On Not Understanding Extraordinary Language in the Buddhist Tantra of Japan

Abstract

The question motivating this essay is how tantric Buddhist practitioners in Japan understood language such as to believe that mantra, dhāraṇī, and related forms are efficacious. “Extraordinary language” is introduced as a cover term for these several similar language uses found in tantric Buddhist practices in Japan. The essay proceeds to a critical examination of Anglo-American philosophy of language to determine whether the concepts, categories, and concerns of that field can contribute to the analysis and understanding of extraordinary language. However, that philosophy of language does not contribute to this analysis, as it is constrained by its continuing focus on its founding concepts, dating particularly from the work of Frege. Comparing it to Indic thought regarding language reveals a distinct mismatch, further indicating the limiting character of the philosophy of language. The analysis then turns to examine two other explanations of tantric language use found in religious studies literature: magical language and performative language. These also, however, prove to be unhelpful. While the essay is primarily critical, one candidate for future constructive study is historical pragmatics, as suggested by Ronald Davidson. The central place of extraordinary language indicates that Indic reflections on the nature of language informed tantric Buddhist practice in Japan and are not simply cultural baggage.

1. Introduction

Visualizing the syllable A (ajikan, 阿字観) is the second of two practices preliminary to starting the fourfold training (shido kegyō, 四度加行) of a Shingon priest (acarya, ajari, 阿闍梨). A variant form of this practice is also taught as a stand alone, or single practice, particularly for lay adherents to the Shingon tradition. The practice involves gazing at a white circle on a black background in which is inscribed the syllable A in Siddham script, developing an eidetic image of it, and then manipulating the image in various ways.

The syllable A is a seed syllable (bīja) mantra and is explained as having a threefold significance: it is the origin of all being, present in all being, and the destruction of all being. To someone familiar with Hindu conceptions, that triad of meanings would be reminiscent of the relations between the three deities Brahma, Viṣṇu, and Śiva (trimūrti), grouped together as creator, sustainer, and destroyer, respectively. Focusing more specifically on A as a syllable, however, tantric symbolism links it to these three functions through its role in Sanskrit. According to the theory of phonematic emanation (Padoux 1990), A is the first syllable of the phonemic sequence of syllables by which this cosmos is manifested and is therefore the origin of all being. A is also the vowel present in each of the syllables manifested, other vowels being understood as variants on A. Since the sequence of syllables is the sequence of manifestation, A is present in all being. And, as the negative prefix, A is considered to bring an end to all being. This visualization practice employing the Siddham script syllable A is one instance of “extraordinary language” in Japanese tantric Buddhism.

Having introduced the category “extraordinary language,” Section 2 explains how that phrase is being used and why. Section 3 provides the reader with examples of extraordinary language in the tantric traditions and the tantric penumbra in Japan. Section 4 explicates the question motivating this inquiry, my thesis regarding the answer to that question, and the significance that answering the question has. Section 5 then considers whether the modern academic philosophy of language has any tools (concepts, categories, concerns) that can contribute to our understanding extraordinary language and its use in tantric rituals in Japan. Section 6 will examine some of the ways that extraordinary language has been explained in religious studies. The primary focus here is critical, in preparation for more adequate studies in the future, ones that appreciate the important role of Indic understandings of language in the development of tantric Buddhist praxes in East Asia.

2. Why “Extraordinary Language”?

The use of language forms such as mantra, as for example in ritualized practices, diverges from ordinary forms of language use, which a majority of those working either in the philosophy of language or in linguistics characterize as communicating information. This, however, proves to be problematic when attempting to understand the ways that mantra and similar uses of language are employed in tantric traditions. Since these language forms do not function in the ordinary, communicative fashion, I will refer to them as “extraordinary language” in the literal sense of “outside the ordinary.” The idea that language is identical with communication is one of the fundamental constraints on the philosophy of language, which takes the discursive or propositional form as basic. The effects of this constraint will be examined more fully below.

An important constraint on an interpretive project such as this one is that we should not presume that the function of extraordinary language is uniform across all usages. We are, after all, examining phenomena that stretch over perhaps as much as five millennia, from Vedic to contemporary times. Even limiting our consideration to tantric use means that we are attempting to understand across a millennium and a half and at least five linguistic and religious cultures. Instead of assuming uniformity, the assumption should be just the opposite—usages should be considered in their individual context, though at the same time within a framework that allows for systematic comparisons. Further, the goal of a single, comprehensive explanatory theory of mantra and other related forms is a chimera—specific context and use are determinative, and those are stunningly variant. Nor can we presume a single lineal developmental history, such as might be constructed as running from Vedic, to tantric, to East Asian Buddhist tantra. Our attention here is on extraordinary language in the tantric traditions and tantric penumbra of Japan, which although still a rather broad swath does provide a largely coherent historical and cultural context.

3. Extraordinary Language in Japanese Buddhist Tantra and the Tantric Penumbra

The Shingon and Tendai traditions of Japan inherit the variety of language forms found in Indic tantra, employing mantra (shingon, 真言), dhāraṇī (darani, 陀羅尼), and vidya (myōshu, 明呪) in ritual contexts, as well as in written forms—the latter usually in Siddham script. Such practices have spread well beyond the specifically tantric traditions into the tantric penumbra in Japan. Such uses are found on the continent (Copp 2011), which is the proximate source for usages in Japan. As described above, visualization of the bīja mantra A has been promoted among Shingon adherents as a standalone practice. Other mantra and dhāraṇī practices were also promoted by Shingon teachers. For example, the Clear Light mantra (kōmyō shingon, 光明眞言) was promoted by Myōe Kōben (明恵, 1173–1232) in medieval Japan and continues to be taught today (Unno 2004). Since both mantra and dhāraṇī are discussed in this essay as instances of extraordinary language, distinctions between the two that might be appropriate for other kinds of inquiries will not be emphasized.

Similarly, dhāraṇī are used in rituals and ceremonies of the Zen tradition, such as in the ritual of “invoking the sage” (shukushin, 祝聖; (Mohr 2008, pp. 207, 215); regarding a different pronunciation and meaning of these characters see (Mohr 2008, p. 323, n. 1)). Also, many sects, including Zen, use funerary markers that are both shaped like stupas representing the five elements of Indian cosmology (pañcacakra, gorinsotoba, 五輪率都婆) with inscriptions in Siddham script of the five seed (bīja) mantras for the five elements. Across Buddhist Asia, a variety of other such uses of the written form are found, such as on amulets, stele, and prayer wheels (Copp 2014). Found in the Pure Land traditions of Japan, the practice known as nenbutsu (念仏, buddhānusmṛti) is a recitation of the name of Amitābha Buddha (阿弥陀, Amida) in the form of the expression NAMO AMIDA BUTSU. Within the Shin tradition (眞宗, Shin-shū), however, this phrase is distinguished from a mantra and explained as a spontaneous expression of gratitude for Amida’s assurance that each will be born in Sukhāvatī after death. Despite this sectarian delineation, it is functionally and structurally similar to mantra and apparently originated as a mantra, with the sectarian interpretation a later addition. Likewise, in the Nichiren sects, the name of the Lotus sūtra is recited in the form of the phrase NAMU MYOHO RENGE KYŌ—a practice known as daimoku (題目) and an instance of the cult of the book (Schopen 1975). Both of these practices—reciting the name of a Buddha and the name of a text—have Indic analogues. Like Shin-shū, Nichiren-shū is part of the tantric penumbra in Japanese religious culture. This is also evidenced by the object of cult veneration (gohonzon, 御本尊, also mandara gohonzon, 曼荼羅御本尊) in Nichiren-shū, which is a maṇḍala-like arrangement of written forms, combining Chinese characters and siddham script mantra that originates with the founder, Nichiren (Dolce 1995, 2002).

The history of the transmission of tantric Buddhism to Japan crosses at least two cultural boundaries—that between India and China, and then between China and Japan. In both instances, the transmission of practices employing extraordinary language faced a twofold set of obstacles. The first is the differences between the religious cultures of South Asia and East Asia. The second is the problems in both transliteration and translation from Sanskrit, through Chinese, to Japanese (Abé 2005, p. 292). Responses to the latter include developments in the technical lexicon and grammar and in orthography, which for the written forms of amulets, talismans, and funerary markers suggests an area of future focused research extending to Japan (for example, Ambros 2001) that has already been done on these forms in China (for example, Robson 2008, Copp 2014). As evidenced by these selected examples, the tantric traditions of Japan provide a particularly rich source for the study of the use of extraordinary language.

4. Question, Thesis, and Significance

The question motivating the larger project of which this is a part (Payne 2018) originates from attempting to understand the extent to which extraordinary language became a common element found in the explicitly tantric Buddhist traditions in Japan, in other traditions, and in the popular religious culture of Japan. These latter can be described as the penumbra of Esoteric Buddhism—metaphorically, areas in which two sources of light interact.

The question of the overall project is “How did Esoteric Buddhist practitioners in Japan conceive of language, such that extraordinary language forms originating in Sanskrit were considered effective?” My thesis is that this understanding of language as efficacious was not simply based on a claim that Sanskrit was conceived to be the sacred language (Williams 2008). That has not been a part of Buddhist thought since its beginnings, given the well-known injunction by the Buddha that the dharma be taught in the language of the audience population. Without a traditional commitment to a simplistic conception of Sanskrit as an inherently sacred language, the transmission of practices involving extraordinary language alone would not have been sufficient, and Indian understandings of language needed to be transmitted in conjunction with the practices. The corollary of this is that understandings of language were not incidental baggage carried from India through China to Japan but were an integral part of Esoteric praxis.

The significance of this thesis is that it challenges two conceptions of Buddhism fairly widely shared in religious studies and popular religious culture. First, the understanding of Buddhism as a religion, when religion is understood in such modern Euro-American terms as essentially the individual quest for transcendence. Instead, it is more appropriate to understand that during important periods of its history, Buddhism has constituted a civilization, i.e., not a cultural form, something located within a society, but as something much more comprehensive. At other times, of course, it has not had the same definitive role and has instead constituted one of several different, not infrequently antagonistic, institutions within a society. In contrast to both of these, the conception of Buddhism in terms of an individual quest for transcendence has marginalized the importance of understandings of language and other aspects of Buddhism as a culture, such as medicine (Gyatso 2015) and embryology (Andreeva and Steavu 2016), astronomy and astrology (Kotyk 2017), and other cultural technologies. Second, this challenges the rather widely shared understanding that practice and doctrine, or theory and method, exist in isolation from one another. This is found, for example, throughout much of Buddhist modernism and conditions both popular and academic representations of Buddhist praxis. In this view, practice and doctrine can be understood without reference to one another—one version of this that accords neatly with conceptualizing Buddhism as an individual quest for transcendence is that practices are context-neutral and value-free mental technologies of liberation.

5. Philosophy of Language

5.1. Language as Communication

It is almost universally accepted in modern studies of language (i.e., both modern philosophy of language and much of linguistics) that there is an identity between language and communication. This goes largely unquestioned and so becomes axiomatic: that is, it is presented as simply a fact, one that does not require further explanation or justification. This exclusive focus on language as communicative in turn then limits the language uses under consideration to those that accord with this presumption—suggesting that something of a petitio principii fallacy is at work. A key instance of this identification of language and communication is that, given the understanding that to communicate fundamentally means to convey information, the declarative form is taken as fundamental to all other forms of language use.

In his succinct overview of modern academic philosophy of language, Scott Soames explicitly tells us that the philosophy of language as the field is known today takes the declarative sentence and the transfer of information as the starting point of inquiry (Soames 2010). The two are directly related as form and function—declarative sentences have the form that they do in order to fulfill the function of communicating information. This understanding of language then also determines the nature of the third basic concept: meaning. Meaning is understood to be the information communicated. This in turn constitutes a particular theory of mind, one in which non-linguistic or pre-linguistic meaning is encoded in language, communicated from one person to another, and then decoded out of language. This model informs a great deal of the thinking about language, though it has been the subject of serious critique (Deacon 2011). Thus, for example, Henriette de Swart says,

Speaker and hearer use language to communicate. Typically, communication involves a message encoded by the sender into some kind of signal and sent through a channel to the receiver, who receives the signal and decodes the message. In a regular speech situation, there is a speaker who is the sender, and the message is what she wants to get across to the hearer, who is the receiver. The signal consists of sound waves, which encode words and sentences, the channel is the air, and the hearer decodes the sound waves.

(de Swart 1998, p. 1)

Anything other than the information, the meaning, that is received in the process of communication is “noise.” This can be referred to as the “encoding–decoding” theory of language, which is at the same time a theory of mind, also serving for example as a model of memory in some theories.

The encoding–decoding model, however, necessarily runs into the problem of the homunculus (Deacon 2012, pp. 69–72) or what Daniel Dennett has called the “Cartesian theater” (Dennett 1991, pp. 101–39). Who or what conceives the thought to begin with, that is, prior to its being encoded, and who or what understands the thought after it has been decoded? And, if one identifies language and communication, what is the nature of the message to be communicated—is it itself linguistic in nature, or is it conceptually formulated in something like “mentalese”? This is not to assert that there may be some, perhaps more basic, level of cognition that functions to understand the pre-encoding and post-decoding meaning, but the majority of presentations of the encoding–decoding model do not attempt to address the issue (for alternative conceptualizations, see (Lohmar 2012)).

Taking the declarative form as basic, however, is simply stipulative. It is not empirical, but an arbitrary starting point. It is a decision about what to take as fundamental, and once that step is taken in establishing the nature of language, the problems that the philosophy of language attempts to address largely follow from that decision. That decision was made at the historical origins of philosophy of language. Such an interest in origins of the philosophy of language might be considered misguided, an instance of the genetic fallacy. However, as we will see, the philosophy of language continues to be defined by reference to the ways its tasks were determined at its origins.

5.2. History of the Philosophy of Language—In Brief

The philosophy of language continues to work largely within the constraints imposed by its own history. Philosophy of language was born from the work of Gottlob Frege (1848–1925) in the nineteenth and early twentieth century, despite some authors pointing to earlier discussions (such as Plato’s Cratylus and Sophist) (Lamarque 1997, pp. 1–9). As T. Givón notes with sweet irony at the opening of his work on syntax, “Unfortunately, in tracing one’s intellectual lineage as far back as one can see, one often succeeds most admirably” (Givón 2001, I: 1). In other words, such claims are to be understood as constructs initiated for various rhetorical reasons, such as claiming legitimacy on the grounds of a supposed ancient origin, rather than as empirically objective statements of genealogy (in contrast, see (Crivelli 2008)). Despite locating ideas about language in the writings of Plato, we note that Heck and May say that “Frege’s contributions to philosophy of language are so numerous and so fundamental that it is difficult to imagine the field without them” (Heck and May 2008).

The concerns, concepts and categories identified by Frege and his followers, e.g., Russell and Whitehead, remain foundational for the philosophy of language. Since its inception, the focus on the propositional form and the information it conveys has motivated the issues of the truth of statements as perhaps the central organizing concern for the philosophy of language. “Analytic philosophy of language may be said to begin with Frege’s determination that the fundamental thing about the meaning of a sentence is its truth-value” (Morris 2007, p. 231). In pursuit of this conception of meaning a fundamental distinction is made between sentences (what is spoken or written) and propositions (what is thought or conceived). While this might initially seem clear-cut and unproblematic, Ruth Kempson has demonstrated that it is not a simple distinction (Kempson 1977, p. 36).

Further issues then cluster around this central concern with the truth of statements. John Searle explains that “The philosophy of language consists in the attempt to analyse certain general features of language such as meaning, reference, truth, verification, speech acts, and logical necessity” (Searle 1971, p. 1). These concerns have also informed the philosophy of religious language, which we consider next.

5.3. Philosophy of Religious Language

William P. Alston succinctly divides religious language into either speaking to God or speaking about God (Alston 2007). While this is not exactly a universally appropriate place for such an inquiry to start, it does demonstrate the continuing determinative role of theology for the study of religion, including religious language, in Euro-American contexts. Its relevance to this inquiry arises when the standard encoding–decoding model of language is applied to extraordinary language. What concept or thought is being communicated? To whom is it being communicated? The philosophy of religious language in theistic traditions simply extends the standard model by saying that religious language is communication with God or with a god. But mantra and dhāraṇī, at least in Japanese Esoteric Buddhist ritual performances, are not communicative, they are not addressed to anyone or anything. Alston’s understanding of “religious language” is constrained by a theological commitment in such a fashion as to make it irrelevant to understanding mantra and dhāraṇī.

Dan R. Stiver has suggested that historical forms of the philosophy of religious language can be divided into three general understandings. The first is the negative way (via negativa) in which “all words must be denied or negated in order to understand Ultimate Reality truly” (Stiver 1996, p. 15). The second is univocal, that is, since revelation is definitive, there can be no equivocation (Scotus). And the third is an interpretative orientation based on analogy (Aquinas). Although the author claims to be compiling a survey of the philosophy of religious language, the scope of the study is limited by the same kind of theological commitments as is noted above regarding Alston. We note that there is one exception to the exclusive focus on Christianity in Stiver’s work. That exception is a one paragraph discussion of Zen kōans, which are explained as an instance of the negative philosophy of religious language that uses language instrumentally to overcome language. He says that a kōan is “a seemingly nonsensical riddle that is to be the means to satori, or Enlightenment” (Stiver 1996, p. 19). Indeed, Stiver goes on to claim, rather unhelpfully, that “The nonsense is the point” (Stiver 1996, p. 20).

Theistic traditions can extend the concept of language as communication to religious language by simply extending the model to include communicating with a deity. However, when we consider the actual use of ritual language in tantric practices, it is not always—and indeed perhaps rarely—directed to a deity in the same sense. Mantra may be directed to objects or uttered in conjunction with actions, making those actions ritually effective.

5.4. Philosophy of Language and Indic Understandings of Language

Since one of my claims is that understanding the use of extraordinary language in Japanese tantra is best approached from Indic understandings of language, we next consider the question of whether the philosophy of language is conceptually powerful enough to subsume those understandings.

Like Searle, Scott Soames attributes the key developments in the philosophy of language to Frege and Russell. These include the development of symbolic logic, the analysis of quantification, the application of logical ideas and techniques to the semantics of natural language, the distinction between sense and reference, the linking of representational content to truth conditions, and the compositional calculation of the contents of compound expressions from the semantic properties of their parts [which] are all due to Frege and Russell. Philosophy of language, as we know it today, would not exist without them.

We can take each of these key developments in sequential order so as to consider the relation between philosophy of language and Indic ideas of language. We can here only treat each in a summary fashion.

• symbolic logic:
  Douglas D. Daye has suggested that the standard argument structures employed in Indian thought indicate a quasi-variable, but symbolic logic as such requires the use of full variables (cf. Daye 1979, 1988).
• quantification:
  It is possible to interpret the four options of the catuskoti, the fourfold arrangement of possibilities (is: “all A are B,” is not “all A are not B,” both “some A are B,” neither “some A are not B”) as having different quantifications, but as with variables, this was not developed in the direction indicated by Soames.
• semantics as logic:
  There is a concern with semantics, but this is not formalized in logical terms; for an example of the logical treatment of semantics, see the work of Heim and Kratzer (1998).
• sense and reference (Sinn and Bedeutung):
  Mark Siderits has asserted that “It is generally accepted that Indian philosophers of language do not posit sense as a component of the meaning of an expression in addition to its reference” (Siderits 1991, p. 65). He goes on to claim, however, that “recognition of something sense-like is forced” on some thinkers, despite the tradition being “dominated by an extreme realist view of all semantic properties” (Siderits 1991, p. 65). Ołena Łucyszyna has examined in some detail the difficulties of ascertaining the view of the Sankhya school on meaning. She indicates that there is a general absence of clarity about whether meaning is referentially a particular existing entity, or a generalization (Łucyszyna 2016, pp. 305–7), though in either case there does not appear to have been any clear distinction correlating to that of sense and reference.
• representational content as truth conditions:
  There is no corresponding inquiry in Indian thought on language.
• compositional calculus:
  This is an area with a well-developed correlate, particularly in the works of Bhartṛhari regarding the location of meaning in words or in sentences (sphoṭa). Siderits refers to this as the “problem of sentential unity.” He suggests that there are two views generally found in Indian thought. The first is “that each word in a sentence completes its semantic function by introducing its denotation, the task being performed independently of the semantic contributions of the other words” (Siderits 1991, p. 10). The second is “that the semantic function of a word is to denote its referent in relation to the denotata of the other words occurring in a sentence in which that word occurs” (Siderits 1991, p. 10).

Thus, we can see that the categories, concerns, and concepts of the philosophy of language, such as those summarized by Soames above, do not map directly onto those of Indian thinkers. The problematic character of attempting to understand Indic thought regarding language within the framework provided by the concerns, concepts, and categories of Anglo-American philosophy of language has in fact already been noted by several authors.

McRea and Patil call attention to this in their study of Jñānasrīmitra’s exposition on exclusion, apoha. They note for example, that although “Modern scholars have tended to view the theory as primarily a theory of meaning … since its origins in the work of the sixth-century Buddhist philosopher Dignāga, the theory of exclusion was used to address a much broader range of philosophical problems. Indeed, for Dignāga and his successors, it formed the basis of their account of all conceptual awareness” (McCrea and Patil 2010, p. 1).

B.K. Matilal summarized the difference between Anglo-American philosophy of language, i.e., “modern analytical philosophers” and “classical Indian philosophers.” Citing Dummett (1980, p. 442), Matilal notes that modern analytical philosophers are concerned with “the fundamental outlines of an account of how language functions—it is in this sense that the philosophy of language is to be regarded as the foundation of all the rest of philosophy. This overall concern with how our language works was not the chief concern of most classical Indian philosophers.” (Matilal 1990, p. 4). For the latter “language acquired a metaphysical importance” (Matilal 1990, p. 4). Although Bhartṛhari comes close to the concerns of modern analytical philosophers, in that he “sometimes asked questions similar to those of post-Fregean philosophers, what he supplied as answers may not engage the attention of modern analytical philosophers. Hence we must conclude that the basic attitude as well as the conclusion was very different, as far as classical India was concerned” (Matilal 1990, p. 4). Bhartṛhari’s understanding of language is deeply influenced by the identity of language and being and consequently epistemology as well. “If the language principle (śabda-tattva) is the stuff the universe is made of, it follows that its sentient manifestations, and in particular human beings, are never divorced from language, for their capacity for knowledge (in fact their very consciousness and self-awareness…) is innately infused with language” (Vergiani 2016). Similarly, Vergiani notes that Bhartṛhari “presents his views on language and cognition within the framework of a grand nondualist metaphysical vision, agilely moving from the ultimate essence of reality to the technicalities of Pāṇinian grammar, not unusually in the context of the same argument” (Vergiani 2016).

In an examination of Sanskrit grammar, Houben gives a survey of the study of semantics in which he gives a qualified agreement with Frits Staal’s suggestion that “Indian linguists in the fifth century B.C. knew and understood more than Western linguists in the nineteenth century A.D.” (Staal 1988, p. 47). Houben suggests that Sanskrit authors may have been “ahead of modern developments … in the field of semantic theory and the development of systems of knowledge representation” (Houben n.d., p. 10). Having said that he goes on to remind us that

But with all this it should not be forgotten that … semantics in the Sanskrit tradition never became the well-defined domain of a separate discipline. The exegetes, logicians and grammarians that dealt with semantic issues were first of all disputing on the ontological and epistemological status of linguistic, semantic, and ‘real-world’ entities.

(Houben n.d., p. 10)

In pursuing comparative studies, such as different systems of thought regarding the nature of language, it is important to avoid simply identifying corollaries between two systems of thought and concluding on the basis of those similarities that the two are the same, or further that because they appear in that way to be the same, one can be subsumed within the concepts, categories, and concerns of the other. Much of what constitutes modern philosophy of language as understood by its contemporary practitioners fails to correspond to premodern Indic ideas of language, and that which does is at best rather approximate. This intellectual limitation to the founding conceptualizations of the philosophy of language has made it a markedly “parochial” specialization with little or no attention to comparative studies. By “parochial,” I mean a highly constrained understanding as to what concerns, concepts, and categories are relevant to a philosophical inquiry into language, which serves as a self-limiting constraint on the scope of the field. More specifically for our inquiry here, the concerns, concepts, and categories of the philosophy of language do not have even approximate analogues in tantric uses of language.

In conclusion to this section we can quote a passing comment from Johannes Bronkhorst. In an essay on the concept of sphoṭa, the meaning of a sentence as a whole, Bronkhorst points out that

The modern study of sphoṭa and related issues is contaminated by ideas borrowed from Western philosophy and linguistics to the extent that a major intellectual effort is required to understand these concepts once again in their own cultural context.

(Bronkhorst 2005, p. 5)

6. Explanations of Tantric Language

Having examined the philosophy of language as a possible resource for understanding the use of mantra and similar language forms in Japanese tantric Buddhism and determined that its founding assumptions make it irrelevant to our concerns in this study, we now turn our attention to other explanations. As suggested to me by Ronald Davidson, there has been a shift from seeking tools of analysis in the philosophy of language to exploring the field of linguistics for this inquiry (Davidson 2016).

While there may be others, I will here focus on three approaches to extraordinary language that have been developed in the Euro-American study of tantra: magical language, performative language, and pragmatics.

6.1. Magical Language

The idea that extraordinary language use can be adequately interpreted under the category of “magical language” has been proposed, claiming that the latter is a meaningful category that includes mantra. While it is unnecessary to review the tendentious history of the category of magic as a scholarly term in religious studies in full here, the juxtaposition between magic and religion, and between those two and secular reason (see Comaroff 1994, p. 303) does inform representations of tantra and of the uses of extraordinary language.

When considering the category of magic, we note the postmodern insight that all usages of categories are located, that is, their meanings and significances change depending on the social, historical, intellectual and religious context within which they are used (for an uncontextualized use, see (Beyer 1978)). As noted by Patton Burchett, the category “magic” and its connotations in contemporary academic discourse is formulated in the nineteenth century, as part of attempts to distinguish religion from magic and superstition (Burchett 2008, p. 808). Religion was defined either as rational or as grounded in immediate experience, while magic and superstition were presented as irrational and grounded either in natural fear or in fear induced for the maintenance of social power.

Application of the category magic beyond its rhetorical functions in modern discourse is doubly problematic. First, the sources upon which the characterizations of magic have been constructed are those of the ancient Near East and medieval Europe. When universalized (sublated) as categories applicable to traditions outside the historical sources of the concept, such applications are problematic, and would require adequate theorization before being deployed in analyses—what is meant by the term, and what makes it appropriate as a universal category?

Second, there is also the semiotic pairing of religion and magic as opposites. In this rhetoric, the language of religion is represented as properly rational and discrete from the world to which it refers, while the language of magic improperly confuses words and their referents. This latter confusion supposedly then leads to the delusion that by manipulating words, one can control the world, a way of distinguishing between religious and magical language that remains current. Clearly, we are here in the realm of high modernism in which we moderns are distinguished from those primitives by our instrumental rationality (see Samuels 2005, p. 345, n. 12).

There is more to this, however, than simply a division between “we moderns” and “those primitives.” Robert Yelle’s work indicates that Protestant concerns with definitive language came into conflict with Indic practices, such as the use of mantras. The religious engagement of Protestant missionaries contributed to interpreting Indian religious practices as confusing words and reality, which is indeed an identity asserted by some tantric metaphysical theories, but then conflating that confusion with the distinction between “true prayer and illicit magic” familiar from Reformation history (Yelle 2012, p. 116).

6.2. Contortions Needed to Avoid Connotations of “Magical Language”

Thomas Kasulis has attempted a comparative study of conceptions of language in Plato and Kūkai. Despite writing as if the comparison is between two figures, Plato and Kukai, Kasulis is actually engaged in a three-way comparison. The invisible third term of the comparison is contemporary philosophy and philosophy of language. Implicit in Kasulis’ writing is the presumption that intellectual history is a teleological development, leading up to the form of philosophy found in the present in the West. This conception of intellectual history is reminiscent of the historiography of Hegel and Comte. Given the implicit teleology, present-day forms are functionally normative, and constitute a suppressed term of the comparison. He says, for example, that today a discussion about the nature of language “would involve an analysis of word, meaning, and reference. To the extent that the Cratylus inquires into the connection between names and the things to which they refer, the discussion is modern” (Kasulis 1982, p. 394). This teleological basis is evident when Kasulis asserts that both Plato and Kūkai “tried to go beyond the simply magical into the philosophical” (Kasulis 1982, p. 392).

Kasulis opens by suggesting that philosophy must necessarily consider language.

Early in its development, a philosophical tradition will consider the nature of language, for language is, after all, the medium of philosophical expression. To be truly philosophical, an inquiry must have at least a rudimentary theory about the relationship between words and nonlinguistic reality. This does not mean that every cultural tradition will make the same initial decision about this relationship. We only require that a preliminary theory of language be logically consistent and a reasonable reflection of at least some aspect of language as used in everyday life.

(Kasulis 1982, p. 392)

Rhetorically, this serves to make whatever is said about language by Kūkai philosophical, without critically reflecting on the significance or consequences of attributing that category to Kūkai’s thought. Kasulis also makes a rhetorical move establishing the meaningful comparability of Kūkai and Plato by setting both in opposition to what he simply claims is an archaic view of language, which is seemingly his own speculative construct of a unitary conception that “words are generally believed to have a creative, magical power” (Kasulis 1982, p. 392). In support of his suggestion that there is such an identifiably coherent “archaic view,” he provides one instance each from the Bible, Ṛg Veda, and Daodejing.

In doing this, Kasulis first divides the field of study into magic and philosophy and then simply asserts that Plato and Kūkai both belong on the same side of that divide. This seems then to be taken as obviating the necessity for any further consideration of context—both are philosophers, so other factors may be passed over in silence. One of the consequences of Kasulis’ attempt to make Kūkai a philosopher, rather than a magician, is that then by definition, Kūkai’s view of language is not magical. This, however, is simply a petitio principii fallacy. Kasulis’ comparison is enabled first by decontextualizing both Plato and Kūkai. He then recontextualizes them as figures to be viewed with “both reverence and tolerance” (Kasulis 1982, p. 393) given that they initiate the philosophy of language in the history of thought, which is itself cast as a single, unilinear progressive sequence culminating in our present understandings.

While apparently never made explicit, the absence of discussion of the limitations of the philosophy of language seem to imply that it is presumed to be universally applicable. It may be that it can be applied to any language, within the confines of the original issues that organize the field today. That kind of universality of application, however, is different from asking whether the structure of the field matches other approaches to the study of language. Indian thought provides a very useful comparison for this latter question, as it is itself a highly developed approach—one that indeed begins systematic study of language several centuries earlier than did Euro-American thinkers.

6.3. Performative Language

Describing mantra as performative has been one approach to explanation within the context of philosophy of language. We will examine a specific instance, that of Harold G. Coward and David J. Goa, after introducing the concept of performative language.

The concept of performative language was introduced by J.L. Austin in his 1955 lectures How to Do Things with Words (1962, 2nd ed. 1975). Influential in bringing this approach to the attention of scholars in religious studies was Tambiah application of it to ritual (Tambiah 1985). This approach challenges the conception of sentences as being meaningful by reference, that is, as answering the question “Does this sentence give an accurate representation of some aspect of the world?” This is the characteristic of declarative sentences, that is, statements of some fact—what Austin called “constative,” also called propositions in the literature. But despite their grammatical form often making them look as if they are declaratives, rather than describing some aspect of the world, performatives are an action, that is, uttering a statement is itself taking the action that otherwise be thought to be described in the statement (Austin 1975, p. 70).

Familiar examples of performative language are promising and apologizing—“I’m sorry to have been late this time, and assure you that I’ll be punctual next time.” Such actions are—rather evidently it seems to me—to be ones taking place in the social, intersubjective world, that is, as part of a communicative relation between people. To my mind, however, it is necessary to distinguish between performative speech acts and the actions performed by different kinds of speech acts. The failure to distinguish these two, admittedly easily confused, dimensions leads to the frequent and I believe mistaken equation of performative language use as such and speech acts in general.

Coward and Goa employ the idea of performative language in their explanation of mantra, and do conflate performative language use and speech acts in general. Discussing the use of mantra in the Ṛgveda, for example, they say

Mantras spoken in ritual activity actually do something. Thus, even from their earliest conception in the Rgveda, mantras are classical examples of what are now called “speech acts.” However, whereas modern thought emphasizes the performative nature of the speech act at the expense of its communication function, the Rgvedic seers are emphatic that while the spoken mantra can do things, the basis for that power is located in its revelation of truth (ta, satya). Although mantras from later Indian tradition can be of a nonsensical nature, in the Rgvedic period it seems clear that a mantra must reveal meaning. Indeed the original sense of mantra as a vehicle for reflection and revelation becomes operationalized by the late Rgvedic seers and priests. “Mantra is the tool, the mechanism, for yoking the reflective power of the seer into the machinery of ritual.”

(Coward and Goa 2004, pp. 16–17; internal references to Austin, Tambiah, and Findly elided.)

We cite this at length in order to exemplify some of the many problems involved in such an attempt to interpret mantra by reference to the idea of performative language use, as well as incidentally some other claims made about mantra.

First, what do the authors mean when they say, “actually do something”? The authors do not specify what kind of “actual effect” the mantra is supposed to have. While an interpretation of such actions according to academic conceptions of magic would conclude that the intended actual effects are physical or material in character, this does not seem to be the understanding of some members of the tradition itself. For example, the question of what effects mantra produce appears to have been of concern to Patañjali as well. George Cardona tells us that

in the course of a rite, the Agnīdh officiant is supposed to heat up potsherds in which offerings will be heated by covering them with coals, and as he covers them the Agnīdh is supposed to utter the mantra bh´ṛgūṇām áṅgirasāṁ tápasā tapyadhvam (Vājasaneyisaṁhitā 1.18) ‘Heat up with the ascetic heat of the Bhṛgu, of the Aṅgiras.’ As Patañjali remarks, the officiant addresses the potsherds with this mantra after he puts them on the fire, although even without the mantra the fire, whose very action is to burn would heat them. A restriction is provided, whereby the act thus performed produces the good results desired, felicity.

(Cardona 1990, p. 11)

Second, what do Coward and Goa mean when they claim that modern theory emphasizes performative at the expense of communicative? (there seems to be instead resistance to this division; and they seem to have confused speech acts as a general category with performatives).

Third, what is “truth” (ta, satya) in their usage? Presumably they do not mean epistemic truth, since it is unlikely that truth conditions could be formulated for mantra. Given the Indic context, perhaps they mean something rather like “actuality” or “true being”? In the background, I take it, is the concept of an act of truth (P. saccakiriya, or saccavacana, Skt. satyavacana).

Ronald Davidson comments on this kind of claim regarding the “truth” of a mantra, saying

the idea of a mantra-dhāraṇī as an act of truth expresses several somewhat problematic questions. First, it is evident that many of them cannot be truth in the sense of truth-conditional semantics, for at least some do not exclusively express statements comprehensible in natural language. Second, their presumption of efficacy appears dependent on the narrative of initial expression, and the needs of the individual in the subject case of the narrative … the direction of fit is to refashion the world into a vessel that is in accord with the intention and desire of the speaker, who speaks the words of the Buddha through his own mouth.

(Davidson 2014, p. 51)

Fourth, what do they mean by “reveal meaning”? Are they equating “truth” (or as others do, purpose) with “meaning”? What kind of meaning are they intending to claim mantra reveal? Fifth, in the quote from Findly, it is the “reflective power of the seer” that seems to be operational. But they do not explain what that is—how is reflection powerful? How does such power effectuate mantra? The assertions made by Coward and Goa appear to be dense with significance, but upon critical examination, so much is left unexplained as to leave the reader with no significance. Perhaps the philosophically most contentious issue is their assertion that performative language also communicates meaning.

Whether performatives are also informative is deserving of serious reflection. While some authors maintain that a performative simultaneously functions to inform, in my own view this position conflates an action with its interpretation. If we are to take a performative as an action seriously, then just as we would separate the action of throwing a ball from interpreting the act to mean that the shortstop intends to catch the batter out at first, we should also separate verbal actions from their interpretations. Since meaning is contextual, it is analytically important to separate action—even verbal acts—from interpretation regarding the meaning of such an act. Much confusion in this area seems to result from failing to distinguish between an analysis—in which action and interpretation are separate—from a phenomenology of speech acts in which action and interpretation may be experienced as necessarily conjoined.

John Searle and Daniel Vanderveken specify that “In general an illocutionary act consists of an illocutionary force F and a propositional content P” (Searle and Vanderveken 2009, p. 1). However, when mantra as an entire category are identified as performative, as by Coward and Goa, we are outside the purview of Searle and Vanderveken’s initial qualifying clause, “in general”—performatives having a unique kind of illocutionary force that is not propositional. If we employ the distinction made by Searle and Vanderveken, then being performative per se—doing something in ritual—does not equate to having propositional content, i.e., meaning with truth value.

For the study of extraordinary language used in ritual, propositional content needs to be analytically separated from performative illocutionary force, i.e., what is done in a ritual, including uttering a mantra, is not propositional content—an action is not a proposition. One can interpretively state a proposition about what is done, but that is a second step, i.e., a separate speech act. As a separate speech act, it and its propositional content cannot be identified with the mantra and its performative illocutionary force. Making this distinction between performative language and propositional content points to that set of speech acts that fall outside what Searle and Vanderveken have indicated to be the case “in general.”

Consideration of this distinction is made difficult by the tendency of many writers to say of all speech acts that they perform some action. But a careful distinction needs to be kept between performing an act of saying something and performing an act by saying something. In other words, there is a need to clearly distinguish between performative speech acts and the performativity, i.e., illocutionary force, of speech acts. As just noted, Searle and Vanderveken indicate that speech acts have illocutionary force. Although speech acts have performativity in that sense, this does not mean that they are performative speech acts in a narrow sense as first introduced by Austin. To claim that identity is a fallacy of equivocation—two different meanings being confused/conflated because the same word is used for both. That conflation so diminishes the significance of the category of performative speech acts that it no longer has any meaning. Under this degraded concept of performative as identical with performativity, all speech acts come to be thought of as performative. Unfortunately, the blurring of these two meanings is found in Austin’s original work (Hornsby 2008, p. 905). When the performativity of speech acts is confused with performative speech acts, the significance of the difference between a (verbal) action and its (verbal) interpretation becomes lost.

6.4. Pragmatics and Extraordinary Language

As mentioned above, Ron Davidson has suggested that the tools of linguistics are more appropriate for the study of extraordinary language than those of the philosophy of language. Davidson chooses pragmatics. Ray Jackendoff explains pragmatics by contrasting it with semantics, “one may view ‘semantics’ as providing the part of the conceptual structure of an utterance that is related directly to linguistic expression, and ‘pragmatics’ as providing the part that arises through inference, heuristics, world knowledge, and understanding of the context” (Jackendoff 2010, p. 8). Given the character of Davidson’s inquiry, he suggests that historical pragmatics provides entrée to understanding the use of dhāraṇī. While pragmatics per se generally focuses on spoken source materials, historical pragmatics, which has the same goals, depends entirely on written texts as sources. Davidson’s hope is that a number of areas of investigation will be facilitated by this kind of approach to dhāraṇī texts: “the changing nature of Buddhist discourse, the macro social environment of India, the intermediate social environment of Buddhist monasteries and the support communities, the smaller environments of Mahayanist groups employing such phrases and placing them in their scriptures, to mention but the more urgent” (Davidson 2014, p. 10).

In his examination of the ways that dhāraṇī are talked about, Davidson identifies nine different categories of pragmatic context. These are the following:

(1)

precipitating assertives, descriptions of the situation in which a dhāṛaṇī is to be taught;

(2)

directives, assertions that one should use the dhāraṇī;

(3)

commissives, promises to teach the dhāraṇī;

(4)

precedent or predictive assertives, previous or future uses of the dhāraṇī;

(5)

benefit assertives, descriptions of the benefits to be derived from use of the dhāraṇī;

(6)

hybrid warning assertives, warnings to those who do not adhere to or accept the efficacy of the dhāraṇī;

(7)

expressives, praise, homage, etc., regarding the dhāraṇī;

(8)

perlocutionary expressives, descriptions of the joy and amazement of the audience receiving the dhāraṇī; and

(9)

the mantra-dhāraṇī itself (Davidson 2014, p. 11).

Davidson maps these different pragmatics to corresponding Sanskrit grammatical forms, demonstrating that the categories are reflections of actual textual usages.

One of Davidson’s goals is to establish historical pragmatics as a tool for studying the historical development of dhāraṇī texts. His second goal is more broadly methodological. It is “to determine if nomothetic linguistic tools can be effectively brought to bear on the evaluation of dhāraṇī texts, rather than inventing, on the spot, a typology of dubious validity, neither really emic nor etic, or invoking emic typologies exclusively” (Davidson 2014, p. 52). What is of value here is that such a method provides consistency across various inquiries into the use of dhāraṇī, and more broadly extraordinary language in general. While the specific categories identified by Davidson may not “translate” into languages other than Sanskrit, only by consistency across different studies can progress in our understanding of how extraordinary language was talked about as ritually effective be made. In relation to the question motivating my own research, how did Japanese tantric practitioners understand language in such a fashion that validated the use of extraordinary language originating in Sanskrit as effective? Understanding how extraordinary language is presented in the textual tradition is an important factor.

7. Conclusions

The concerns, concepts, and categories that inform the philosophy of language preclude its application—or imposition—onto the study of extraordinary language use in tantric traditions, such as those of Japan. This becomes clear when we consider the question, how did Esoteric Buddhist practitioners in Japan conceive of language so that mantra, dhāraṇī, and other kinds of extraordinary language originating in the cultural context of Sanskrit could be thought to be effective? This is not simply a case of some unsophisticated and undifferentiated “archaic view of language” a là Kasulis. Instead of dismissing this as just a magical confusion of word and world, as will be explored in more specificity in additional research (Payne 2018), this use in Japanese tantra is actively informed by Indic conceptions of language.

The use of extraordinary language in Esoteric Buddhist practices in Japan indicates that Indic understandings of language formed part of the larger culture of Buddhism and accompanied practices employing extraordinary language. In the study of Japanese Buddhism in European and American scholarship, little attention has been given to this kind of intellectual transmission from India through China to Japan.

The central role of mantra, dhāraṇī, and related forms of language use in the tantric traditions is well-established, so much so that tantra and mantra are sometimes identified with one another. Mantra is so central to tantric praxis that the names used for tantric Buddhism include “mantranaya” (the method of mantra) and “mantrayāna” (the way of mantra). The centrality of language for these traditions indicates that further research focused on such usages is warranted. To date, however, the philosophy of language has not formulated any significant contributions toward the study of such language use because it is constrained by its own original formation. More promise is held though for the application of the analytic tools of linguistics. In addition to the application of pragmatics, as per Davidson, cognitive linguistics, socio-linguistics, historical linguistics, and semiotics (Rambelli 2013; Yelle 2013) may be of use in further studies.

The complexity of the subject of extraordinary language necessitates a careful balance between overarching concepts that provide a basis for meaningful comparisons, and attention to the actual use and function in specific contexts. More generally, neither a single definition nor a single explanation of mantra, dhāraṇī, and so on can be a meaningful goal of our studies.