭa does not hold, because knowledge of an artha is not a gradual, but rather a sudden phenomenon (37) ( Table 8.3).

The Sequence is a Property of Phonemes
The sequence is just a property of phonemes, so no entity needs to be postulated in addition to the phonemes (38) ( Table 8.4).

The sphoṭa does not Qualify as śabda
Phonemes are the only basic linguistic elements, other fictional entities are not required (39); śabda is said to be of two types, articulated and not articulated, and the sphoṭa is neither of these two (40) ( Table 8.5).
(40) is one of the rare cases, in NM 6, in which Jayanta quotes from his own tradition.  While in short words such as "cow" there is an impression of unity due to the brevity of the sound, there are many long words, such as "Devadatta" where the differences of sounds composing the word are clearly audible (41). The Vaiyākaraṅas proposed the reductio ad absurdum that if sentences are made of words and words are made of phonemes, also phonemes must have parts. But parts of phonemes are never perceived: a phoneme is either fully perceived, or it is not perceived at all (42) (Table 8.6).   9 Re-use from Nyāya Sources In Tables 9.1-3 there is a small sample of re-uses from the NS and the NBh, as a term of comparison with the above-listed sources from other traditions. In Table 9.1, three quotations of the NS are shown, as an example of the many occurring in the NM. In Table 9.2, there are some re-uses from the NBh without an explicit mention of the NBh author. Finally, in Table 9.3, a few NBh re-uses with an explicit attribution are listed.

Conclusions
In Tables 10.1, 10.2 and 10.3 some characteristics of all the above re-uses are summarized for a better appreciation of specific patterns. The sequential number is shown in bold typeface when the re-use is a quotation. The context shows how re-uses constitute the back-bone of the whole argumentation in favor and against the sphoṭa. It would be possible, indeed, to make sense of the main structure of the sphoṭa section just by following the concatenation of arguments present in the quotations. The sequence of the arguments, however, differs from that found in previous sources, so Jayanta's claim of his role as a re-arranger seems corroborated by this survey.
In the re-uses surveyed here there are always three subjects involved: besides Jayanta himself (the Nyāya exponent), in fact, a dialog is staged between a symbolic re-user and the utterer of a re-used source, in the present case alternatively the Vaiyākaraṅa or the Mīmāṁ saka. The symbolic re-user is in some cases explicitly

True
False True mentioned, and in others clearly identifiable in the flow of the argument, according to his role in the debate either as proponent (pakṣin) or opponent (pratipakṣin). Although the Mīmāṁ saka role in the staged debate is that of presenting a second antithesis (uttarapakṣa) and not a final verdict (siddhānta) one cannot fail to notice how Jayanta's views are very close to the Mīmāṁ saka's, so much so that in the present list of re-uses, insofar as the truth value of the statements, Jayanta almost invariably agrees with the Mīmāṁ saka. In these conclusive tables, the truth values seen from Jayanta's viewpoint are in bold face, together with the truth values of those who agree with him.  Throughout the NM, when not explicitly credited to the author, quotations are almost invariably introduced by tad uktam, or, more rarely, yathoktam. The overwhelming percentage of quotations thus introduced is in verse or sūtra form. This tad uktam pre-quote expression can be considered Jayanta's marker for a precise quotation, so much so that it may be used as a valuable clue to identify quotations from undetermined or lost works. Also, it is a signal that the specific tradition derived from the context, in the above cases Vyākaraṅa and Mīmāṁ sā, is symbolized in the quotation. Unlike Ś abara and Kumārila, Bhartṙhari is never explicitly credited. Since Jayanta's views are certainly closer to the Mīmāṁ sakas's, such explicit credits could by themselves be not only tokens of respect, but also of the trustworthiness of the source.
The almost invariable concomitance of the tad uktam formula with quoted verses and aphorisms may have several reasons. It is possible that verses were more suited for quotation purposes, for mnemonic reasons, and that therefore Jayanta thinks more appropriate to introduce them with an explicit "so it was said". It is also possible that verses and aphorisms, since they were more easily memorized and thus faithfully transmitted, enjoyed a special status as śabdapramāṇa in Jayanta's perspective.
It is remarkable how differently re-uses from Nyāya sources are treated. First of all, in these cases the tad uktam pre-quote is not systematically present anymore. There is a substantial amount of uncredited quotations from the NBh, which is all in prose, except for rare passages (see Table 10.3). Perhaps these passages were so well known to Jayanta and his audience to make irrelevant the necessity of credits, or perhaps in Jayanta's understanding only versified passages were worthy of credits, but in some cases they could also be unconscious re-uses creeping in the NM. 14 In the context of versification, I found (48) particularly interesting, where a prose passage of the NBh is integrated and completed in metrical form by Jayanta.
In relation to re-use in Nyāya literature, in any case, it seems that the impact of the oral tradition, and thus of massive portions of works committed to memory, must be taken into account.
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.