THE ROUTLEDGE HANDBOOK OF PHILOSOPHY OF RELATIVISM

term used to frame the process of transition from an apartheid regime to a new “rainbow nation” and an African renaissance. There is no consensus on what ubuntu actually means, and the precise content of the concept is still contested. The translations range from “humanity” and “charity” to “common sense” and “generosity.” Often ubuntu is seen as a concept enshrined in a traditional philosophy of life, although one needs to further differentiate between ubuntu as a moral quality of a person or as a way of living. The core meaning of the concept ubuntu is frequently expressed using the Zulu-Xhosa aphorism “umuntu ngumuntu ngabantu” – “A human being is a human being through other people.” meaning that every human being needs other people in order to be human; every person is part of a whole, integrated into a comprehensive network of mutual dependencies. The aphorism expresses “the African idea of persons: persons exist only in relation to other persons. The human self . . . only exists in relationship to its surroundings; these relationships are what it is. And the most important of these are the relationships we have with other persons” (Shutte 2001, 23). Thus, the aphorism refers to the deep relational character of ubuntu and underlines at the same time that human beings (umuntu) are a “being


Introduction
Thinking about classical Indian philosophy in light of relativism is a challenging hermeneutic task. There are no readymade volumes in the classical literature that we can identify under this category. Siderits argues along these lines that "the cultural factors that make relativism a pressing issue for us were largely absent from the classical Indian context, so that the various forms of relativism do not receive philosophical scrutiny in the Indian tradition" (2016,24). The fundamental problem in thinking about Indian philosophy through relativism is not that there are no readymade texts but that scholars refrain from engaging relativism, as if it is taboo or a disease that philosophers need to stay away from (e.g. Siderits 2016, 31, 35). My own approach to relativism is relativistic, as I believe that endorsing relativism in one respect does not require one to be relativistic in all accounts. Just like any other "ism," relativism should be handled as a device to fathom human nature and to help humanity negotiate a perplexing, complex social reality. When we open ourselves to read classical Indian materials through the lens of relativism, we encounter a wealth of materials. Dialogues recorded in Vedic literature epitomize cultural fluidity, diversity and an openness to perspectives. Traditions have adopted perspectivism to make sense of an otherwise bewildering variety of commentarial literature with conflicting interpretations. The problem then is we encounter a semblance of relativism and can be easily misdirected. Before we assign epistemic relativism in the Jain "multiperspectivalism" (anekāntavāda) or moral relativism in the Mahābhārata or meaning relativism in Bhartṛhari's philosophy of language, we need to carefully define the categories and explore the parameters. Cultural pluralism was a norm in classical India and every region dealt with religious differences. Everyday society also incorporated linguistic differences and grammarians such as Patañjali were keenly aware of dialectical variations even within a single language. Combined with polytheism and panpsychism, India is founded upon the co-existence of different and Review Copy -Not for Redistribution Manuel García-Carpintero -University of Barcelona -27/07/2020 at times, conflicting viewpoints. Written in this cultural milieu, texts such as Bhagavadgīta endorsed different soteriological approaches by necessity, to combine multiple methods for liberation. It is not possible to address all these issues in a few pages. I therefore limit myself to re-examination of some of Nāgārjuna's claims, keeping in mind both classical and contemporary interpretations. I explore, in particular, the doctrine of "two truths" and Nāgārjuna's interpretation of the "viewpoints" (dṛṣṭis). In so doing, I am open to drawing parallels and initiating a cross-cultural dialogue on relativism. In conclusion, this conversation boils down to relativism leading to truth skepticism on the one hand and pluralism and hierarchical truth predications on the other.

Nāgārjuna on viewpoints (dṛṣṭis)
Nāgārjuna (150-250) is one of the major Buddhist philosophers and the founder of the Mādhyamika school. Scholars have primarily read his philosophy for its dialectical methods, rejection of substantialism, and interpretation of the doctrine of "emptiness" (śūnyatā). Most importantly, he is known for his pioneering doctrine of "two truths" (dve satye) and deconstruction of "viewpoints" (dṛṣṭi). Nāgārjuna introduces a unique logical method that reduces the opponent's viewpoints to absurdity (reductio ad absurdum) to defend his position that there is no inherent nature (svabhāva), whether by ontological truth claims regarding substance, or epistemic claims regarding reality -including the limits to human rationality. Nāgārjuna explores any proposition in terms of fourfold possible extremes (koṭi), eventually proving it absurd to adopt any one of those extremes.
Regarding the inherent nature (svabhāva) of being and things, Nāgārjuna posits and then refutes that: (1) Things have inherent nature ("is" thesis).
(2) Things do not have inherent nature ("is not" thesis).
(3) Things simultaneously possess and lack inherent nature ("is and is not" thesis).
(4) Things lack both the inherent nature and the lack thereof (not -"is and is not" thesis).
Regarding causality, he likewise proposes as categories that: (1) Things emerge because of the internal factors ("svataḥ" or "from within" thesis).
(3) Things emerge due both to the internal as well as the external factors ("dvābhyām" or "from both" thesis). (4) Things emerge without any cause ("ahetutaḥ" or from "no cause" thesis).
In rendering this thesis of an "intrinsic nature" (svabhāva) absurd, Nāgārjuna establishes the doctrine of "emptiness" (śūnyatā). Examining this discussion historically, what he says is that, just like aggregates do not have their own intrinsic nature (the position that the Abhidharma school has endorsed), so also do the building blocks of the manifest reality, the so-called dharmas, not have any inherent nature. The tricky part is that he is not advocating this last statement as his thesis. The argument is if the emptiness of inherent nature were a thesis, this would be tantamount to endorsing absolutism by another name. Therefore, the negation of intrinsic nature is just a negation. The problem is that this understanding only partially captures the way Nāgārjuna has been historically understood. Reading Nāgārjuna Review Copy -Not for Redistribution Manuel García-Carpintero -University of Barcelona -27/07/2020 is perplexing for both the classical commentators and contemporary scholars alike. The following verse is ground-zero of our investigation: The teaching of the dharma(s) by the Buddha relies on two truths: the limited conventional truth and the truth as it is.
(MMK XXIV.8) 1 There are obviously two different ways to understand this passage. It can mean that phenomenal truth exists and only applies to conventional reality and that absolute truth transcends language and concepts. This understanding of a hierarchy of truth does not reject truth claims, and can be interpreted in two different ways: first, that there are two tiers of truth, or second, that there are different sets of truths. In another possible interpretation of "two truths," this verse can also be explained by truth that is conceived of in the "covered" (saṃvṛti) state. For example, a truth such as seeing a sand dune as mirage or a rope as a snake, does not amount to actual truth due to its origination within a state of delusion. As a result, this view asserts that truth only exists corresponding to the way the entities are (parama-artha-taḥ). Therefore, a correspondence theory of truth underlies this interpretation. And if this position is followed, Nāgārjuna would not be making any anti-foundational claim in the exalted sense. This reading, however, would contradict Nāgārjuna's own proclamation that there is no "inherent nature" (svabhāva), as this would simply be replacing one form of absolutism with another. This would also contradict Nāgārjuna's direct statement that openly rejects absolutism regarding emptiness (śūnyatā): It is not our fault that you resort to emptiness. No foundation (sa = adhilaya) can be established on emptiness.

(MMK XXIV.13)
If what is described in terms of [the entities] lacking their inherent nature is the very being of the lack of the inherent nature, this would negate the lack of inherent nature and only the being of inherent nature would be established.
(VV 26, see Bhattacharya et al. 1978) 2 Keeping these straightforward stanzas in mind, Siderits argues that the term paramārtha or "the way the things are" does not confirm any ultimate truth, but on the contrary, "the ultimate truth is that there is no ultimate truth" (Siderits 1989, 231). Garfield confirms this same interpretation: Suppose that we take a conventional entity, such as a  (Garfield 2002, 38-39) There are two possible responses to the preceding statements, and both were historically applied by Nāgārjuna. One response is to reject such a claim, demonstrating circularity in its logic, arguing that even this amounts to a truth claim. The other is to apply linguistic or conceptual tactics to interpret negation while keeping open the possibility of speaking about the truth. The current conversation on relativism claims a central place in this shift from a correspondence theory Review Copy -Not for Redistribution Manuel García-Carpintero -University of Barcelona -27/07/2020 of truth. Whether to understand Nāgārjunian claims as metaphysical or semantic is not a new quandary. So far, recent discussions and arguments are a flimsy replica of the debate between the Prāsaṅgika and the Svātantrika readings sustained over millennia. 3 The dilemma though is if this is a rejection of the absolute truth, and the conventional is not the "truth" per se, there is no truth to defend. With this view, the category "truth" would be fictitious, like rabbit-horn. And if this is only the rejection of absolute truth but not of relative truth and therefore interdependent truth, this would mean that truth is always relative, perspectival, and this position is not a rejection of "truth."

Truth: metaphysical or semantic
If what Nāgārjuna meant is that there are two truths, this would be a metaphysical theory, a theory about the ultimate nature of reality. The semantic interpretation recognizes this proclamation as not about the nature of reality but about the nature of truth. Siderits explains that, "all things are empty [means] that the ultimate truth [has] no ultimate truth -there is only conventional truth" (Siderits 2003, 11). This would help to separate truth claims from metaphysical reality and we could say, the statement "Rāvaṇa had ten heads" is true based on narratives, irrespective of the possible existence of such a monster. Returning to the position of "two truths," a semantic interpretation claims that no statement can be ultimately true. Siderits argues further, "Given that dharmas must be things with intrinsic natures, if nothing can bear an intrinsic nature, then there is nothing for ultimately true statements to be about; hence the very notion of ultimate truth is incoherent" (2003,(11)(12). It appears Siderits draws from Hilary Putnam to develop a thesis that requires the rejection of any singular truth regarding the nature of reality that would presuppose a model of metaphysical realism. The target is to reject "emptiness" (śūnyatā) itself as a metaphysical claim. And this position omits the demolition of such a premise by the logical fallacy of circularity. To say that "there is no final truth about reality" would also apply to the claim that "all things are empty," which of course one would expect the Mādhyamika philosophers to reject. And historically some have taken this route. Siderits, however, suggests that even the claim "all things are empty" is only conventionally true.
Re-contextualization of the claims is necessary to establish any form of relativism based on the aforementioned position. To say that truth is only conventional, the conclusion derived from Siderits' reading, opens up a potential space for multiple perspectives in which all retain a degree of validity. This, however, is not what Siderits proposes and it deviates from Nāgārjuna's position, as it yet again underlies a supposition on the truth per se; specifically that, in an underlying metaphysical claim, even absolute truth can only be relatively revealed. The rejection of absolute truth does not, however, confirm the validity of viewpoints (dṛṣṭi), as has already been argued. To assume all that can be spoken of truth are just viewpoints does not mean the same judgment can't be true in one perspective while false in another. Nāgārjuna is not proposing that the human encounter with reality is mediated by language or culture. But if we were to read that "two truth" theory affirms perspectives, while not discrediting the category truth in the ultimate sense, we can derive that truth is relatively revealed in different modes. We can now engage G. Ferraro's (2013) arguments with this new accommodation to address relativism.
Ferraro argues against this semantic reading, maintaining that Nāgārjuna's doctrine of "two truths" upholds "two visions of reality on which the Buddhas, for soteriological and pedagogical reasons, build teachings of two types" (2013, 563). Emptiness (śūnyatā), in this reading, is in fact Review Copy -Not for Redistribution Manuel García-Carpintero -University of Barcelona -27/07/2020 "equivalent to supreme truth." To make his claim, Ferraro first divides the metaphysical claims into two groups: (1) a realistic metaphysical reading that considers "supreme truth an existing and somehow characterizable dimension," and (2) an anti-realistic metaphysical reading that denies the "existence of supreme truth" and affirms "existence exclusively of ordinary reality" (2013, 566).
Now the argument is that whatever applies to our pedagogical approach also applies to the use of language: our use of language or words are relational, and while our objective may be to speak the "truth," given that there are metaphysical truths to be conveyed by language, our approaches can vary. Consequently, we can derive that the conventional is a necessary step, that we can discuss truth only conventionally. And since it is counterintuitive to conceive of the "conventional" as being a single perspective, the discourse on truth automatically becomes perspectival and relational. This claim, therefore, could reject both the metaphysical claim, and the validity of the so-called supreme truth. The fundamental divergence in this interpretation with Siderits and Garfield (2013) arises due to confusion between metaphysical and semantic interpretations. Siderits and Garfield argue that semantic interpretation does not interpret "two truths," but demonstrates that truth is a semantic property. In the Buddhist historical context, if reality is analyzed based on dharmas or essential factors, the emptiness doctrine says that even dharmas lack inherent characteristics and thus are devoid of intrinsic nature. In this sense, what Siderits and Garfield propose only negates the reality of what is proposed as a higher reality of dharmas.
There is not much new to add, except to point out that contemporary conversations are enriched with nuances borrowed from a global philosophical discourse. While we should persist in the hermeneutic task, our first loyalty goes to reading the texts the way that they are. One can be relative about different interpretations but not about the actual words. And when we look back to the texts themselves, we encounter that the term satya for example, is not just for the truth but also for reality. Derived from the Sanskrit verbal root asa, the term only describes the mode of sat, or that which is. Nāgārjunian terms for the so-called two truths are saṃvṛti and paramārtha, where the first does not translate to "relative" but "covered," and it also means "covering." There is nothing "supreme" in the paramārtha either, as the term is a compound of "parama + artha" with the first being in the superlative of para meaning the other, and thus meaning the last or the final, and artha referring to both "meaning" and "reality." Multiple interpretations of the same text lead to hermeneutic relativism. Even when we ignore the examples where the same commentator derives different meanings from the same passage, texts come with multiple commentaries with contrasting meanings. For a reader, there are always options in determining meaning. A relativistic hermeneutic approach, however, does not open a text to anarchy in meaning. Even the skeptics such as Jayarāśi were not skeptical about reason per se. And the openness of interpretation only meant that readers needed to be openminded about perspectives as far as the semantic power of words can accommodate. When we read Sanskrit literature, we not only come across multiple commentaries, we even encounter different interpretations in the commentary written by the same author who composed the original text. Buddhist literature is no exception to the phenomenon of different interpretations for the same passage or the same author composing the text and its commentary. All in all, there is no taboo for a multi-façade-interpretation as far as classical exegesis is concerned. If we give credit to Nāgārjuna for being the philosopher that he is, it is not hard to conceive that he is aware of both possibilities, and is leaving the text open-ended regarding Review Copy -Not for Redistribution Manuel García-Carpintero -University of Barcelona -27/07/2020 the ways it can be read. The interpretative prowess within the context of MMK is epitomized in the commentarial literature.
For our current purpose, let us say Nāgārjuna makes a realistic metaphysical claim that he considers a two-tier truth theory. Even if this does not directly confront relativism, we can accept that teaching methods and what is described are relative to the audience. In other words, our words can mean what they mean based on external factors. Accordingly, the teaching of the four noble truths (suffering, origination, cessation of suffering, and the means to end suffering) relates to "adopting the limited perspective"; while teaching emptiness relates to "following the supreme truth." On the other hand, if we follow Siderits and Garfield (2013), we are left with just perspectives and can only confirm relativism. However, these are very different types of relativism. The relativism that fits better with perspectivism should not be conflated with relativism regarding rationality. Even following Siderits and Garfield in this regard, there is no need to confirm that all epistemic claims are equal or that knowledge is a norm of assertion governing rational inquiry (see Walsh 2015). Whichever position, reading Nāgārjuna in light of relativism remains valid. However, if we mean "hard" relativism, we can argue along the lines of what Siderits says: The Prāsaṅgikas, with their no-theory approach to conventional truth, would be forced to accept the relativism about rationality that such evidence seems to suggest. But the Svātantrikas could, I think, be pluralists without being relativists: pluralists in admitting a plurality of possible canons of rationality, no single one of which is ideally suited to uncover the ultimate nature of reality; but they could not be considered relativists in that one such canon may quite straightforwardly be said to be better than another. (Siderits 2016, 35)

Moving beyond Nāgārjuna
Even more important than asking, "why did Nāgārjuna start with causation?" (Garfield 1994) would be to ask, "why did Nāgārjuna end his masterpiece with 'viewpoints' (dṛṣṭi)?" Rejection of causality grounds the Mādhyamika philosophy. Deconstruction of the "viewpoints," on the other hand, destabilizes the entire philosophical enterprise. The imprints of Nāgārjuna are visible in the lines of Śrīharṣa, a prominent Advaita philosopher who lived one millennium after Nāgārjuna. 4 By critiquing other viewpoints, Nāgārjuna is not proposing his own thesis, which would be counterintuitive. He himself cautions, "the victorious ones have proclaimed that there is no foundation as there is emptiness of all views. However, to whom emptiness [itself] is a view, they are considered incorrigible" (MMK XIII.8). It is therefore not the case that Nāgārjuna is rejecting the theory of causality; he is rejecting the viewpoints, and the first among them happens to be the theory of causality. For him the fundamental human problem is not the lack of theories but our obsession with them Another key position to derive relativism comes from Maṇḍana Miśra. For him, our everyday reality is composed of our own ignorance (avidyā) and the individual subjects are the locus of this metaphysical ignorance. This position results in saying that all we can encounter by means of our cognitive faculties and semantic analysis are just the perspectives, each conditioned by our own preconceived notions, and filtered by means of the habit patterns (saṃskāras). Every individual, in this paradigm, projects his own world. Accordingly, each has his own conceptualized truth, guided by one's own presuppositions and misconceptions. 5 Since all that we can communicate regarding the nature of reality is mediated by our concepts, which in turn are the Review Copy -Not for Redistribution Manuel García-Carpintero -University of Barcelona -27/07/2020 conditions from our past experiences, this thesis does lead to some form of relativism. Expanding upon the philosophy of Maṇḍana, the non-dualist philosophers (Advaitins) argue that collectively shared experiences are what they are because subjects having homogenous experiences do share a common history. Borrowing their own example, this is similar to multiple subjects having the dream of a snakebite and coming to the conclusion that they all dreamt of the same snake. Just as dream experiences are subjectively circumscribed, so also are other experiences. Even our experiences of pain and pleasure corresponding to certain stimuli are rooted in habitual tendencies that constitute some experiences as painful and others as pleasant. This does involve bodily memory. This is to say that we are not able to escape our corporeal and psychological horizons in our pursuit for grounding our experience. What are we left with then? Just our "viewpoints" or "perspectives" (dṛṣṭis). However, this is as far as their agreement goes, as the Advaita philosophers are not relativistic with regard to the absolute reality of the being equated with consciousness (sat-cit). For them, every mode of experience and every perspective underlies the same principle of being and consciousness. For them, being and consciousness are a logical necessity for every is affirmation or negation. They see this as something that cannot be rejected by means of negation, and for them, the foundational being and awareness is not yet another perspective but only the possibility for the perspectives to be, and not the truth of all the truths, but merely the categorical possibility that makes us think about truth in general. Our everyday modes of experience, accordingly, do not negate experience as a category. The argument here is that subjects can bracket the factors that condition experience, including the ego, and enable being in a mode that is not subject to conceptualization. This is not to say that there is nothing real; this is not surrender to any form of nihilism. This is a proclamation that any truth-claim is relative, or perspectival.
One may conflate this position with Kantian transcendental idealism. And some early scholars reading Nāgārjuna such as Tiruppattur R. Venkatachala Murti have found comfort in such a charge. Following this, just as the objects we intuit in space and time are appearances, the mental states that we intuit in introspection are likewise appearances. We can nevertheless think of things in themselves using categories, as they affect our sense faculties. This conflation, however, misses a major distinction whether it be a Nāgārjunian or Advaita position: the entire philosophical endeavor cannot be isolated from the goal of "apprehending the way things are" (yathābhūtārtha-darśana), or "direct apprehension" (sākṣātkāra). When scholars say that our experiences are shaped by our habit tendencies (saṃskāra) and that all we experience, conceptualize, and verbally express are mere copies of the way things are and that what it actually is cannot be expressed; this is never meant to conclude that we are incapable of overcoming our own subjectivity. The resultant position advocates some variations of semantic and epistemic relativisms, while retaining the possibility of different types of metaphysical realism.
In the discourse on relativism, the Jain "multiperspectivalism" (anekāntavāda) is sometimes imagined to be relativism itself. This, however, is not the case. In its most systematized form, this doctrine for any given situation consists of sevenfold possibilities: (1) It may be.
(2) It may not be.
(3) It may and may not be.
(4) It may be but is not describable. (5) It may not be while being indescribable. (6) It may both be and not be while being indescribable. (7) It may simply be indescribable.

Review Copy -Not for Redistribution
Manuel García-Carpintero -University of Barcelona -27/07/2020 This is not a thesis that truth is relative to individual subjects, or that everyone has her own truth conditioned by her language and culture. Another way this has been confused is by equating it with perspectivism. "May be," to begin with, is not proposed as yet another perspective, and none of these are individually circumscribed to be true. This is rather saying that truth is manifold, or that each of these constitutes a part of the truth that is revealed only globally when all aspects have been analyzed. Another way this has been understood, is as a form of pluralism. It seems appealing to argue that there are multiple perspectives to the truth, but in fact, what the doctrine is saying, is that while different doctrines make different truth-claims, none of these have the total picture of the reality when accepted individually. That is, there is a truth claim when the totality of the possibilities is accepted, but not that truth is only revealed as a perspective and that all of them have some sort of validity if taken individually. What this implies is that one who has all the perspectives has the truth. And this can be better explained as "mosaicism": that each component of a mosaic comprises a necessary element for constituting the truth, but no single piece of the mosaic alone can reveal the truth the way it is.

Conclusion
It would be wrong to equate any of the aforementioned positions with relativism. But fortunately, there are many kinds of relativism and when engaging Sanskrit philosophical literature, we may have encountered a different variety, or varieties of relativism that are not just antecedent to contemporary forms of relativism. What applies to most Indian traditions, is that being relative about truth is not to deny the category "truth" but to assert that our rationality and comprehension of what is true is relative, and that there are external factors to underscore our ways of reasoning or our grasping of what we consider to be true. But in all accounts, truth as a category underpins this assumption. Different subjects from varied cultural backgrounds might share different values and different systems of judgment and from a meta-gaze we may see relativism in their perspectives. However, this does not apply, that subjects endorsing such views consider them as relative. Each and every cultural subject has their own unique experiential and epistemic horizon that is for them the only truth. Those who are capable of distinguishing their personal perspective from among other viewpoints, are subjects possessing a "meta-gaze" and in some regards, are the liberated (mukta) subject, able to transcend their own subjective horizon.
The preceding discussion provides a framework for re-contextualizing moral relativism in the Mahābhārata. Overall the text teaches non-violence (ahiṃsā) although every page of it is saturated with the blood of the antagonists and heroes. A small section from it, the Bhagavadgītā, epitomizes the tension between relative and absolute perspectives on morality, vividly portrayed as the clash between the individual duty of a warrior (Arjuna) to fight, and the universal dharma of non-violence. There is no relativism about non-violence: this is the single most absolute upon which the other absolutes such as truth (satya) and "not stealing" (asteya) are founded. In this tension between the universal and individual dharmas, Arjuna recognizes the necessity to perform his individual dharma. Is this a simple justification for a war? If the book is teaching anything, it is that individual perspectives or truths triumph over global perspectives, but this can be allowed when and only when the global perspective is at peril. Ahiṃsā, it seems, is not always capable of defending itself. Not by choice but as the final resort when all options have been exhausted, Arjuna is left to decide between a lesser evil of confronting violence with violence, or a greater evil of avoiding it. This isn't because a warrior wants to kill or craves fame but because those being killed and raped are unable to defend themselves with a mere vow of non-violence. A warrior allows himself to act within the universal dharma so that the others can uphold it and Arjuna chooses his personal truth: as a warrior he has to fight. The difference in perspective is, Review Copy -Not for Redistribution Manuel García-Carpintero -University of Barcelona -27/07/2020 prior to the teachings, there is Arjuna a prince, a brother, and a husband deeply wounded by the atrocities of his enemies, while after the teachings, there is just a warrior who recognizes his role, his moral responsibility which makes the global sense of morality possible.

Notes
1 Refercens to Nāgārjuna's Mūlamādhyamakakārikā are cited using the abbreviation MMK, number of the chapter and verse or half-verse, e.g. "MMK XXIV.8." Please find the full reference in the bibliography under Kalupahana (ed.) (1986). 2 Refercens to Nāgārjuna's Vigrahavyāvartanī are cited using the abbreviation VV and the verse, e.g. "VV 26." Please find the full reference in the bibliography under Bhattacharya et al. (1978). 3 The classical analysis of "two truths" is complex. Candrakīrti, for example, divides saṃvṛti as real empirical and unreal empirical in order to make a distinction between the conventional and erroneous objects. Bhāvaviveka makes a distinction between the conceptualized and actual truths when addressing the paramārtha. For further analysis of the Prāsaṅgika-Svātantrika distinction, see Dreyfus and McClintock (2003).