Language and Extra-linguistic Reality in Bhartṛhari’s Vākyapadīya
- 91 Downloads
Relation between language and extra-linguistic reality is an important problem of Bhartṛhari’s linguistic philosophy. In the ‘Vākyapadīya,’ this problem is discussed several times, but in accordance with the general perspectivist trend of Bhartṛhari’s philosophy each time it is framed through different concepts and different solutions are provided. In this essay, an attempt is undertaken to summarize the variety of different and mutually exclusive views on language and extra-linguistic reality in VP and to formulate the hidden presuppositions on which the actual viewpoints expressed in the kārikās are based. As a result, the following approaches are formulated: (A1) Language is coextensive with external reality. (A2a) Language, designated as kalpanā/vikalpa, is distinct from reality. (A2b) Language refers to the secondary/metaphorical reality (upacārasattā/aupacārikī sattā). (A3) Language and reality somehow correlate, because otherwise, practical/linguistic activity (vyavahāra) would be impossible. The origin of these approaches and their affinities with different schools of Indian philosophy (Nyāya-Vaiśeṣika and Buddhist Pramāṇavāda) are examined. Approach (A3), according to which correlation between language and reality is functional and not ontological, seems very close to Dharmakīrti’s concept of arthakriyā. This approach accords with Bhartṛhari’s perspectivist philosophical strategy. It enabled him to explain how effective linguistic activity is possible, capturing language in its dynamic aspect, without limitative static ontological constructions.
KeywordsBhartṛhari Vākyapadīya Indian linguistic philosophy Arthakriyā Language and extra-linguistic reality
Any study of Indian philosophy, the obvious challenge of understanding philosophical ideas aside, is inevitably accompanied with solving hermeneutical problems. There is a constant search for the exact counterparts for Sanskrit terms and notions, as well as for the general conceptual means appropriate for expressing Indian doctrines in terms of Western philosophy. The problem of language and extra-linguistic reality in Bhartṛhari’s ‘Vākyapadīya’ undoubtedly gives rise to this kind of complexities, which means that this problem cannot be fully solved without paying close attention to the conceptual means applied and the terms in which the problem is formulated.
Bhartṛhari (~ 450–510 CE) was a prominent proponent of Indian linguistic philosophy. His main opus was the ‘Vākyapadīya’ (VP), an extensive work composed in verses (kārikās)—a sort of compendium of different philosophical views and doctrines. Being a grammarian, Bhartṛhari discussed ontological, epistemological, and semantic problems underlining the central role of language in the universe and in all cognitive processes. It was a distinctive feature of Bhartṛhari’s philosophy that he was generally open to different points of view and admitted different, often mutually exclusive opinions. J. Houben designated this methodological principle as ‘perspectivism’1 stressing that Bhartṛhari usually avoided ‘attempts to establish the superiority of one view over the other, even if one would expect Bhartrhari to have a “natural” preference for one of two opposed views’ (Houben 1997, p. 341).2
The word refers to the universal, which inheres in the object.
The word refers to the substance, of which the object is made.
The word refers to the metaphorical reality (aupacārikī sattā), which is different from absolute reality.
This evident difference of interpretations implies that the problem of language and reality in Bhartṛhari’s philosophy needs a further elaboration. I generally agree with Bronkhorst that VP contains more than one solution. In this paper I provide an alternative classification that is based on the reconstruction of the presuppositions which underlie the views on language and reality expressed in the kārikās. It seems promising to develop a classification not as an enumeration of actual views mentioned in VP, but as a model that comprises different theoretical approaches towards the problem of language-reality relation. Such a classification would be more comprehensive, and it might help in demonstrating that in certain cases different views formulated in different terms rest upon similar theoretical presuppositions.
In the light of the hermeneutical challenges mentioned at the beginning of the paper, it seems evident that attention should be paid also to the very concepts of ‘language’ and ‘(extra-linguistic) reality.’ So, prior to the exposition of the classification of the views, particularly important is to determine what is meant with the terms of ‘language’ and ‘(extra-linguistic) reality’ in Indian and in contemporary Western philosophical discourse, and whether these concepts in both traditions correspond to each other.
Language and Reality: Sanskrit Terms and Their Western Counterparts
Language and reality were the topics widely discussed by different schools of Indian philosophy. The influence of grammar on other philosophical traditions, as well as the general ‘linguistic bias’ of Indian philosophy, made it inevitable that the problem of relation between language and reality was solved in a number of ways, though sometimes in an implicit manner. However, we should not forget that the terms ‘language’ and ‘reality’ belong to contemporary discourse and have no direct counterpart in Indian philosophy, and even in Western tradition, these terms are too broad and are not to be defined out of context exhaustively.
It was probably since Sophists that language (logos) was comprehended as a separate object of philosophical discourse, a distinct entity that differs in scope and extent from extra-linguistic reality. In the subsequent centuries and millennia, different approaches to the problem of language and reality can be traced in Western philosophy. A naive popular view on language as a direct counterpart to external reality (‘words as names of things’)—similar to what Saint Augustine observed in his ‘Confessions’—developed into realist schools. Later on, the emphasis on empirical reality led to critique of natural language and to positivists’ projects of formal languages. Another tendency in philosophical investigations was to consider language as the conceptual basis of worldviews. Another general view on language was functional. According to it, words do not refer to any extra-linguistic entities, be they referents or mental ideas—meanings are the usages of the words, and language is fundamentally expressive rather than descriptive.3 From this vague outline it seems quite obvious that even in Western tradition the concepts of language and reality are somewhat ‘slippery’. Using these terms in a study of Indian philosophy, one should specify their exact meaning in each case.
On the other hand, one can hardly find exact counterparts for these terms in Indian philosophy and, particularly, in VP. Sanskrit terms vāc, śabda, kalpanā, and vikalpa in some contexts do correspond broadly to the Western concept of language, but their primary meaning is different. Vāс (‘speech’) and śabda (‘word,’ ‘linguistic expression’) can be considered close to linguistic/semiotic concepts of language (cf. Matilal 1990, p. 132). Kalpanā and vikalpa4 may correspond to philosophical interpretations of language in the style of Humboldt or Heidegger. The concept of extra-linguistic reality is even more problematic. It implies that there exists a reality outside of language, though not every school of Indian philosophy would agree with this point. An object of a linguistic expression is usually rendered with the term artha that in some contexts can be translated as ‘meaning,’ ‘referent,’ or ‘intentional object’; sometimes also as an ‘extra-linguistic object’. A phenomenal object treated independently from language may be referred to as dravya (‘thing,’ ‘material object’). The latter term also refers to the substance that constitutes material things. The supreme reality that is beyond limits and conceptualization is usually called paramārtha (‘the ultimate artha’).
A point of view of a grammarian for whom everything reflected in language is believed to be existent.
A more refined view according to which only physical things and the language-principle are admitted.
The ultimate tier of ontology that admits only the language-principle (Aklujkar 1970).
In the latter case, phenomenal world appears a manifestation of the language-principle (Brahman), whereas the opposition of language and extra-linguistic reality appears irrelevant.
So, ‘the problem of language and extra-linguistic reality’ is not an exact wording in VP, but rather a conceptual outline applied for analyzing Bhartṛhari’s philosophy in terms of contemporary philosophical discourse. Consequently, language and reality are not fixed concepts—rather they can be considered a starting point for the investigation undertaken in this paper. In what follows, I will discuss different models of language-reality relation in Bhartṛhari’s philosophy, every time specifying the exact meaning of the terms and the general philosophical paradigm from which a particular model derives.
Three Approaches to the Problem of Language and Reality in VP
The problem of language and extra-linguistic reality is discussed several times in VP, but each time it is framed through different concepts, and different solutions are provided. An impossible task of solving all semantic problems in a single consistent theory was probably the reason why Bhartṛhari—in his usual perspectivist manner—developed a number of standpoints, each of them clarifying a distinct aspect of language-mind-external reality continuum. Disagreements between different views concern the nature of the meaningful linguistic unit (sound, word or sentence),5 the ‘thing-meant’ (universal, individual object, substance or mental entity), as well as the nature of language and its interrelation with cognitive processes.
The variety of views exposed in VP is based on different models of language-reality relation. Not all of these models are formulated explicitly; some of them just remain presuppositions to be reconstructed. Therefore, the scope of my paper cannot be confined with actual standpoints expressed in the kārikās or in the commentaries. On the contrary, I am going to analyze the presuppositions on which the actual viewpoints expressed in the kārikās are based and examine different views scattered throughout VP in order to determine how they correlate to each other.
(A1) Language is coextensive with external reality.
(A2a) Language, designated as kalpanā/vikalpa, is distinct from reality.
(A2b) Language refers to the secondary/metaphorical reality (upacārasattā/aupacārikī sattā).
(A3) Language and reality somehow correlate, because otherwise practical/linguistic activity (vyavahāra) would be impossible.
(A2a) and (A2b) evidently repeat solutions (3) and (4) of Bronkhorst’s classification. I bring them together as subdivisions of a single approach, as it will be argued further that both of them imply a similar model of language-reality relation. Bronkhorst’s solutions (1) and (2)—words referring to universal inherent to it, or to the substance of which the object is made—are omitted in my classification because both statements have more than one interpretation in VP. The view on universals as meaning of words can be reduced to the approach (A1)—in case universals are considered real entities; or to the approach (A2a)—in case they are believed to be of a mental nature. The same is true for the material substance. The view that all the words refer to the substance can be considered a variation of the approach (A1). But the view that distinct words refer to unreal forms of the substance (asatyopādhi)—the limiting factors of the consciousness that make us perceive the indivisible permanent dravya as different impermanent objects7—can be reduced to (A2a).
In the following sections, the approaches listed above will be discussed. Particular stress will be made on the reconstruction of the philosophical presuppositions of every approach as well as on the relation with the views on language and reality elaborated by other schools of Indian philosophy.
Language Is Coextensive with External Reality (A1)
An inquiry concerning the nature of language may start from a naive view that words are ‘names of things,’ with the realm of words and the realm of things considered to be isomorphic. From this view, a realist thinker can conclude that all referents of linguistic expressions are real. In other words, language and extra-linguistic reality are coextensive—approach (A1). In the history of Indian philosophy, this view was characteristic to such schools as Nyāya and Vaiśeṣika. There are also instances in VP where the arguments are based on a similar view; however, Bhartṛhari has never formulated this approach in explicit terms. So, prior to the discussion of the relevant textual passages in VP, let us turn to the schools that elaborated this view in more detail.
The origin of the realist view on the direct correspondence between language and extra-linguistic reality can be traced in the early Vaiśeṣika. The initial intention of the Vaiśeṣikas was to ‘calculate’ the phenomenal reality in order to represent its plurality as derived from the finite number of ‘categories’ (padārthas).8 Praśastapāda enumerated six padārthas: substance (dravya), quality (guṇa), action (karma), universal (sāmānya), particularity (viśeṣa), and inherence (samavāya). However, in the ‘Vaiśeṣikasūtras’ only the first three are explicitly designated as padārthas, whereas the latter could have been introduced later (Halbfass 1992, p. 75). The parallels between dravya, karma, and guṇa and three parts of speech—noun, verb and adjective respectively—seem quite suggestive (Bronkhorst 1996, p. 121; Bronkhorst 2011, p. 26), the more so as the same three padārthas were listed in Patañjali’s ‘Mahābhāṣya’ (Mbh 1880, 1).9 This can be considered an evidence that the schema of padārthas in the early Vaiśeṣika was stipulated linguistically, with each padārtha corresponding to a distinct part of speech.
Still, one should remember that the early grammarians did not distinguish between nouns and adjectives, as there is no morphological distinction between them in Sanskrit.10 In Pāṇini’s ‘Aṣṭādhyāyī’ (A) the terms ‘guṇavacana’ (A 4.1.44, 4.1.44, 5.1.124) and ‘viśeṣaṇa’ (A 1.2.52, 2.1.57) referred to the syntactic and semantic role of a subanta (word with a case-ending) in a sentence and not to the distinct parts of speech. Patañjali enumerated his analogues to padārthas in order to classify extra-linguistic objects/referents of words (Mbh 1880, p. 1). This means that his classification derived primarily from philosophical and not from linguistic context, as the distinction of dravya and guṇa had no direct counterparts in Pāṇinian grammar. So, the question of whether the first three padārthas of Vaiśeṣika were formulated under the influence of language or because of some logical and philosophical reasons remains open. From the point of view of Pāṇinian tradition, the one-to-one mapping between padārthas and parts of speech is not as obvious as it may seem to a modern reader.
Whatever the origin of this idea could be, parallelism between language and extra-linguistic reality was the basic idea for realist Vaiśeṣika and Nyāya. Praśastapāda stated explicitly (Pbh 1895, p. 16) that the six padārthas are existent, knowable and nameable, as they possess the attributes of being (astitva), nameability (abhidheyatva) and knowability (jñeyatva).11 Praśastapāda did not explain whether the concepts of astitva, jñeyatva, and abhidheyatva are coextensive and did not clarify the relationship between astitva and abhidheyatva, i.e., being and name-ability.12 Most probably he was not generally interested in semantic problems (Halbfass 1992, p. 158). Later on, the Naiyāyikas discussed possible implications of this statement in detail—with reference to an additional padārtha of abhāva, i.e., something nameable and knowable, but at the same time non-existent (Perrett 1999). In any case, Praśastapāda’s thesis concerns padārthas only. It does not imply that everything that is nameable exists.13 Still, there are examples of how this implication—though not stated explicitly—was implicitly accepted by the Vaiśeṣikas who wanted to justify the existence of some concepts. E.g., already in the ‘Vaiśeṣikasūtras’ (VS), the existence of such substances as time and direction is justified by means of such words as ‘earlier,’ ‘later’ (VS 2.2.6), ‘nearer,’ and ‘further’(VS 2.2.12). The existence of inherence as a separate ontological type of relation (samavāya) is proved by means of the utterance ‘iheti’ ‘[this occurs] here’ (VS 7.2.29). In these cases, the linguistic reality is implicitly considered to have the same validity as empirical reality, with words being the means to prove the existence of the referents (сf. Lysenko 2005).
These kārikās belong to the passage VP I.51–71 where the problem of the integral vs. compound nature of the word is discussed. Bhartṛhari claims here that, apart from the thing-meant, the word refers to its own form (śabdasvarūpa), which thus becomes a distinct signified. That is to say, each word is a signifier of itself—of an integral word that has a signifier and a signified aspects. Most probably, this concept was introduced in order to explain from a psychological point of view the way in which understanding of a verbal utterance is accomplished—so that before understanding the actual meaning of the uttered words, one has to recognize them as words in general (VP I.54, 57–58).
Before connection with the signified, a signifier has [the own form] as the meaning and [thus] accounts for the use of the sixth and the first ending.
The first ending is prescribed after the signifier because of its meaningfulness. [The words] ‘[this is the signifier] of this’ express the difference [of the signifier] from its referent.14
Along with functional/psychological justifications of śabdasvarūpa, Bhartṛhari additionally justifies it in VP I.67–68, relying on the verbal utterance ‘asyeti.’ In this utterance, the genitive form ‘asya [śabdasya]’ refers to the word that has its own form as the meaning. It should be noted that these kārikās justify svarūpa on a functional/psychological level and not on the ontological—as it could have been in the case of Nyāya/Vaiśeṣika. The same observation can be made with respect to VP II.119 interpreted by Bronkhorst as another example of this approach.15 According to this kārikā, the words that denote supernatural concepts, e.g., apūrva ‘remote/invisible concequence,’ devatā ‘god,’ and svarga ‘heaven,’ refer to their objects in the same way as words that refer to ordinary things like cows. Still, it does not follow from the context that this kārikā refers to the ontological status of the referent and not to a psychological process of understanding of a linguistic utterance, with artha standing for ‘external object’ and not for ‘meaning’.16
In any case, it is evident that for Bhartṛhari, the way of reasoning based on the coextensiveness of language and reality was not the predominant one. Especially since Bhartŗhari, contrary to the Vaiśeṣikas, did not appeal to words in order to justify philosophical categories. His use of the approach (A1) in the instances discussed above had an intrinsic connection with linguistic problems.
Language, Designated as kalpanā/vikalpa, Is Distinct from Reality (A2a)
The second approach of my classification is based on a sharp distinction between language and reality. Reality in this case is understood as ultimate or transcendental, whereas language is considered a means responsible for creating a false or irrelevant worldview. The parallelism between language and extra-linguistic reality is negated here. Language can be characterized as kalpanā/vikalpa, generally meaning ‘conceptual construction.’17 Both terms were used before Bhartṛhari, so prior to discussing vikalpa/kalpana in VP, let us follow their history.
Vikalpa was defined in the ‘Yoga-sūtras’ I.9 as an invalid knowledge that follows words and has no (real) object.18 In Buddhist epistemology, perception (pratyakṣa) was radically distinguished from language: perceptual experience was believed to be independent of any conceptual activity. The realm of language and thought was looked upon as integral and was referred to as vikalpa/kalpanā. This attitude can be traced in the early Yogācāra texts (such as the ‘Madhyāntavibhāga’).19 Later on, it was elaborated by the Buddhist logic school of Pramāṇavāda. Dignāga (~VI CE) glossed kalpanā as ‘association of a name, genus etc. [with a thing perceived].’20 Dignāga’s follower Dharmakīrti (VII CE) claimed in the ‘Pramāṇavārttika’ (PV II.2) that the validity of a linguistic utterance is conditioned by the speaker’s intention and does not depend on the reality of the object. Accordingly, the reality of referents of the words differs from the reality of empirical objects.
It is well known that Bhartṛhari was close to the Buddhists and significantly influenced different Buddhist traditions, from Dignāga and his followers to Jñānaśrībhadra, the eleventh century commentator on the ‘Laṅkāvatārasūtra.’21 The critique of language that Bhartṛhari exposed in the ‘Saṃbandhasamuddeśa’ (‘Chapter on relation’ of VP) was literally borrowed by Dignāga in his ‘Traikālyaparīkṣā’ (‘A critical investigation of three times [past, present and future]’). More problematic is the question of the particular Buddhist texts and thinkers that were familiar to Bhartṛhari. There are reasons to believe that he could have been influenced by some works of Vasubandhu and Nāgārjuna (Lindtner 1993; Houben 1995, pp. 53–58). But as in the case of Praśastapāda,22 apparent parallels can be drawn only with posterior Buddhist works, in the first place with those of the Buddhist Pramāṇavādins (Dignāga and Dharmakīrti).
In the second part of the ‘Saṃbandhasamuddeśa’ (VP III.3.52–88) Bhartṛhari criticized the general idea that there exist different objects (arthas), to which words may refer. From this idea follows the view on semantic relation as the correspondence between words and their referents that was also criticized in VP. VP III.3.54 claims that ‘a word is based on a cognition which does not reflect the object in its entirety’.23 Language behavior is based on untrue divisions of reality (VP III.3.73). Those who see reality as it is perceive the true artha (i.e., Brahman) as free from the division into the subject (draṣṭā), object (dṛśya), and the process of perception (darśaṇa) (VP III.3.72). Thus, that what is usually called language is nothing more than vikalpa. It does not express the true reality, but correlates with its untrue subdivisions into multiple objects.
Bhartṛhari’s concept of vikalpa/kalpanā is close to that of the Buddhists and probably emerged under their influence.24 At the same time, the view on reality as the single partless artha evidently contradicts the Buddhist doctrine of kṣaṇikavāda that considers only the momentary to be real. Bhartṛhari’s ontology can be generally characterized as holistic, whereas the Buddhists generally tend to atomistic/discrete ontological solutions.25 In the light of this general difference between Bhartṛhari’s and the Buddhists’ ontologies, all the more striking is the methodological proximity in their attitudes towards language and its relation with reality and cognitive processes.26 In both cases, language is considered not as the counterpart of external reality. On the contrary, language is believed to be a part and parcel of cognitive processes. This can be explained with the general shift from ontology and objects of cognition (prameya) to epistemology and modes of cognition (pramāṇa) in Indian philosophy of the middle of the first millennium CE, which occurred in the first place in the works of the Buddhist philosophers and Bhartṛhari (cf. Franco 2013; Desnitskaya 2016). Considering language as vikalpa/kalpanā does not necessarily imply that empirical objects are false or non-existent. Rather, this indicates that a study of semantic problems operates with entities of different kind which cannot to be described satisfactory in terms of a realist ontology.
Language Refers to the Secondary/Metaphorical Reality (upacārasattā/aupacārikī sattā) (A2b)
As a crystal placed near colored objects acquires their hue, similarly a word, fixed in this secondary mode of being, becomes related with different contradicting and non-contradicting properties (VP III. 3.40–41). Thus, the reality of referents is believed to be metaphorical or subordinate to empirical objects. The concept of upacārasattā helps solving a number of semantic problems widely discussed by Indian philosophers. Indeed, if it is presupposed that every word refers to a certain extra-linguistic object, such utterances as ‘hair’s horns’ or ‘Devadatta makes a pot’ become problematic, as in reality hairs have no horns, and a pot referred to in the sentence does not exist yet. But if it is agreed that words refer to metaphorical reality, semantic paradoxes of these types would be escaped (cf. VP III.3.42–48).
The objects of words, while being designated, [acquire] a distinct, secondary mode of being, that shows the individual form of all objects in all states.27
The primary meaning of upacāra is ‘metaphor,’ ‘secondary/figurative meaning.’28 For the first time met with in Mbh (Mbh 1880, p. 38), this term has been extensively used in different intellectual traditions—in semantic, epistemic, and hermeneutical contexts.29 Especially important it was for Buddhist philosophers of the Yogācara school who also criticized realist theories of meaning and assumed that referents of words are independent of extra-linguistic phenomena. Despite the evident proximity of the attitudes, the Buddhists’ account of upacāra differs from Bhartṛhari’s, as Buddhist thinkers introduced upacāra in different methodological perspectives and elaborated this concept basing on different philosophical implications. Asaṅga underlined the performative function of upacāra; Vasubandhu employed it mainly in a hermeneutical context; Sthiramati described the figurative referential relation in terms of causality (Tzohar 2011, Tzohar 2017).
The presuppositions on which Bhartṛhari’s concept of upacārasattā is based are very close to those of kalpanā.30 Upacārasattā and kalpanā (1) both are understood as linguistic reality that is different from empirical reality and (2) both are considered as somewhat subordinate to the mukhyasattā, the true ontological being. The concepts also seem proximate, as they often are described in similar terms—upacāra’s interconnection with mental processes is underlined. Vasubandhu used the term ‘upacāra,’ characterizing it as transformation of consciousness (vijñānapariṇāma) (Houben 1995, p. 258). Helārāja also referred to upacāra as mental being (buddhisattā) (VP, Kā III 1963: 153) or produced by buddhi (VP, Kā III 1963: 159). He characterized buddhi with regard to upacāra as related to vikalpa (vaikalpikī) (VP, Kā III 1963: 150).
So, the concepts of upacārasattā and vikalpa are evidently related. But there is also a difference between them that most probably derives from the difference in the philosophical contexts in which each of the concepts is introduced. The concept of vikalpa/kalpanā underlines the separateness of language from external reality in order to overcome language, get rid of it. It is estimated negatively as something inferior to the true reality, no matter in which terms this true reality is defined. Upacārasattā, on the other hand, is considered a positive entity. It is introduced in VP in order to demonstrate that referents of linguistic utterances somehow exist, though their mode of being differs from the mode of being of empirical objects. The concept of upacārasattā helps to solve some philosophical problems and semantic paradoxes. But it is mentioned in the kārikās only once, so one can hardly consider it Bhartṛhari’s main view on language and reality.31
Language and Reality Somehow Correlate, Because Otherwise Practical/Linguistic Activity (vyavahāra) Would Be Impossible (A3)
There is yet another solution met with throughout VP, though mainly as a presupposition to be reconstructed. According to it, there must be a certain parallelism between language and external reality because otherwise ordinary practical/linguistic activities (vyavahāra) would be impossible. Contrary to the approach (A1), Bhartṛhari does not rely here on a verbal utterance that might correspond to the real state of affairs. He simply appeals to the fact that some activity can be effective, and this would be unexplainable, unless there exists a relation between words and external referents. So, this approach considers the correlation of language and reality to be functional and not ontological.
The term yogyatā, translated above as ‘correlation,’ refers to relation with an undetermined ontological status.33 This understanding of yogyatā accords with the concept of the functional permanence (vyavahāranityatā) that was introduced by Bhartṛhari in the ‘Mahābhāṣyadīpikā’ (a commentary on Patañjali’s ‘Mahābhāṣya’) as opposed to the ontological permanence (paramārthanityatā) (MbhD 1987, p. 18).34
As there is the beginningless correlation between indriyas and their objects, the same beginningless correlation is the relation between words and objects.32
In different parts of VP, we can also find other examples of this approach, e.g., VP III.1.11 claims that the referents of words (in their being arthajāti or śabdajāti) are determined by usage.36 VP III.1.95 says that the existence of universals should be admitted because otherwise language activity would be impossible.37 In VP I.44–70, Bhartṛhari discussed different views on the structure of the word—Houben summarized them as two models: the ‘two-level model’ and the ‘two-capacity model’ (Houben 1995, p. 69–75). The functional approach to these models is expressed in VP I. 59:
[Only] if there is [semantic] relation, a word can stipulate the understanding of the meaning. Therefore it is understood that relation exists.35
Two aspects of a word, distinguished artificially and perceived as separate, stipulate different activities (grammatical procedures) without contradiction.38
Different views may correspond to different kinds of activity. On the other hand, activity can be explained from different attitudes by means of different doctrines (cf. VP I.75).
Though the approach (A3) has not been formulated as a general philosophical strategy, it seems very intrinsic to Bhartṛhari’s philosophy because it generally accords with his perspectivist tendency. This approach enables to explain how effective linguistic activity is possible, capturing language in its dynamic aspect and stressing its role in cognitive processes. Proponents of the old realist schools (such as Vaiśeṣika and Saṃkhya) strived to discover the permanent principles of the universe. They based their ontological systems on concepts with fixed and unambiguous meaning (universals, particles, categories of different kind). It was an implicit defect of these philosophical systems that they ignored the dynamic nature of the universe and especially the dynamism of language activity and cognitive processes. Bhartṛhari’s approach (A3) is fundamentally different, as it rejects static ontological constructions, and considers language the basis of cognition and, consequently, the basis of philosophical worldview. Language is understood primarily as linguistic activity, whereas reality is described in functional, not in ontological terms. Ontological dimension of reality is left to another mode of description designated paramārtha (highest reality), the opposition of vyavahāra and paramārtha being evidently Buddhist-inspired.39
Vyavahāra and arthakriyā
There is one more parallel between Bhartṛhari and the Buddhists. The functional approach (A3) brings to mind Dharmakīrti’s concept of arthakriyā as a criterion of the validity of a cognitive act. At the same time, these concepts are not entirely synonymous, as they were introduced in different philosophical context and with different purposes. So, the concept of arthakriyā needs to be considered in detail.
Padārthasya kriyā ‘the action of a thing, its causal efficiency’;
Arthavatī = prayojanabhūtā kriyā ‘action that serves a purpose, purposive or useful action.’
The second meaning is doubtlessly older, being introduced in the context of Dharmakīrti’s epistemological investigations (Nagatomi 1967, pp. 53–55). Dharmakīrti appealed to arthakriyā, in order to justify the validity of inference (anumāna), the second of the two pramāṇas (means of valid knowledge) in Buddhist epistemology. Inference, contrary to perception (pratyakṣa), has no direct access to the real particular object (svalakṣaṇa); therefore, it can be considered erroneous (bhrānta). Still, it is a valid cognitive act, insofar as it leads to the fulfillment of the purpose (arthakriyā), i.e., is pragmatically successful. Verbal knowledge is regarded a kind of inference, so its validity is also based on arthakriyā. The Buddhists consider semantic relation as purely conventional and language as belonging to the realm of kalpanā.
The interpretation of arthakriyā as arthavatī kriyā is primarily epistemological, with artha understood as the aim/object determined by the cognizer who is engaged in practical activity (Nagatomi 1967, p. 63). Still, Dharmakīrti evidently was not a pure pragmatist (cf. Dreifus 1995). The two aspects of arthakriyā, distinguished by Nagatomi, are inseparably interconnected. In arthakriyā understood as causal efficiency, artha means primarily ‘object.’ The ultimately real object of perception is the momentary svalakṣaṇa which is characterized by the causal power (arthakriyāśakti/arthakriyāsāmartyā) (NB 1.15).40 Most probably this interpretation of arthakriyā was introduced in order to harmonize epistemology with traditional Buddhist ontology, i.e., the concept of momentariness (Nagatomi 1967, pp. 64–69).
Here, vyavahāra seems similar to arthakriyā in its epistemological interpretation (arthavatī = prayojanabhūtā kriyā), the more so as Dharmakīrti amidst the vast scope of general meanings of vyavahāra, used this word mainly in the sense of linguistic communication and purposeful action (Dunne 2004, p. 257), which probably was not without Bhartṛhari’s influence. Lindtner puts it even broader, claiming that the very concept of arthakriyā appeared under the influence of Bhartṛhari: ‘When Dharmakīrti defines reliable knowledge in terms of arthakriyā he certainly has Bhartṛhari in mind, even though the term can be traced back as far as Nāgārjuna’ (Lindtner 1993, p. 204).44
Bhartṛhari’s vyavahāra is devoid of ontological interpretation that arthakriyā possesses. Bhartṛhari, with his holistic worldview, was not generally interested in the problems of causality, especially to the extent that Dharmakīrti was. Bhartṛhari also was not a pure pragmatist, as he possessed an ontological worldview. For him, pragmatism was one of the possible approaches to be applied in his perspectivist study.
Bhartṛhari and Dharmakīrti were evidently interested in different matters. Language and linguistic activity were the central point of Bhartṛhari’s inquiry, whereas for Dharmakīrti, it was a secondary issue in his study of pramāṇas. Similarly, there is a difference between Bhartṛhari’s attitude (A3) and Dharmakīrti’s concept of arthakrīya. Bhartṛhari relied on vyavahāra, in order to demonstrate that every philosophical doctrine is inextricably connected with its extra-linguistic context and is justified only in its pragmatic terms. Dharmakīrti’s arthakriyā is primarily an epistemological concept. Still the methodological proximity of their solutions is quite obvious. The common view on language as vikalpa, shared by both thinkers, is supplemented by the doctrine of pragmatic activity which justifies the validity of linguistic utterances.
So, Bhartṛhari’s vyavahāra and Dharmakīrti’s arthakriyā differ in their scope, ontological background, and the role in the philosophical traditions they belonged to. Still they are similar in a methodological perspective, as both describe epistemological processes in functional terms.
In accordance with his perspectivist attitude, Bhartṛhari obviously tended to present the full range of views on language and extra-linguistic reality. So, it is hard to find the only true solution of the problem of language and extra-linguistic reality in VP, especially because some of the theoretical solutions are not explicitly expressed and can be only reconstructed as presuppositions of some other doctrines. Therefore, in this paper, I concentrated on typological structure of these reconstructed solutions. Different solutions, at first glance seeming mutually exclusive, can be typologically close to each other (as in cases with kalpanā/vikalpa (A2a) and upacārasattā (A2b)).
Some solutions indeed resemble the ideas of other philosophical schools (Nyāya/Vaiśeṣika and Buddhist Pramāṇavāda), but this proximity should not be overestimated. Bhartṛhari’s general position was meta-philosophical in the sense that, contrary to other schools, he did not identify himself with some single doctrine but regarded each of the doctrines as possible in a certain functional perspective. This statement may seem a paraphrase of (A3). In the section (A3), this approach was exemplified by instances of verbal or cognitive activity, which presuppose theoretical solutions. But the general application of this fundamental approach can be found in the text of VP considered as a whole—where different ontological and epistemological solutions appear to be justified, as each provides an answer to a certain theoretical question stipulated by a certain mode of cognitive activity.
Bhartṛhari’s position was meta-philosophical, but he did not formulate his methodological approach in explicit terms. Most probably, he did not possess conceptual means that would be appropriate for such a description. Mutually exclusive approaches expressed in VP belonged to different types of philosophical discourse. The gap between them emerged due to the paradigmatic shift from ontological to epistemological problematic that was characteristic to Indian philosophy in the first millennium CE. Realist solutions belonged to the old ontological mode of philosophical thought, whereas mentalist and functional solutions were rooted in the more recent epistemological problematic. However, the idea of historical change and historical development remained alien to Indian philosophical discourse. As a result, ideas elaborated in the frame of different methodological approaches and on the basis of different presuppositions were generally considered mutually exclusive. Bhartṛhari’s perspectivism was a solution of this collision. Though never formulated as an explicit philosophical strategy, it helped escaping from static ontological structures and capturing the dynamic nature of actual linguistic activity.
For the first time in the studies of Indian philosophy, this term was introduced by W. Halbfass, but mainly with respect to the Jainas and the Vedāntins (Halbfass 1988, pp. 268–269, 414).
It should be noted, however, that perspectivism is not an explicitly formulated doctrine in VP, but a conceptual means introduced for a study of Bhartṛhari’s philosophy in terms of Western philosophical discourse. Therefore, the concept of perspectivism can be criticized or elaborated further. Halbfass in his response to Houben’s paper noted that the question remains, whether Bhartṛhari’s ‘doctrinal tolerance’ was ‘simply part of an inclusivistic strategy to uphold and promote his own basic ideas and systematic commitment?’ (Halbfass 1997, p. 474–475). Cardona in the extensive review criticized Houben for exaggerating the role of perspectivism and neglecting the opinion of the commentators (Cardona 1999). Though this is not the proper place to discuss this argument in detail, it should be noted that Cardona constantly refers to Bhartṛhari as ‘Pāṇinīya,’ basing his arguments on the implication that there was a single successive ‘Pāṇinīyadarśana’ (a philosophical school of grammatical philosophy) to which ancient grammarians, Bhartṛhari, and his commentators must have belonged. Being the traditional view of late doxographers, this implication evidently needs to be justified in historical terms (cf. Todeschini 2010, fn. 14). In any case, the emic approach, according to which the answer to every question must be found in later commentaries, and the etic approach, based on the study of traditional texts in terms of modern philosophy, are obviously complementary and do not necessarily exclude each other.
For the latter idea and its origin in Herder’s philosophy, cf. (Forster 2010, p. 16, 51).
On the meaning of these terms, see below, section (A2a).
I will not discuss this point in this paper. For the general exposition of the problem, cf. (Raja 1963).
The most obvious example of this can be found in the ‘Jātisamuddeśa’ (‘Chapter on Universal’) and the ‘Dravyasamuddeśa’ (‘Chapter on Substance’) of VP, where two mutually exclusive doctrines occasionally appear to be isomorphic. In VP III.1.6 and III.2.2-5, Bhartṛhari comes to the conclusion that every word has its own referent, be it universal (jāti) or substance (dravya), whereas in VP III.1.33 and III.2.11, he considers all universals and material objects (dravya) as manifestations of the unique principle (sat or sattā).
On the latter view and its affinity with Mādhyamaka, cf. (Ogawa 2017).
As Aklujkar put it: ‘…some of the ancient systems of thought can be models rather than direct enumerations of certain entities. One can, for example, look upon Sāṃkhya and Vaiśeṣika as attempts to create models made of what must be minimally assumed if the world is to be explained—a kind of creation of ideal language’ (Aklujkar 2001, p. 455).
Viz. dravya, guṇa, kriyā; and the fourth—ākṛti (form)—corresponding to sāmānya.
In the Aṣṭādhyāyī, both were considered subanta, i.e., word with a case-ending (suP). Yāska and ancient grammarians classified them as nāman ‘name’ (Abhyankar 1961, p. 201).
Though Praśastapāda lived after Bhartṛhari, he formulated and justified the trends in the attitude towards language and reality which can be traced in the earlier Vaiśeṣika. Having this in view, it seems relevant to rely on his work while discussing VP.
Śrīdhara explains abhidheyatva as abhidhānapratipādanayogyatvaṃ and specifies its close relation with jñeyatva: bhāvasvarūpam evāvasthabhedena jñeyatvam abhidheyatvaṃ cocyate (Pbh 1895, p. 16).
According to Potter’s general interpretation undertaken in the broad context of Nyāya, Praśastapāda’s thesis deals not with ordinary language and common sense, but with an ‘ideal language,’ the formulation of which was the part and parcel of Naiyāyikas’ philosophy. It is with concern to this language that Praśastapāda’s thesis is ‘fully reversible: whatever exists (in the broadest sense) can be known and can be spoken of, named; furthermore, if something is known or if it is named, then it exists. Likewise, if something is nameable, it is knowable, and vice versa.’ (Potter 1968, p. 275–277)
prāk saṃjñinābhisaṃbandhāt saṃjñā rūpapadārthikā |
ṣaṣṭhyāṃ ca prathamāyāṃ ca nimittatvāya kalpate ||
tatrārthavattvāt prathamā saṃjñā śabdād vidhīyate |
asyeti vyatirekaś ca tadarthād eva jāyate (VP I.67–68)
On the other hand, in the ‘Mahābhāṣyadīpikā’ on A 2.1.1. vt. 5, Bhartṛhari claims explicitly that on the basis of these words, the existence of their referents can be inferred (Bronkhorst 2011, p. 33).
śabdajñānānupātī vastuśūnyo vikalpaḥ. YS I.9.
On the concept of vikalpa in the ‘Laṅkāvatāravṛtti’ cf. (Unebe 2000).
Cf. fn. 11.
akṛtsnaviṣayābhāsaṃ śabdaḥ pratyayam āśritaḥ (VP III.3.54).
Cf. (Bronkhorst 2011, p. 113).
Though, Bhartṛhari’s monistic approach could have been not completely alien to Buddhist Pramāṇavādins. In the ‘Traikālyaparīkṣā,’ there is a verse borrowed from the Vṛtti on VP I.1 (or according to Frauwallner, initially from Bhartṛhari’s lost ‘Śabdadhātusamīkṣā’), in which the original ‘(amṛtam) brahma’ (imperishable Brahman) is replaced with Tibetan rnam shes (vijñāna), with the quite distinct tendency to understand vijñāna (consciousness) as the monistic principle (Lindtner 1993, p. 201). On Buddhist monistic approach, cf. also (Wayman 1999).
Comparing Bhartṛhari’s and Jñānaśrībhadra’s views on vikalpa, Unebe comes to similar conclusion (Unebe 2000).
vyapadeśe padārthānām anyasattaupacārikī |
sarvāvasthāsu sarveṣām ātmarūpasya darśikā (VP III.3.39)
As an alternative view on upacāra, Tzohar also considers VP II. 285–297, which elaborates an analogy between figurative meaning (referred to with the term gauṇa) and perceptual error (Tzohar 2011: 51–85). However, this passage in VP is evidently pragmatically oriented and deals with a distinct epistemic problem, not with the problem of reference in general. So, in the frames of the classification of approaches to the problem of language and reality presented in my paper, the passage VP II. 285–297 can be rather considered an example of Bhartṛhari’s approach (A3).
Cf. (Bronkhorst 2011, p. 112).
Cf. Houben, ‘the value of the theory of Secondary Existence in Bhartṛhari’s thought… should not be overrated. It is a theoretical solution for problems which arise if certain theoretical choices are made (emphasis on individual word instead of sentence as a whole, and on word meaning as substance instead of universal). At other places in VP different choices are made, and the problems asking for a solution as the postulation of Secondary Existence do not even arise (or they arise in a different form)’ (Houben 1995, p. 272).
indriyāṇāṃ svaviśayeṣv anādir yogyatā yathā |
anādir arthaiḥ śabdānāṃ saṃbandho yogyatā tathā (VP III.3.29)
In a similar sense this term was used in VP I.100.
Kaiyaṭa in his commentary on Mbh designated these two kinds of permanence as pravāhanityatā and kūṭasthanityatā (Mbh, PA 1986, pp. XX, XXIV).
sati pratyayahetutvaṃ saṃbandha upapadyate |
śabdasyārthe yatas tatra saṃbandho ‘stīti gamyate (VP III.3.37)
vyāpāralakṣaṇā yasmāt padārthāḥ samavasthitāḥ (VP III.1.11)
vyavahāro na kalpate (VP III.1.95)
bhedenāvagṛhītau dvau śabdadharmāv apoddhātau
bhedakāryeṣu hetutvam avirodhena gacchataḥ (VP I.59)
On ontological and epistemological trends in VP, cf. (Desnitskaya 2016).
For a detailed analysis of Dharmakīrti’s ideas on perception with respect to arthakriyā cf. (Katsura 1984).
In the Vṛtti on VP I.135, arthakriyā is glossed as lokavyavahāra.
In the Vṛtti on VP I.75 vyavahāra is glossed as śāstravyavahāra (‘śāstric talk’), which can be considered the most general case of purposeful activity for an Indian philosopher.
bhinnadarśanam āśritya vyavahāro ‘nugamyate |
tatra yan mukhyam ekeṣāṃ tatrānyeṣāṃ viparyayaḥ (VP I.75)
On possible affinities of Dharmakīrti’s arthakriyā with philosophy of Nyāya cf. (Kanō 1991).
I thank Professor A. Aklujkar for his valuable comments on the early version of this paper, the anonymous reviewers for their suggestions, and Nicolai Vorobjev for language support.
This work was supported by the Russian Science Foundation (RSF) under Grant 16-18-10427.
- A= von Böhtlingk, O. (Ed.) (2001). Pāṇini’s Grammatik. Delhi: Motilal Banarsidass.Google Scholar
- Abhyankar, K. V. (1961). A dictionary of Sanskrit grammar. Baroda: Gaekwad’s Oriental Series.Google Scholar
- Aklujkar, A. (1970). The philosophy of Bhartṛhari’s Trikāṇḍī. Dissertation: Harvard University.Google Scholar
- Bhate, S. (1994). Bhartṛhari on language and reality. Asiatische Studien: Proceedings of the first International Conference on Bhartṛhari, 47, 67–73.Google Scholar
- Bronkhorst, J. (1996). Sanskrit and reality: the Buddhist contribution. In J. Houben (Ed.), Ideology and status of Sanskrit: contributions to the history of the Sanskrit language (pp. 109–135). Leiden: Brill.Google Scholar
- Bronkhorst, J. (2011). Language and reality: on an episode in Indian thought (Translated from the French by M. S. Allen & R. Raghunathan). Leiden/Boston: Brill (Brill’s Indological Library, 36).Google Scholar
- Chaturvedi, M. (2009). Does language map the reality: Bhartṛhari’s view. In V. Chaturvedi (Ed.), Bhartṛhari: language, thought and reality (Proceedings of the International Seminar Delhi, December 12—14, 2003) (pp. 205–214). Delhi: Motilal Banarsidass.Google Scholar
- Dreyfus, G. (1995). Is Dharmakīrti a pragmatist? Asiatische Studien, 49(4), 671–691.Google Scholar
- Dunne, J. D. (2004). Foundations of Dharmakīrti’s philosophy. Boston: Wisdom Publications.Google Scholar
- Forster, M. N. (2010). After Herder: philosophy of language in the German tradition. Oxford: Oxford University Press.Google Scholar
- Franco, E. (2013). On the periodization and historiography of Indian philosophy. In E. Franco (Ed.), Periodization and historiography of Indian philosophy (pp. 1–34). Wien: de Nobili.Google Scholar
- Griffiths, P. J. (1994). On being Buddha: the classical doctrine of Buddhahood. New York: SUNY.Google Scholar
- Halbfass, W. (1988). India and Europe: an essay in philosophical understanding. New York: SUNY.Google Scholar
- Halbfass, W. (1992). On being and what there is: classical Vaiśeṣika and the history of Indian ontology. New York: SUNY.Google Scholar
- Halbfass, W. (1997) Research and reflection: responses to my respondents. IV. Topics in classical Indian philosophy. In E. Franco, & K. Preisendanz (Eds.), Beyond orientalism: the work of Wilhelm Halbfass and its impact on Indian and cross-cultural studies. (p. 471–486). Poznań Studies in the Philosophy of the Sciences and the Humanities, 59. Amsterdam/Atlanta: Rodopi.Google Scholar
- Herzberger, R. (1986). Bhartṛhari and the Buddhists. An essay in the development of fifth and sixth century Indian thought. Studies of classical India, 8. Dordrecht: Kluwer.Google Scholar
- Houben, J. E. (1995). The Saṃbandha-samuddeśa (chapter on relation) and Bhartṛhari’s philosophy of language: a study of Bhartṛhari Saṃbandha-samuddeśa in the context of the Vākyapadīya, with a translation of Helārāja’s commentary Prakīrṇa-prakāśa. Groningen: Egbert Forsten.Google Scholar
- Houben, J.E. (1997) Bhartṛhari’s perspectivism (1): the Vṛtti and Bhartṛhari’s perspectivism in the First kāṇḍa of the Vākyapadīya. In E. Franco, & K. Preisendanz (Eds.), Beyond orientalism: The work of Wilhelm Halbfass and its impact on Indian and cross-cultural studies (pp. 317–358). Poznań Studies in the Philosophy of the Sciences and the Humanities, 59. Amsterdam/Atlanta: Rodopi.Google Scholar
- Kanō, K. (1991) On the Background of PV II 12ab—the origin of Dharmakīrti’s idea of arthakriyā. In E. Steinkellner (Ed.), Studies in the Buddhist Epistemological Tradition: Proceedings of the Second Dharmakīrti’s International Conference, Vienna, June 11–16, 1989 (pp.119–128). Wien: Verlag der Ősterreichischen Akademie der Wissentschaften.Google Scholar
- Lindtner, C. (1993). Linking up Bhartṛhari and the Bauddhas. Asiatische Studien/Etudes Asiatiques, 47, 195–213.Google Scholar
- Lysenko, V. (2005) Le principe de correspondence: la version de Prashastapada. In Categories de langue et categories de pensee en Inde et en Occident (pp. 79–90). Paris: L’Harmattan.Google Scholar
- Matilal, B. K. (1986). Perception: an essay on classical Indian theories of knowledge. Oxford: Clarendon Press.Google Scholar
- Matilal, B. K. (1990). The word and the world: India’s contribution to the study of language. Delhi: Oxford University Press.Google Scholar
- Mbh = Kielhorn, F. (Ed.) (1880) Patañjali’s Vyākaraṇa-Mahābhāṣya. Vol. I. Bombay: Government Central Book Depot.Google Scholar
- Mbh, PA = Joshi, S. D., & Roodbergen, J. A. F. (Ed. & Trans.). (1986). Patañjali’s Vyākaraṇa-Mahābhāṣya. Paspaśāhnika, introduction, text, translation and notes. Publications of the Centre of Advanced Study in Sanskrit, Class C, No. 15. Poona: Poona University.Google Scholar
- MbhD = Bronkhorst, J. (Ed.) (1987). Mahābhāṣyadīpikā of Bhartṛhari. Āhnika I. Poona: Bhandarkar Oriental Research Institute.Google Scholar
- Nagatomi, M. (1967). Arthakrīya. Adyar Library Bulletin, 31, 52–72.Google Scholar
- NB = Stcherbatsky, T. (Ed.) (1918). Nyāyabinduṭīkā of Dharmottara. Petrograd: Bibliotheca Buddhica.Google Scholar
- Ogawa, H. (2017). Two truths theory: what is vyavahāra? Language as a pointer to the truth. Journal of Indian Philosophy, 1–21.Google Scholar
- PBh = Dvivedin, V. P. (Ed.) (1895). The Bhāṣya of Praśastapāda together with Nyāyakandalī of Śrīdhara. Benares: Lazarus & Co.Google Scholar
- Potter, K. (1968). Astitvajñeyatvābhidheyatva. Wiener Zeitschrift für die Kunde Süd- und Ost-Asiens und Archiv für Indische Philosophie, 12-13, 275–280.Google Scholar
- PV = Sanskrityayana, R. (Ed.) (1953). Pramāṇavārtikabhāṣyam or Vārtikālaṃkāra of Prajñākaragupta. Varanasi: Banaras Hindu University.Google Scholar
- Raja, K. K. (1963). Indian theories of meaning. Madras: Adyar.Google Scholar
- Todeschini, A. (2010). Bhartṛhari’s view of the pramāṇas in the Vākyapadīya. Asian Philosophy, 20(1), 97–109.Google Scholar
- Tzohar, R. (2011) Metaphor (Upacāra) in early Yogācāra thought and its intellectual context. Dissertation: Columbia University.Google Scholar
- Urban, H. B., & Griffiths, P. J. (1994). What Else remains in Śūnyatā? An investigation of terms for mental imagery in the Madhyāntavibhāga-corpus. Journal of the International Association of Buddhist Studies, 17(1), 1–25.Google Scholar
- VP, Kā I = Iyer, K. A. S. (Ed.) (1966). Vākyapadīya of Bhartŗhari with the commentaries Vŗtti and Paddhati, Kāņḍa I. Poona: Deccan College.Google Scholar
- VP, Kā III = Iyer, K. A. S. (Ed.) (1963). Vākyapadīya of Bhartŗhari with the commentary of Helārāja. Kāņḍa III, part I. Poona: Deccan College.Google Scholar
- VS = Jambuvijayi, M. Śrī (Ed.). (1961). Vaiśeṣikasūtra of Kaṇāda with the commentary of Candrānanda. Baroda, Oriental Institute (GOS 136).Google Scholar
- Wayman, A. (1999). Does the Buddhist momentary theory preclude anything permanent. In S. Katsura (Ed.), Dharmakīrti’s thought and its impact on Indian and Tibetan philosophy: Proceedings of the Third International Dharmakīrti Conference, Hiroshima, November 4–6, 1997 (pp. 433–440). Wien: Verlag der Ősterreichischen Akademie der Wissentschaften.Google Scholar