1 Introduction

A scholar unfamiliar with Buddhist philosophy of language would upon closer inspection find both similarities and notable differences from corresponding discussions in the Occidental intellectual tradition. While both traditions agree on the philosophical importance of elucidating how, if at all, language can be about the world, the Buddhist approach to this question has a distinctive orientation that makes it markedly different from corresponding motifs in Western philosophy. To a considerable extent, this could be attributed to the specific ontological and epistemological assumptions on the impermanent and momentary nature of everything that comes to be. Noting that language by contrast operates with generally applicable categories, Buddhist philosophers tend to conclude that language must be incapable to capture the fleeting nature of reality. Clinging to the categories postulated by language, therefore, becomes a shackle that reinforces the reificatory tendencies of the mind (kleśa), and hence keeps us locked in a mode of perpetual dissatisfaction (saṃsāra). At the same time as reality is ultimately ineffable, Buddhist philosophers nevertheless insist on the indispensability of language (śabda) and conceptual thinking (kalpanā) in spiritual practice (mārga). To go beyond language and attempting to reach the ineffable character of reality (tathātā) still requires language, which places it in the position of being both a help and a burden on the path to enlightenment.

In this paper, we elucidate the core points on language in the Buddhist tradition, and, where appropriate, provide contrastive comparisons with discussions of similar issues in Western philosophy. Though common ground can be found on some key issues, Buddhist philosophy provides a unique account of language and conceptual thinking as distorted and ultimately misleading in its apprehension of reality. These comparisons between the respective traditions therefore primarily serve the purpose of providing an initial orientation of Buddhist thought, and should not be seen as direct parallels. We begin in Sect. 2 with recapitulating Buddhist philosophy of language. To do so, we depart from the central epistemological and ontological assumptions of momentariness and impermanence in Buddhist philosophy. The soteriological ambition of Buddhist contemplation is also highlighted as a characteristic feature for determining the validity of philosophical doctrines. We then provide an overview of how these assumptions and ambitions come to be expressed in Buddhist philosophy of language, particularly noting how mainstream Buddhist thought tends to equate linguistic meaning with conceptuality and the nominalist inclination of considering language as taking part in the formation of (ultimately fictitious) universal categories (sāmānya). By contrast, the Madhymaka school rejects that linguistic expression have meaning independent from context, and instead emphasises how language is best characterised in terms of the instrumental and social functions it serves. This section lays the necessary groundwork for Sect. 3, where we dive deeper into the role of language as simultaneously a burden to be overcome and a necessary starting point for meditation (dhyāna). We discuss how ineffability is approached in two different ways, emanating from the different analyses of language in mainstream Buddhist philosophy and Madhyamaka. These two readings are critically compared to the early and the later Wittgenstein’s treatment of what the limits of language imply for philosophical practice. Section 4 concludes with a discussion on language and ineffability in both philosophical traditions.

2 Philosophy of Language in the Buddhist Tradition

2.1 Context and Background

Buddhist philosophy of language, as part of Indian philosophy, emerged during what is commonly referred to as the Classical period of Indian history, which falls between the inception of so-called schools of tenets (darśana) from c. the fifth century BCE and the Moslem raids in the eleventh and twelfth centuries (Radhakrishnan 1951, pp. 21–32). During the Classical period, preceded by the Vedic period beginning with the Aryan conquest of the Indian subcontinent (Panikkar 1960, pp. 14–41), major indigenously Indian religious systems such as Buddhism and Jainism were formed. This was also the period when their canons of doctrinal texts were composed, and religious infrastructures established. The philosophical elaborations of religious and spiritual views, so-called darśanas, are traditionally divided into the orthodox āstika schools, which accept the authority of the Vedas, and the non-Hindu nāstika schools, which reject it. The most influential āstikas are the following six schools: dualistic Sāṃkhya based on the writings of Kapila (c. fifth century BCE), ritualistic Mīmāṃsā derived from the teachings of Jaimini (c. third century BCE), theistic and atomistic Vaiśeṣika, which grew out the activity of Kaṇāda (c. second century BCE), spiritualistic Yoga inspired by the work of Patañjali (c. second century BCE), the idealistic doctrine of Vedānta going back to Bādarāyaṇa’s writings (c. third century BCE) and Nyaya, the school of logical realism founded by Akṣapāda Gautama (c. second century CE). The best known nāstikas include philosophies formulated within the two non-Hindu religious movements, Buddhism and Jainism, as well as the materialistic and anti-intellectualistic Cārvāka, which takes Brhaspatisūtra as its foundational text. Although the original text was lost, it is supported by extensive discussions and commentaries in post-Vedic literature.

The significance of darśanas, apart from the content of their respective philosophies, lies in that they delimited the scope of reflection during the Classical period. They did so by providing thinkers with a cognitive map, which helped them locate their own views against a larger philosophical background. Such a way of philosophising expresses, as argued by the prominent Indologist Georges Dreyfus, “the commentarial style of Indian scholarly tradition”, whereby:

… all philosophical activities rely and are intended to validate the framework given by the tradition; … philosophical problems are not discussed only on the basis of their philosophical merits but in relation to and under the form of commentaries to some basic text formative of the tradition (Dreyfus 1997, p. 4)

The “commentatorial style” captures an essentially scholastic sentiment of Classical Indian philosophy. Usually, authors put forward a position as an elaboration of views found in their own darśana and in opposition to those in other darśanas. In fact, the presence of an antagonist often plays a more important role in articulating a view than reference to one’s own tradition. For example, the formulation of the doctrine of śūnyatā (“emptiness”) by Candrakīrti (c. 600– c. 650 CE) of the Madhyamaka school of Buddhist thought, is mainly done by rejecting the positions attributed to his adversaries. The reasoning follows the format of catuṣkoṭi, or tetralemma, where argumentation examines four logical possibilities of a proposition: (a) something is, (b) is not, (c) both is and is not and (d) neither is or is not (Jayatilleke 1967).Footnote 1 Candrakīrti lays out his view of emptiness by critiquing the idea of substantially existent (svabhāva) cause and effect. In doing so, he rejects the views that (a) cause and effect are essentially the same, which is attributed to Sāṃkhya, (b) that they are essentially different, which is attributed to Vaiśeṣika, (c) that they are both the same and different, which is identified as the Jain position (d) and finally that there is no causation, which is attributed to Cārvāka (Candrakīrti 1979; cf. Cabezón 1994, p. 181; Rizzi 1988, p. 12).

Similarly to representatives of medieval scholasticism like William of Ockham, John Duns Scotus and Thomas of Erfurt, Indian philosophy had a keen interest in language and its relation to the world and mind (Cabezón 1994). In India, there was a strong grammatical tradition, which goes back to Pāṇini (fourth century BCE) and his students, who were mainly concerned with codifying rules for Classical Vedic Sanskrit (Żywiczyński 2004). This intellectual movement formed a basis for discussing more philosophically oriented problems, such as the ontological and epistemological foundations of meaning or the expressibility of reality. The problem that organised philosophical debates about language was formulated by Patañjali (see above) in the treatise on Sanskrit grammar Mahābhāṣya: “What is the object of words?” The Indian philosophical tradition worked out two solutions to the problem, attributed to the ancient grammarians Vājapyāyana and Vyāḍi, who both probably lived shortly after Pāṇini (Herzberger 2011).

Vyāḍi argues that words directly apply to individual substances (dravya); hence, he asserted that reference is the primary function of languages. To give a standard example from the Indian literature, the function of the word “cow” is to pick up an appropriate individual in the context where the word is uttered (Hiriyanna 1938). The word performs this function by eliminating its contradictory, i.e. everything that is “non-cow” (Dreyfus 1997, p. 207). Vājapyāyana takes an opposite view and claims that words perform the referential function only through the mediation of universalsFootnote 2 (sāmānya, lit. “generality”), defined as properties (viśeṣaṇa) common to a number of individuals (dravya; Tenpa Gyaltsen 1999, p. 21). In his view then, the main function of language is significative, whereby e.g. the word “cow”, primarily indicates the property “cow-ness”, which is then used to identify individuals depending on whether they possess this property or not (Hiriyanna 1938). The above difference leads to different views on the cognitive role of language. For Vājapyāyana, language mainly serves to identify commonalities (saṃsarga) between individuals, e.g. to identify that the attribute “cow-ness” is common to all domesticated ungulates of the genus Bos (Hiriyanna 1938). Vyāḍi sees language mainly working on the principle of exclusion (bheda), whereby words do not primarily serve to identify commonalities but to distinguish things to which a word, such as “cow”, can correctly be applied from all other things like horses, dogs, cats, etc. Rightly or wrongly, Western scholarship designated Vājapyāyana’s and Vyāḍi’s respective positions as “connotationism” and “denotationism” (e.g. Dreyfuss 1997). We will follow this usage here, without critical discussion whether these terms are appropriate or not.Footnote 3 With the notable exception of Madhyamaka (see Sect. 2.5), the major Buddhist schools of thought (siddhantha) accept Vājapyāyana’s connotationism but factor in the notion of exclusion derived from Vyāḍi’s proposal. Before discussing Buddhist proposals concerning the nature of meaning, we must introduce some of the basic ontological and epistemological assumptions of Buddhism.

2.2 Basic Ontological and Epistemological Assumptions in Buddhist Philosophy

The Buddhist ontology is committed to the view that existence (sat) is only attributed to things (bhava) that are impermanent (anitya). Impermanent things are those that exhibit the four characteristics of arising, abiding, decaying and disintegrating (Tenpa Gylatsen 1999, p. 8). In a stronger sense, impermanence is equivalent to momentariness (kṣaṇika), whereby things manifest the four characteristics within the span of one moment (kṣaṇa).Footnote 4 Accordingly, only momentary things are ultimately existent (paramārthasatya), while all other things, such as external, common-sense objects (tables, chairs, trees, etc.) or self (jīva), are wrongly assigned the property of existence by a mind deluded by ignorance (avidya). For example, Vaibhāṣika, the dualististic Buddhist school of thought, posits that there are two types of existents: (i) configurations of particles (paramāṇu), which arise, abide, decay and disintegrate within one moment, and give rise to similar configurations of particles in the next moment; (ii) momentary manifestations of consciousness (mānas), which give rise to subsequent momentary manifestations of consciousness (Ponlop Rinpoche 1998, pp. 39–52). To account for the seeming stability of things, existence is defined in terms of causal effectiveness (arthakriyāsamartha), i.e. the ability of a thing to produce later instants of its own continuum (Ponlop Rinpoche 1996). By teaching us that the stability of common-sense objects and our own mental processes are ultimately an illusion (māyā), the Buddhist ontology is committed to considering “change [as] more fundamental to reality than stability” (Dreyfus 1997, p. 60). As Dreyfus notes, such a process-based ontology is different from the typical emphasis on existence in Western philosophy. At least since Parmenides, existence and permanence have typically been equated with ontological categories such as forms, substances and universals considered metaphysically more fundamental than variation and impermanence. In terms of its commitment to change as a pivotal ontological assumption, the Buddhist position resembles the Pre-Socratic philosopher Heraclitus, who famously argued against permanence, instead claiming that everything remains in perpetual change.Footnote 5

The doctrine of momentariness is fundamental for Buddhist epistemology. Most Buddhist schools of thought, with the notable exception of Madhyamaka, accepts Nyaya’s theory of pratyakṣa, commonly translated as perception (Żywiczyński 2004, pp. 52–54). Pratyakṣa is a mental activity (mānas) that is (i) produced from the coming together of consciousness and an external object, (ii) non-verbal (avyapadeśya), (iii) non-erroneous (avyabhicāra)Footnote 6 and (iv) definite (vyavasāyātmikā), in the sense that it correctly identifies the category to which the apprehended object belongs (Dreyfus 1997, p. 334). Buddhist epistemology accepts the first three criteria and understands them as moments of consciousness which apprehend momentary manifestations of external objects, i.e. sense perceptions (indriyapratyakṣa) of colours and shapes, sounds, etc. (Tenpa Gylatsen 1999, pp. 3–7). According to Nyayaikas, the last characteristic expresses the ability of pratyakṣa to apprehend universal properties: based on sense data, mind is able to form a perceptual judgement, whereby it for instance recognises that the bulbous, splay-bottomed shape and the colour grey constitute a jar (Dreyfus 1997, pp. 345–348). Such a view on the nature of pratyakṣa reflects Nyaya’s realistic position on universals (sāmānya, see above), which claims that universal properties (such as “jar-ness”; sāmānya) inhere (viśesa) in all individual jars. A likely Western ally to the Nyaya view on universals would be the immanent (or moderate) realism most famously espoused by Aristotle. On such a view, universals really exist, but do so as particularised within individual entities.

Apart from purely epistemological concerns,Footnote 7 Nyayaikas’ motivation to include definiteness as one of the properties of pratyakṣa, considered in Indian philosophy as a paragon of cognitive validity, was to support their realistic solution to the problem of universals: if pratyakṣa apprehends universals, then – their argument runs – universals must necessarily exist (literally, are substantially established; dravyasiddha; Dreyfus 1997, p. 154). To defend its antirealistic position on universals, Buddhist epistemology analogously rejects that definiteness is a property of pratyakṣa. Universals are permanent phenomena (nitya), which do not arise, abide, decay and disintegrate, and hence, in view of Buddhist philosophers, they are non-things (abhava), i.e. entities which are causally ineffective in the sense of being unable to produce later moments of their continua (in the way that particles or moments of consciousness can; Tenpa Gyaltsen 1999, p. 2 and see above). Universals are objects of kalpanā, a form of consciousness distinct from pratyakṣa, which sometimes translates as “imagination” but is more commonly rendered as “conception” in philosophical contexts (e.g. Cabezón 1994; Dreyfus 1997; Sherab Gyaltsen 1997). The Buddhist account of kalpanā stresses its mistaken character (bhrānta). Most Buddhist philosophers underline a causal connection between perceptual experiences communicated by pratyakṣa and the workings of kalpanā (Żywiczyński 2004, pp. 64–65). Commonalities between certain perceptual experiences, such as the colour of tomatoes, blood, fire-engines, etc., give rise to the universal “red”. However, due to the power of ignorance, these commonalities and resultant universals are interpreted by kalpanā as real existents. It is this reificatory tendency of kalpanā that locks us in the experience of perpetual dissatisfaction (saṃsāra); hence, the essence of the Buddhist path (mārgā) to the liberation (nirvāṇa) from saṃsāra consists in uprooting ignorance by eliminating this reificatory activity of kalpanā. To insist on a separation between perception and conception has parallels that go back at least to Plato’s distinction between the sensible and the intelligible world. In complete opposition to Plato, however, Buddhist accounts considers perception to be epistemologically primary. This rather aligns with both empiricism and the primacy of perception in phenomenology. Such a comparison should however be read with caution and limited just to deriving conception from perception, since neither empiricist nor phenomenological analyses would propose conceptual activity to be mistaken or erroneous.

2.3 The Soteriological Goal of Philosophical Contemplation

The mistaken character of kalpanā and how liberation can be achieved bring us to the soteriological aspect of Buddhist philosophy. The ultimate intent behind philosophy (literally “right understanding”, sammāditthi)Footnote 8 is to lead a practitioner towards spiritual betterment, which culminates in enlightenment (bodhi; Reugg 1980). Thus, all philosophical proposals put forward by Buddhist scholarship should not only be considered on its intellectual merits but, primarily, as a means of effecting spiritual progress (Gymatso Rinpoche 1997). Such a soteriological orientation of philosophical reflection is captured by the pedagogy of ascending level of analysis (Dreyfus 1997), which developed in Buddhist monastic universities in India around seventh century BCE and was adopted by Tibetan scholarship (Cabezón 1994). Accordingly, Buddhist schools of thought (siddhanta) were ordered from realist Vaibhāṣika, through hypothetical realist Sautrāntika and idealist Cittamātra to the emptiness-doctrine of Madhyamaka (Ponlop Rinpoche 1998).Footnote 9 The four schools of thought were thought to offer more and more refined insight, culminating in Madhayamaka, which was considered by many late Buddhist philosophers in India and all Tibetan scholars as Buddha’s definitive teaching. Accordingly, once a student mastered the Vaibhāṣika doctrine and meditated on the meaning of its teaching (dhyāna), he or she would go on to study Sautrāntika and so on, in a process of increasing spiritual insight by eliminating subtler and subtler veils of ignorance.Footnote 10

To fulfil its soteriological function, philosophical reflection must in some way be connected to reality. Establishing this connection is a problem for Buddhist epistemology, which – as shown above – is in general suspicious of conceptual thinking and its ability to connect to reality in an undistorted way. According to Buddhist epistemology, pratyakṣa directly (svasaṃvēdya) and reliably apprehends sense objects, and hence it constitutes a means of valid knowledge (pramana; Dignāga 1956a,b; cf. Klein 1986, p. 91). However, pratyakṣa is ineffable (anirdeśya) in the sense of being the non-reflective experience of mere-sensing and as such it cannot combat the reificatory proclivities of mind (Dharmakīrti 1956; Cabezón 1994, p. 129). The ultimate spiritual achievement of the Buddhist path (buddhatva), the state of Buddhahood, is a state free from kalpanā, or conceptual thinking; initially, however, a practitioner must rely on conceptual analysis (Cabezón 1994, pp. 138–139). In fact, most Buddhist teachings (dharma) such as selflessness (anātma), impermanence (anitya) or emptiness (śūnyatā), must be ascertained through inferential thinking (anumāna) before they can be apprehended by meditative insight (i.e. yogic perception, yogipratyakṣa; Gyamtso Rinpoche 1998, pp. 15–16). For this reason, Buddhist philosophy accepts at least some forms of conceptuality, and specifically inference (anumāna), as means of valid knowledge (pramāṇa; Dignāga 1956b). Thus, Buddhist epistemology accepts two forms of valid cognition: perception and inference. The former is grounded in reality but lacks cognitive content; the latter depends on conception, whose operation, as already noted, involves a cognitive mistake of reifying commonalities into as if real universals. How then does Buddhism explain the utility of conception if its objects do not have a grounding in reality communicated by pratyakṣa?

The majority of Buddhist philosophers argue for a causal connection between conception and reality. In doing so, they employ arguments reminiscent of a nominalist position in the Western debate on universals. The most complete explanation of this connection was elaborated by Dignāga (c. 480 – c. 540 CE; 1956a, b) and Dharmakīrti (c. 6 h century CE; 1956) into the apoha theory (lit. “exclusion”). They start with the observation that conception is an active process, which highlights similarities and downplays differences recorded at the perceptual level. This process, i.e. what gets highlighted and what gets downplayed, is regulated by one’s mental propensities (vāsanā; Asanga 1956) and results in schematic representations (pratibimba; lit. “reflection”) of objects. However, schematic representations are too diverse to render a universal; the representations of, for instance, different trees do not on their own give rise to the uniform concept of “tree-ness”. For this to happen, representations must be subsumed under a single word (nāma); in our example, representations of different trees render the universal “tree-ness” only when connected to the name ‘tree’. In this way, “tree-ness” as the proper object of kalpanā is a constructed universal that individual trees are (falsely) assumed to share (Dreyfus 1997, p. 227; Żywiczyński 2004, pp. 70–71).

In this way, Buddhist epistemology attempts to demonstrate the validity of conception and its soteriological function. Conception, although operating on fictive universals, nevertheless has an indirect access to reality communicated by perception. For this reason, the sharp distinction between pratyakṣa and kalpanā is blunted by the intermediary of mental representations, which, though elaborated in the process of highlighting similarities and downplaying differences, are not arbitrary thanks to their immediate connection with percepts. By considering universals as coming about through a mental process of creating universal categories, the Buddhist view seems to have a close affinity not just with nominalism but particularly with a conceptualist position in the Western discussion over universals (Żywiczyński 2004, pp. 70–71).

Another important point is the role of language in the formation of universals. As noted, linguistic labels (nāma) standardise mental representations and in effect lead to the emergence of universals. On the one hand, such a proposal grants language a creative role in the epistemological process. On the other, the social character of language enables the transformation of perceptual experience into types of experiences that are communicable. Language is then indispensable for turning private experiences into something communicable and hence possible to share with others. Viewed in such a way, language offers the possibility for transcending the limitations of perception and attaining a kind of intersubjective and socially shared validity (a point also touched on by for instance Merleau-Ponty 1969; see Żywiczyński 2004).

The apoha theory formulated by Dignāga and Dharmakīrti not only attempts to identify a connection between conception and reality but also to underline the antirealist position of Buddhism on universals. Their foundational assumption is a necessarily dichotomous character of conception (Żywiczyński 2004, pp. 68–69). While pratyakṣa directly apprehends its objects, kalpanā identifies its objects, i.e. universals, by eliminating (apoha) what they are not. Somewhat similar to the structuralist view of oppositions as integral to linguistic categories, the establishment of universals is based on an exclusionary process where a category like “tree-ness” is established by the identification of what cannot be subsumed under it (i.e. all “non-trees”; Dreyfus 1997, pp. 123–124). Importantly, as explained by Tibetan scholar Gorampa (1429–1489; 1968), the view that kalpanā is based on exclusion (apoha) does not pertain to the psychological but to the epistemological level of explanation. Accordingly, in order to entertain the universal “tree-ness”, it is not necessary to eliminate an infinite number of objects that are not trees (e.g. houses, cows, humans, etc.). Rather, the very operation of kalpanā entails exclusion: in order to identify a category like “x” necessarily involves the identification of “non-x”. This is why Gorampa argues that establishing the dichotomy between “x” and “non-x” is required for knowing whether we use “x” correctly. Dreyfus explains the negative nature of concept formation in terms of validation:

Let us consider [an] example, ’this is a cow.’ In this expression we have two elements, the subject and the predicate expressions. The predicate expression, here the word ’cow,’ does not signify a particular cow but cowness in general. This property (cowness) is not some positive entity that can be found in the world but only a difference that conceptually obtains between certain individuals. What about the subject expression? The validation of the subject expression is a bit more immediate since it refers to a real individual. This reference, however, is not direct. The subject expression does not refer directly to a real individual, but only locates the conceptual individuation of this individual within the conceptual space by eliminating the other possible subjects of predication. Hence, for both expressions, the validation is negative, although in different ways. (Dreyfus 1997, pp. 224–225)

We could attempt to establish a category like cow in terms of the different physical properties it might have. Such a procedure would however presuppose that we know the definition of these properties, which is similarly reached by a process of excluding those things that lack these properties. To overcome this problem, category-formation can instead by seen as an enterprise that necessarily creates dichotomies. At a later stage, the fundamentally fictitious nature of concepts is reified into something believed to exist independently, when it is in fact just an effect of a process of exclusion (Dreyfus 1997, p. 225; Żywiczyński 2004, p. 69).

2.4 Universals, Language and Meaning

Having established some basic ontological and epistemological assumptions and the soteriological goal of philosophy, we are now in a position to examine how the Buddhist tradition approaches language. One key assumption about language is its tight integration with kalpanā, to the extent that they are often treated as equivalents, whereby statements about kalpanā can be isomorphically translated into statements about language (Żywiczyński 2004, pp. 73–74). The equivalence is most visible at the level of lexical semantics, where the meaning of a word (artha) is identical with the object of kalpanā, i.e. a universal (sāmānya; Cabezón 1994, p. 119). In this regard, as already noted, Buddhist philosophers follow Vājapyāyana’s connotationism. They insist that words cannot refer to external objects because these are momentary; hence, it would absurdly follow that the term “cow” would have to be learnt anew whenever we apply it to a particular cow (Dreyfus 1997, p. 217). However, this would contradict the very principle of language: we learn words in order to subsume individuals under categories; the meaning of words must therefore be based on some intensional properties, which – on the Buddhist account – are afforded by universals. Since universals are fictive, the Buddhist theory of meaning struggles to show how words can perform the referential function, which is analogous to the problem found in the Buddhist account of kalpanā, and the solution to it is elaborated along the same lines. Lexical meanings are arrived at in a process of exclusion: “… words exclude objects from being included in a class to which they do not belong” (Dreyfus 1997, p. 221); for example, the word “cow” serves to exclude a potential referent (let’s say the cow named “Bossie”) from the category of non-cows. In this view, there are no meanings that positively capture the nature of objects but agreed-on fictions constructed through exclusion which we use to categorise the world. There is however a causal, indirect connection between the world and those agreed-on fictions, which ensures that words can be used to approximate reality (Żywiczyński 2004, pp. 76–79). In this way, the Buddhist account of lexical meaning integrates the intuitions of Vājapyāyana’s connotationism, whereby universals constitute the proper object of words, and Vyāḍi’s denotationism, whereby reference is based on a process of exclusion (Żywiczyński 2011).

The most complete version of such a model integrating denotationism and connotationism was put forward by Dignāga in Pramāṇasamuccaya (1956b). It was later developed by Tibetan scholars, who were particularly interested in the theory of the linguistic sign. In this tradition of Buddhist epistemology, two types of universals are postulated: the object universal (arthasāmānya) and the term universal (śabdassāmānya). The former is the proper object of kalpanā, which has emerged from one’s mental representations related to a particular domain of reality, in the way that for instance the perceptual experience of certain shapes and colours has given rise to the universal “tree-ness” (Tenpa Gyaltsen 1999, p. 22). The term universal is the vocal image of a word which signifies the corresponding object universal; in our case, it is the vocal form /triː/. In this way, Buddhist thinkers try to distinguish between unique phonetic tokens of the word from its underlying type (Tenpa Gyaltsen 1999, p. 22). A term universal is instigated whenever one calls up to mind an appropriate object universal, or hears a phonetic realisation of a term universal, which then leads to the identification of the appropriate object universal. A somewhat similar viewpoint can be found in the structuralist definition of the linguistic sign as comprised of two conjoined elements: signifier and signified (Żywiczyński 2004, p. 84). Just as in Buddhist thought, the signifier is not just a physical sound but its psychological imprint, which Saussure (1959, p. 56) calls the “sound-image.” The similarity to structural linguistics can be further deepened by pointing to the conventional and conceptual nature of linguistic signs. On a structural reading, the meaning of a linguistic sign is empty and depends on its “position” within an interrelated system of signification. Without necessarily having a concept of sign meaning as dependent on relations within a web of signs, the Buddhist analyses still reaches the somewhat similar conclusion that the linguistic sign exhibits an empty relation between a signifying and a signified element without an immediate connection to something actually existing.

As noted above, Buddhist philosophy of language focuses on lexical semantics, and on a more theoretical plane considers lexis as the most important element of language. This attitude goes back to the Indian grammatical tradition initiated by Pāṇini, whereby words (nāma) constitute the fundamental building blocks of linguistic communication. Most importantly, the understanding of the sentence (vakya) is compositional: it depends on the understanding of its individual words and the syntactic relations between them. In this way, a finite lexicon and a finite syntax is able to generate a potentially infinite number of sentences (Żywiczyński 2004, pp. 80–81; Dreyfus 1997, p. 267). This word-plus-syntax account of meaning, which dominated the Buddhist reflection on language, was given the most complete elaborate formulation by Dharmakīrti in Pramāṇavārttika (1956). Such a view was famously challenged by the sphoṭa theory formulated originally by Bhartṛhari (c. fifth century CE), who was affiliated both with Buddhism and Nyaya (Radhakrishnan 1951). In Bhartṛhari’s holistic approach to meaning, the most important element of linguistic communication is the sentence (vakya), which he understands not so much as a structural unit but as a unit of thought. Accordingly, the sentence is a physical expression of the speaker’s idea, which is understood “in a flash” (lit. sphoṭa) by the listener who hears it (Matilal 1990). In contrast to the word-plus-syntax approach,Footnote 11 he argues that the meaning of the sentence is not established by combining the meanings of individual words it consists of, but that the meaning of individual words is established post-factum by the deconstruction of the sentence’s original meaning (Sideritis 1985). This can be compared to the type of non-compositional theories that analyse word meaning as dependent on the surrounding linguistic context, as proposed by for instance Wittgenstein (1961 [1921], p. 25), who claims that “[o]nly the proposition has sense; only in the context of a proposition has a name meaning”. One can also note similarities to the type of semantic holism espoused by for instance Quine (1953) and Davidson (1984). In the mainstream Buddhist tradition, Bhartṛhari had a strong impact on Dignāga, who appealed to sphoṭa in elaborating a mentalist version on connotationism (Żywiczyński 2004, pp. 80–81).

2.5 Limits of Language

One of the most hotly debated issues in Buddhist scholarship concerns the problem of ineffability (anirdeśya) – the extent to which reality is beyond the scope of thought and language (see 2.3). The significance of this problem derives from the assumption central to the Buddhist account of meaning that language signifies fictive properties (object universals). An inevitable consequence of such a position is that reality is outside the realms of language. The problem of ineffability is further solidified by the epistemological postulate that only pratyakṣa, both sense perception (indiyapratyakṣa) and spiritual insight (yogipratyakṣa), is able to connect to reality in a direct and undistorted way. But accepting a view that the gap between reality and language is unbridgeable has unwanted consequences, the most important of which is of soteriological nature. As already noted, the ultimate goal of philosophical reflection is spiritual growth. How then is philosophy able to accomplish this goal if its medium, language, can only engage with fictions created by one’s mind? (Dreyfus 1997, pp. 263–264). In our discussion of kalpanā and reference, we saw how Buddhist scholars try to combat this unwanted consequence of ineffability by positing a causal link between reality, on the one end, and kalpanā and language, on the other. Pratyakṣa engages objects as they truly are (vidhipravritti; lit. “collective engagement”); thought and language also do so but in an indirect and distorted way by constructing (in the process of exclusion; apoha) universals out of particular features of real objects. Therefore, the Buddhist stance, in contrast to the total scepticism of Cārvāka, is that linguistic description can afford knowledge into the nature of reality.

A stronger claim of ineffability was formulated within the Madhyamaka school of Buddhist thought, and particularly by Candrakīrti in the two influential treatises – Madhyamakāvatāra (1956) and Prasannapadā (1979).Footnote 12 Madhyamaka is radically anti-mentalist and adjusts its theory of meaning to espouse this position. The mainstream of Buddhist thought, represented by Dignāga and Dharmakīrti, argues that the stability of meaning is guaranteed by the similarity of mental images (or object universals; arthasāmānya) evoked in one’s mind by a linguistic expression. In contrast, Madhyamikas insist that the meaning of a linguistic expression is not reducible to the mental activity evoked by its use. Instead, meaning is the ability to use a linguistic expression in an appropriate context (askepa; see below). Language thus operates primarily at a social level, which both Nāgārjuna (c. 150 – c. 250 CE), the founder of the Madhyamaka school, and Candrakīrti emphasise by claiming that words do not express any propositional meaning (pratijñā) above and beyond the function they serve when used in actual situations (Garfield 2015, p. 267). Hence, similarly to the later Wittgenstein’s (1958, p. 20) instrumentalism (and very much unlike Vyāḍi’s denotationism), Madhyamikas account for linguistic meaning in terms of context dependence: only uses of linguistic expressions in specifiable contexts can be meaningful. The key notion of the Madhyamaka theory of meaning is prajńapti, often translated as “convenient designation” (Sprung 1979). It is explained by Candrakīrti in the following way:

Prajńapti [is a term] or a way of talking which is paradigmatically tenable and useful; a term or a way of talking which guides toward the surpassing comprehension or awareness; an existential hypostisation. The function of prajńaptis [is] making use of the language and ideas of the everyday. (Candrakīrti 1979, p. 272)

Accordingly, there are no entities that could act as true referents of linguistic expressions. To use Candrakīrti’s own example, the expression “person” (puruṣa) is only a description of someone’s psycho-physical history, a useful hypostisation, above which there is no meaning of “person”. In a similar vein, Wittgenstein (1958, p. 3, 20f, 34f, 48) argues that the “everyday” pre-theoretical understanding for correct and appropriate use is the most authentic form of language that is tacitly presupposed when providing semantic theories in terms of sense and reference.

To account for communicative success afforded by language, Candrakīrti appeals to the connotationist doctrine of contextual implication – askepa – traced back to Kumārila Bhaṭṭa of Mīmāṃsā, which posits that interpretation of ideations evoked by linguistic expressions is regulated by context; for example, “rice” is differently ideated when uttered in a rice field than when used in a restaurant (Żywiczyński 2004, p. 175). However, in contrast to connotationism and the mainstream of Buddhist thought, Madhayamaka scholars reject the idea that universals are the direct object of words and argue that the meaning of a linguistic expression can only be considered in terms of its contextual acceptability and its relations to the contextual acceptability of other linguistic expressions (Sprung 1979). Cabezón provides an eminent summary of the radically conventional and dependent character of the relation, or lack thereof, between words and their referents.

[W]ords do have referents, but these referents have no substance to them, being themselves merely labelled entities that depend on other entities, and so on ad infinitum. Every entity depends on other entities in a giant web where the only reality is the interrelatedness of the entities. There is no real substratum to this universe, and the only existence that things can be said to have is a very weak, conventional one that is reflected in the patterns of interconnection, that is, in the usage of language (Cabezón 1994, p. 163)

To effectively combat the reificatory tendencies of language, Madyamaka does not – as the mainstream of Buddhist thought – accept that designata of linguistic expressions are conventional in the sense of being derived from agreed-on fictions, i.e. universals arrived at through the process of exclusion (apoha), as claimed by Dignāga and Dharmakīrti. Instead, according to Madhyamikas, the only kind of existence that designata of linguistic expressions can enjoy is nominal existence, conditioned by how these expressions are used in relation to other expressions. On this view, universals are flatus voci – a convenient description of uses to which we put general terms. The idea of nominal existence (vyavahārasatya) allows the practitioner to take the middle way (madhyama pratipadā) between the extreme of existence (sat), whereby the workings of kalpanā are taken as real, and the extreme of non-existence (āsat), which denies that there is any reality at all.

On the Madhyamaka view, all phenomena and all ascriptions of them are ultimately empty (śūnya)Footnote 13: they lack an inherent essence that could uphold their being independent from something else. Everything owes its existence to something else in an interconnected web of “dependent arising” (pratītyasamutpāda; Garfield 1995, p. 168, 176). It is important to not mistake the rejection of substance for a veiled form of essentialism where the negation inherits its conceptual structure from that which is negated. Instead, Madhymaka persistently insists on the emptiness of emptiness. This is where the resort to the middle way (madhyama pratipadā) becomes a solution to the perpetually interlocked extremes of existence and non-existence. In this way, Madhyamaka also aims to avoid the nihilistic thesis that there is nothing but emptiness and in doing so appeals to the dialectic of two truths: relative truth (saṁvṛitisatya) and ultimate truth (paramārthasatya). Similarly to its treatment of emptiness, Madhyamikas take a well-established but not focal idea in Buddhist doctrine and give it a very emphatic interpretation (Hopkins 1996, pp. 400–401). As argued by Nāgārjuna in Mūlamadhyamakakārikā (1977) and later Candrakīrti in Madhyamakāvatāra (1956), relative truth refers to an unanalytical, naïve attitude towards reality based on linguistic conventions. On analyzing this attitude, one is forced to conclude that each aspect of reality is empty – this realization leads to the experience of ultimate truth (Williams 1989). Although relative truth is fallacious, ultimate truth is impossible without it, as it is arrived at only by a thorough investigation of relative phenomena. The correct Buddhist view (sammāditthi) should then assume a union of both truth, or a union of “form and emptiness” to use Buddha’s dictum from Prajñāpāramitāhṛdayasūtra (Żywiczyński 2004, p. 162). Accordingly, from the Madhyamaka perspective, the philosophical task lies in recognising the (ultimately) radical emptiness of all phenomena without losing sight of their conventional mode of existence.

3 Saying Something About the Unsayable

Contemplation is surrounded by problems and paradoxes borne from its reliance on the ultimately limited capacities of language (Priest 2002). What this shortcoming might entail has been widely discussed by influential thinkers in both the Eastern and Western tradition (as well as in the scholarly literature comparing the two traditions, e.g. Cabezón 1994; Garfield 2015). The topic of the limits of language is far too extensive to fully do justice within the limited frames of this paper, but an overview of Buddhist philosophy of language would however remain partial without at least acknowledging how the characterisation of language is tightly integrated with its limits. Just as mainstream Buddhism and Madhymaka have different views on language, their respective account of ineffability displays corresponding differences. For the former, language operates on fictive properties, which as described in Sect. 2.4 translate into an indirect and distorted picture of reality. In the Madhyamaka school, by contrast, the reificatory tendencies of linguistic conceptualisation make it incapable to ultimately ascertain the empty and dependent nature of all phenomena. To further characterise these two positions on ineffability, we compare them to Wittgenstein’s struggles with the limitations and possibilities of language. This is accomplished by comparing the mainstream view with Wittgenstein’s logically oriented work Tractatus (Wittgenstein 1961 [1921]), whereas – as we argue – the Madhyamaka position can more fruitfully be compared to his later works, most importantly Philosophical Investigations (Wittgenstein 1958). It must however be kept in mind that these comparisons remain limited, since the Buddhist take on ineffability is, as we saw in the previous section, inscribed in the very fabric of language. By contrast, Wittgenstein’s struggle – as most Western treatments – comes from recognising that there is a limit to what language can express (see Gäb 2020 for a detailed conceptual analysis of ineffability).

With that said, Wittgenstein’s approach has on numerous occasions been compared to Buddhist thought (e.g. Gudmunsen 1977; Read 2009). One main reason for the popularity of such comparisons is the way in which Wittgenstein attempts to arm philosophy in a war against itself. As Wittgenstein sees it, one of the chief aims for engaging in philosophical activity is to show how this very activity gives rise to pseudo-problems. His philosophical project thereby has the therapeutic aim of dissolving problems rather than solving them (Wittgenstein 1958, p. 51). It lets one overcome the impression that philosophy is dealing with problems that have solutions in the first place (Read 2009, p. 15), and it is in this regard that the Buddhist and Wittgenstein’s ideas of philosophy appear strikingly similar, as can be seen in the following quote:

Most of the propositions and questions to be found in philosophy are not false but nonsensical. Consequently, we cannot give any answers to questions of this kind, but can only establish that they are nonsensical. Most of the propositions and questions of language arise from our failure to understand the logic of our language. […] And it is not surprising that the deepest problems are in fact not problems at all. (Wittgenstein 1961 [1921], p. 37)

In considering philosophical problems as nonsensical, Wittgenstein engages in a reflection on the prospects of philosophical enterprise to state anything valuable or substantial at all. One should be cautious to thereby infer that philosophy is a meaningless pursuit; rather, “deeply personal and genuinely tormenting problems” (Hudson 1973, p. 479) are meant to be overcome by attending properly to the underlying form and function of language. In Tracatus, Wittgenstein endorses (just as Buddhist philosophers like Dignāga and Dharmakīrti; cf. Section 2.4), a correspondence theory of truth, but gives it a logical-atomistic interpretation where propositions provide composite “pictures” corresponding to (similarly compositely constructed) “facts”; or perhaps more accurately, they provide depictions of possible states of affairs that might be true or false. Formally, for these depictions to correspond to facts, they have to reflect or retain the logical structure of the facts they are representing. Since the very saying of anything meaningful is seen as dependent on the isomorphism between facts and propositions, Wittgenstein argues that there is no way to say (or rather, no way to depict) the structures that make world and logic resonate with one another. The ability of the latter to depict the former cannot be stated; it can only be shown (Wittgenstein 1961 [1921], p. 51). There is no way to account for how language can provide a corresponding depiction of reality, which is something that just has to be taken for granted by a correspondence theory.

The distinction between saying and showing is what ultimately leads Wittgenstein to insist that language is incapable to in a truthful manner articulate its relation to the world. It is this that eventually leads to the famous final seventh proposition of Tractatus, which laconically states that “[w]hat we cannot speak about we must pass over in silence” (Wittgenstein 1961 [1921], p. 151). In a remarkably performative moment for a work devoted to articulating the logical structure common to world and language, the last proposition differs from the structure of the preceding six by having no elaborations in sub-propositions. When nothing more can be said, Tractatus thus performatively resorts to keeping silent. A similar sentiment is echoed in the Buddhist saying that after his enlightenment, Siddhartha Gautama did not utter a single word (Cabezón 1994, p. 171; cf. D’Amato 2008 for an interpretation of this claim). When there are things that cannot even be articulated into meaningful statements, relinquishing the desire to formulate a problem with the impoverished tools provided by language becomes a kind of acceptance. If one were to phrase it in Buddhist terminology, the solution is perhaps to accept that there are things that cannot be said and believing otherwise is bound to lead to perpetual dissatisfaction.

Similar to the mainstream Buddhist position represented by Dignāga and Dharmakīrti, Wittgenstein states that there is a fundamental and principal limitation to language. Both Buddhist thinkers and Wittgenstein reach this conclusion by approaching linguistic expressions as something that aims to correspond to an external reality. To what extent language is successful or not would then depend on its ability to represent reality – conceptually mediated or not (see Sect. 2.4). In Buddhism, the referential function of language fails due to the unbridgeable gap between the reified generally applicable forms of language and the ultimate momentary nature of reality. The problem emphasised by Wittgenstein is quite different: it concerns the ineffable character of the correspondence between language and world that the logical structure is incapable of articulating. While they thereby differ in what way language fails, they agree to move from a linguistically facilitated mode of contemplation to meditative silence. When one has climbed to the top, it is time to throw away the ladder – as the famous parable from Tractatus goes (see Read 2009 for a comparison of the similarity between Wittgenstein’s discarded ladder and the Buddhist notion of ascending levels of analysis discussed in Sect. 2.3).

Linguistically facilitated conceptual thinking is needed to go beyond it, but this does not tell us at what moment speech ceases to signify and when silence is the preferred response. Tractatus provides us with no clues on how to overcome the limitations of language and when to refrain from speaking. While a Buddhist solation also recommends resort to a silence beyond the referential mode of language (i.e. meditative insight; yogipratyakṣa), it has been recognised that this solution introduces new layers of problems – especially so since the limitation of language is linguistically expressed (sometimes called “the inexpressibility paradox”, see Garfield and Priest 2003). It is here that the soteriological aims and practical applications intersect with philosophy. Numerous sūtrasFootnote 14 and meditative techniques (samādhi)Footnote 15 all aim to find a non-linguistic form for overcoming language (Garfield 2015, p. 254). One such practice is the recitation of mantras, where the repeated use of the same sound fills the linguistic channel with something that somehow can keep the mind from falling back into its ultimately feeble linguistic-conceptual activity.

3.1 From Propositions to Use

Silence becomes a revolt against the shortcoming of language by harbouring no propositional content; it makes no claims that are ultimately futile. Despite having a lack of propositional meaning, the silence at the end of Tractatus is clearly not just the negation of speech. Situated within a broader discursive matrix, to refrain from speaking becomes something else than mere silence. In the context of Tractatus, silence is a vehicle for indicating the inadequacy of using language. Responding to the unsayable with silence is not sign-for, but a sign-of this very inexpressibility. This paradoxical mutual interplay between silence and language has been thematised by Buddhist philosophers like the Chinese Madhyamaka scholar Jizang (sixth century CE), who argues that resorting to silence does not solve the riddle of ineffability since silence becomes “articulate precisely because of its place in the larger discourse” (Garfield 2015, p. 256). Garfield (2015) goes on to note that we are now back at a paradox, which when philosophically thematised involves a turn away from the meaning of speech to the silence and absence that it presupposes.Footnote 16 This recognition opens up for a different perspective on language that do not privilege its referential and representational mode.

Such a view is found in the Madhyamaka deconstruction of core assumptions and arguments made by the denotationist and connotantionist models (see Sect. 2.4). Candrakīrti (1979) argues that linguistic items isolated from a specific social context do not have any referential function, hence he rejects denotationism. At the same time, the vehemently sceptic attitude towards anything resembling substantialism makes Candrakīrti reject the connotationist view of universals as objects of linguistic expressions (for a detailed summary, see Żywiczyński 2004, p. 173f). To ascribe linguistic entities with sense independent from their application in specific contexts will lead to their reification as something that by its own nature has meaning, which must be rejected in the light of the Madhyamaka doctrine of emptiness (Dreyfus 209–210). Since language is ultimately – just as everything else – empty, it is utterly incapable of even failing to refer to reality. Based on the two truths doctrine (see Sect. 2.5.), one should nevertheless still recognise the relative truth of language as functionally operational within these limits. This underpins a view of language as a practically applicable instrument without substance (svabhāva) – either in reference to something else or in the words themselves. Such a bifurcated concept of truth serves the therapeutic function of overcoming the ontologising tendency of conceptual thinking anchored in language. Through this shift in attitude towards language, one can then appreciate that there is no existence apart from conventional dependent existence.

Similar to Madhyamaka, the later Wittgenstein insits that language must be viewed against the background of its practically oriented use, where one and the same same expression can serve radically different social functions in different contexts (Wittgenstein 1958, p. 11f; for a deepened discussion on the similarities between Wittgenstein and Madhymaka, see Gudmunsen 1977; Hudson 1973). This variable use is not regulated by an overall meaning; instead, to the extent that there is a meaning it resides exactly in the variable applications across different contexts and situations (Wittgenstein 1958, p. 20, 31f; cf. the Madhyamaka notion of askepa). The pre-theoretical knowledge of knowing how to use language correctly is epistemologically prior, presupposed and – perhaps for those very reasons – forgotten in philosophical reflection (Wittgenstein 1958, p. 48). When language is viewed in this practically oriented manner, the issue of whether it can refer to reality or not can be dissolved.

Even though there are decisive parallels between the later Wittgenstein and Madhyamaka, they account for propositional use in quite different ways. The former relegates the referential form to a specific but not privileged type of “language game”. As we discussed in Sect. 2.5, the latter instead relies on the doctrine of two truths that do not, to our knowledge, have any counterpart in Western philosophy. What this would mean is that at the relative level language works – more or less – as we expect it to do. There are words and expressions that let us do things in practical situations, including referring to things. Upon closer inspection, one will however realise that there is nothing else behind this naïve attitude. Nothing is upheld by any substance, reason or motive, but is empty. Ultimately, linguistic expressions are also empty and therefore cannot refer or strictly speaking accomplish anything else. Being the middle way to enlightenment, it would however remain fundamentally incomplete as long as one has not returned to the everyday world, only now with the fundamental insight that whatever exists does so only in the conventional manner of lacking independent existence (Garfield 2015, p. 261f). When applied to language, we would thus recognise that it does not have the capacity to operate at the level of ultimate truth, but as relative truth it can still serve a multitude of instrumental purposes.

The transition to a social view of language attempts to overcome ineffability by a thorough reinterpretation of language that would free it from its ontologising tendencies. The prospects for being successful in this regard are not just philosophical for either Wittgenstein or Madhyamaka, but can rather be sought in the therapeutic goal of relinquishing the desire for language to express substantially meaningful propositions. Such an approach does however introduce new problems, like the two truths doctrine paradoxically operating as an ultimate truth (as argued by Jizang, cf. Garfield 2015). Similarly, the socially established rules that Wittgenstein identifies as a key trait of language introduces the possibility of scepticism or relativism, since “any course of action can be made to accord with the rule” (Wittgenstein 1958, p. 81).Footnote 17 While these issues fall beyond the scope of this paper, they nevertheless indicate that a non-propositional view of language will also encounter its own limits.

4 Conclusion

This paper offered an overview of Buddhist philosophy of language. It is, as all intellectual traditions, unique in that it grows out from particular sentiments and issues. With its explicitly soteriological goal, Buddhist philosophy makes the path to enlightenment primary, which is reflected in how the reificatory tendencies of language and conceptual thinking impedes this process. Through these categories, language apprehends and connects to reality, albeit in a distorted way. Various doctrines such as apoha or sphoṭa found in the mainstream of Buddhist philosophy try to account for the limited validity of language in ways that neither contradict the basic ontological and epistemological assumptions nor render the soteriolgical goal impossible to reach. In order to get a more truthful grip on reality, one must use conceptual thinking and language, which in a Buddhist context, means a paradoxical reliance on language. In doing so, one is imposing a border between known/unknown, sayable/unsayable, etc. However, in pointing to this limit, one is also saying something – albeit in a negative way – about that which nothing by definition could be said (Priest 2002; Garfield and Priest 2003). As we saw in Sect. 3, this is why Buddhist reflection ultimately wants to transgress language where reflective silence becomes an appropriate response. Ultimately, when silence becomes significative it does however introduce a whole new layer of paradoxes.

Ineffability would then mean that language is ill equipped for articulating certain experiences. In mainstream Buddhist philosophy, this shortcoming follows from the view on language as contributing to the reificatory tendency of creating fictive universals. While this specific articulation is derived from momentariness and the illusion of permanence in Buddhist ontology, it also captures the generally recognised property of language that linguistic forms can be reused at another time for saying something else about a completely different situation. This stands in contrast to a truly authentic event, which, as argued by for instance Badiou (2005 [1988]), is in the broadest sense inimitable and non-repeatable. In On truth and lies in a non-moral sense, Nietzsche expresses this in a way that echoes the Buddhist views of language: “[a] word becomes a concept insofar as it simultaneously has to fit countless more or less similar cases – which means, purely and simply, cases which are never equal and thus altogether unequal. Every concept arises from the equation of unequal things”. The general concepts of language unite fundamentally different phenomena and thereby miss their individuality and particularity, which, in the case of Buddhism, is immediately connected to the momentary (or, in the case of Madhyamaka, ultimately empty) character of all phenomena. This emphasis on linguistic expressions as always inadequate separates Buddhist philosophy from most Western thinking on the ontology of language. Thus, even where similarities are detected, comparison between traditions must always carefully respect the specific context, heritage and problems that a particular tradition is dealing with. While we have pointed to numerous places where Buddhist philosophy can be compared to Western discussions, like universals, sense/reference and limits of language, we also noted several ideas and concepts specific to Buddhism. Among these are the doctrine of two truths and apoha according to which conceptual thinking is based on processes of exclusion. One possible venue for future work would be to carefully scrutinise the underlying premises behind such unique philosophical concepts and outline their contribution in the formation of Buddhist philosophy of language.

Even though Buddhist philosophy rejects that language is capable of ever getting to the particular, it is also noted that there is a general applicability that lets language – however distorted it may be – turn anything into a theme and topic for linguistic elaboration. One can at least try to say something about inexpressible experiences, like the immediate and complex feelings from hearing a certain piece of music. The opposite does however not seem to be true: it is not apparent that a piece of music can be about a dialogic exchange of ideas. One could arguably even deduce the oft-noted peculiar self-reflexivity of language from this more general observation: if language can be about anything, it would follow that it must be able to be about itself (an observation also made by Hjelmslev 1961, p. 109). By extension, this ability of language to turn anything into its theme might also be why it can attempt to transgress the borders of what can be expressed.

The Buddhist approach to language is painfully aware of this gap between the unique and the repeatable. The ultimately real ceases to be within the span of a moment or, in the case of Madhyamaka, amounts to accepting that everything that exists is empty and dependent on something else. Either way, this makes permanence a distorted illusion due to the reificatory tendencies of mind and language. At the same time, it acknowledges that the ability of language to – albeit in an inaccurate fashion – connect to reality. It is because of this that contemplation is possible in the first place. To see language as both liberation and bondage then becomes a necessary step on the path to enlightenment.