The Indian philosophical tradition, as everyone familiar with it knows well, is abundant with robust and brilliantly defended metaphysical, epistemological and ethical positions of all varieties. Its several millennia of development have featured stalwart defenders and innovators of scholastic views, independent and encyclopedic thinkers, trenchant critics of established doctrines and assumptions, mystics, ritualists, grammarians and anti-religious, worldly sages. There may be many who, for good reasons, find themselves drawn to a particular school, thinker or persuasion in the rich spectrum of ideas in Indian thought, who then become compelled to embrace and enhance one of these available perspectives. Others may find room for great syntheses of views under a broadly overarching system, and indeed the penchant for synthesis is one of the powerful achievements of South Asian genius. Still others might, in the long course of exploring the contesting arguments of Indian thought, become “converts,” abandoning an initially favored philosophical orientation for another one that has become more convincing. I am one of those converts. In this second of my planned three posts this month to the Indian Philosophy Blog, I will briefly recall that process of conversion.
As noted in my last post, my initial draw toward the Indian philosophical tradition, during my undergraduate studies, was through the works of Schopenhauer. His own appropriation of 19th century Orientalist depictions of early and medieval Indian texts resulted in several major features of his system. First, his reading of the early Upaniṣads and some Vedāntic glosses inspired him to interpret Kantian transcendental idealism as the way our cognitive processes falsify our knowledge of the world, creating what he called a “veil of māyā.” Instead of being sensorally or conceptually aware of the unitary metaphysical nature of the world as one homogeneous will, we take the world to be filled with heterogeneous phenomena, individuated as “this and that” and “I’ and “other.” Second, his understanding of the formulations of tat tvam asi in the Upaniṣads and the Gitā impelled him to take it as an ethical maxim, a move that, as Paul Hacker showed, had much influence on Vivekanada through Paul Deussen. In this construal, genuine ethical motives are infused with the insight that we are not of different fundamental natures, but are all manifestations of one and the same will, and this insight fuels the compassion required to alleviate one another’s suffering. Third, Jain and Buddhist forms of renunciation were for Schopenhauer manifestations of the “denial of the will,” a final mystical turning away from the violence and conflict to which the will drives all individuated beings.
Though I had misgivings regarding Schopenhauer’s characterizations of renunciation from the start, I found, at first, his construal of idealism fascinating and his thematization of the “tat tvam asi” ethic moving. As I entered my graduate studies, I set out to learn in greater detail the schools, texts and arguments that I thought could bolster idealism, monism and an ethic of compassion akin to Schopenhauer’s. I dove into modern scholarly renditions and interpretations of the Upaniṣads themselves, as well as the rather wide-ranging classical and modern readings of Śańkara’s Advaita, early Buddhist texts and Vijńānavāda Buddhist thought. The overriding commitments which fired these early studies were a strong sense that our proclivities for drawing distinctions between things and self were shaped not by perceptual recognition but by our own cognitive processes and that we are fundamentally desirous beings who can only find solace in compassion and equanimity. I was following a path through Indian thought upon which many Westerners find themselves when they are given primarily Orientalist and Orientalist-inspired directions.
Over the course of my studies, these convictions began to loosen. Both hermeneutic criticism of Schopenhauer’s reception of Indian thought and well-considered Western philosophical critiques of Schopenhauer’s own system helped me to become more objective about this so-called “sage of Frankfurt” and his legacy. During the same period, the appeal of Śańkara’s Advaita to me was progressively undermined. The first weakening was facilitated for me by the 12th century’s Ramānuja and then the 20th century’s J.N. Mohanty and Sengaku Mayeda. Their analyses and critiques of the inconsistencies in Śańkara’s attempt to reconcile the unity of the ātman with his description of the perceptual process brought problems into sharp relief. Next came a brilliant short essay by Daya Krishna, making the case that Śańkara was not a monist at all, but more akin to a Sāḿkhya philosopher who believed in a single metaphysical self. Finally, 20th century critics of Advaita like B.S. Yadav and Francis Clooney revealed an extremely conservative positions on caste taken by Vedāntins, which threw into great question the entire supposition that anything like a modern tat tvam asi ethic prevailed among them. The withering Buddhist arguments against the notion of a fixed ātman as well as Nāgārjuna’s incisive analyses of causation, which seem to destroy both idealist or realist views, also dug the roots out of my youthful attraction to classical Indian “non-dualism.”
In my later graduate work under Mohanty and Yadav, and in the early years of my own teaching career, I began to study the classical Nyāya philosophers, primarily Gautama, Vātsyāyana, Uddyotakara, Udayana and the Nyāya commentary of Vācaspati Miśra. Initially, their texts needed to be understood in order to evaluate the centuries of debates over knowledge, metaphysics, logic and language that were waged against the Naiyāyikas by centuries of Buddhists. At the beginning, neither the Nyāya brand of realism, nor any brand of philosophical realism, were ever appealing to me. Though I have been a lifelong admirer or science, philosophical realism for the young me appeared unraveled by Hume’s skepticism, the verification debates of 20th century analytic thought and the findings of modern quantum theory itself, which suggest that, at the micro-level of reality, our acts of observation alter what we perceive and measure instead of reflecting a static reality. In the classical Indian context, the Vijńānavāda refutations of atomism and Śri Harṣa’s relation regress criticisms did not appear to leave much room for realism either. But then, something important happened. Two of my mentors strongly recommended that I read the late B.K. Matilal’s 1986 masterpiece, Perception: An Essay on Classical Indian Theories of Knowledge. To put the matter simply, Matilal made me into a realist. A qualified realist, to be sure, but a realist nonetheless.
In the years that have passed since my early career, my two main foci of interest in the Indian tradition have been classical Nyāya philosophers and the works of the Madhyamaka Buddhist preceptor Nāgārjuna. Ultimately, I think the formidable criticisms that Nāgārjuna brings against the Nyāya theory of knowing in his Vigrahavyāvartani are satisfactorily met by the ensuing centuries of Naiyāyikas. I am in no way committed to the larger Nyāya ontology, which is certainly too extravagantly furnished, nor to their intermittent theism. However, their characterizations of the perceptual and imaginative processes, their rebuttals of regress arguments, Buddhist part-whole analyses and attempts to destroy externalism still seem to me quite powerfully articulated, not just within the Indian context but beyond it. Nyāya stands as a robust defender, contra Buddhist forms of reductionism, of the view that compound phenomena need not be considered any less real than simple phenomena, and this vindicates the heterogeneity and complexity of existence. On the other hand, despite his epistemological vagueness and at times incomplete argumentation, Nāgārjuna offers us a profound understanding of interdependent causation. His notion of causation is one I consider to be an advancement on early Buddhist theories, to the degree that Nāgārjuna does not merely insist that every event must be causally conditioned, but that causality is multi- and not unu-directional. This has lasting implications not only for our understanding of reality, but for our orientations and responsibilities toward one another. Though Nāgārjuna’s Madhyamaka and classical Nyāya are fierce opponents of one another, I find in both systems of thought philosophical insights of lasting and shaping value.
To return to the larger topic of this review of my journeys in classical Indian thought, I can say that it is those journeys, primarily, that have facilitated one of the great “conversions” I have undergone in my philosophical career. I have in the course of my three decades of interest in Indian thought been changed from an idealist in search of an underlying metaphysical unity of all beings tino a qualified realist who values the heterogeneity and plurality of the world. But that heterogeneity and plurality is tied together not by a nebulous, mysterious unity, but instead by the complex but discernable bonds of causal inter-relation. That inter-relation can indeed be seen as an even stronger basis for compassion than an abstract fundamental unity. This is what I have learned, and continue to learn, from the Indian philosophical tradition.
But, happily, my journeys continue, and may easily lead to further realizations and conversions. Learning never ceases, and changing one’s mind is not a defect, but a sign of growth. And, even more happily, the Indian philosophical tradition is replete with incredible resources for a plethora of insightful philosophical debates and views. And that is why it will always remain a wellspring of ideas and practices for philosophers in the future.
What a great post, Doug. I appreciate the philosophical autobiography of what brought you to this way of thinking, especially with the glimmers of practical implication involved in it (“and this vindicates the heterogeneity and complexity of existence.”)
Many thanks for this post and its optimistic outlook, especially concerning the dynamism of thought.
I hope you will tell us more (here or in a following post) about how you synthesise the Nyāya realism as for complex phenomena and Nāgārjuna’s emphasis on interconnected causation.
Thanks, Elisa, for a fantastic suggestion regarding an attempt to synthesise the validation of compound and real phenomena in Nyāya with the notion of interdependent causality in Nāgārjuna. Indeed, I think there may be a way to create strong mutual support between these ideas.
My next post was going to be about what I see as some challenges in the field currently and some ways we might go about addressing them. And the synthesis you inquire about might be too detailed for me to work out in a reply. But it’s a really stellar idea for an essay! So I might put it in that form soon!
My question below might be related to this, if parts too causally depend for their existence on wholes. Nagarjuna devotes an entire chapter on this in the Mulamadhyamakakarika, I believe. Unfortunately, don’t have time now to check, but I confess I don’t get most of Nagarjuna’s arguments. (The only way I can make sense of them is an uncharitable one.) Hope you can comment on this too.
Hello Douglas, enjoying these posts immensely, and I can relate to some of your experiences.
Reductionist and eliminativist arguments were all the rage back in the 2000s when I was learning about Western metaphysics. For instance, there is Trenton Merricks’s elimination of ordinary midsized objects via the causal exclusion argument, which he adapted from Jaegwon Kim. It seems to me that many of the Buddhist rivals of Nyaya would have liked this argument. Indeed, a key principle Merricks relies on is the so-called Alexander’s dictum, “To be is to have (nonredundant) causal powers”. I recall that Dignaga or Dharmakirti subscribed to a similar dictum.
I wonder, then, if we find something similar to the causal exclusion argument in Dignaga or Dharmakirti, and how Nyaya philosophers did/would/could have responded to this argument as nonreductionists about wholes.
Many other things I’d like to comment on, but will cut it short here for now.
Glad you are enjoying the posts. Thanks for your excellent questions.
Certainly the Dińńaga-Dharmakīrti school would insist on a focused analysis of what has the causal capacity to produce an effect (arthakrīyasamarthya). But their focus in this respect is on perception. They would insist that only a uniquely particular (svalakśana) cause, and not a general feature which is posted only in concepts, could bring about a sensory experience. Their notion of exclusion (aphoa), on the other hand, informs their philosophy of language, according to which the meaning of a general term is only arrived at by excluding other things from the reference of that term, but that its reference cannot be further ascertained than that.
I take the arguments of Kim and Merricks to be somewhat different–focused as they are on how we characterise material phenomena. In Merrick’s formulation, physical events can be reduced to their constituent elements, such that it is causally redundant to attribute the shattering of a window to a baseball when in fact that shattering of the window is cased only by the ball’s atoms acting in concert. This kind of argument is a little more redolent of early Buddhist philosophical attempts to distinguish between asamskṛtadharma-s (uncompounded, that is, “simple” phenomena) and samskṛtadharma-s (compounded phenomena), where all compounded phenomena can be analytically reduced, in terms of real existence, to uncompounded phenomena. O course, the Merricks-Kim arguments have profound implications for perception and philosophy of mind in general, as for them all mental states, though material, can be considered nothing more than the interactions of their most elemental causes.
In the modern context, it seems to me that reductionism is, at least currently, more popular among philosophers of science than it is among scientists. Of course, a neuroscientist often looks for the most basic material causes of mental events. But that doesn’t mean, for example, that a neuroscientist would hold that, in order to understand a behaviour, it is only the neurons, transmitters, receptors, synapses and signaling that matter. To understand our behaviours, we also must understand how neural information is encoded, how it is retrieved, and how it interacts with other neural structures. A physicist, in like manner, would certainly agree that baseballs are fundamentally made of atoms–and subatomic particles. But in order to understand the physics of the shattering window, it is not just the electron resistance between the atoms of the ball and window that matters, but the mass and velocity of the entire ball and the structure of the entire window that would determine the causal outcome. A physicist might even quibble with the way that the reductionist philosopher characterises the causal closure of the physical world. Why reduce everything to matter when matter itself can be considered a certain dynamic configuration of energy, and when mass itself merely results from the way a false vacuum alters the spin of certain originally massless field foci? Even in modern particle physics, then, reductionism as it has been philosophically formulated doesn’t seem to fit the bill.
And this gets to your question about Nāgārjuna in your other post, which I’ll respond to here. Nāgārjuna does have a chapter on the skandha-s or “aggregates” of Buddhist analysis. But actually the chapter itself, and other verse clusters in the MMK, argue against the tendency of early Buddhists to reduce the person to their skandha-s. His argument is basically that, even if one wants to consider skandha-s the causes of “form,” the skandha-s in that case cannot be understood in isolation from the forms that they go into producing. Therefore, the cause-of-form (the skandhas) and the forms they create are inter-dependent and hence “empty,” neither has causal autonomy (svabhāva). Madhyamaka can then be considered an opponent of rival forms of Indian Buddhist reductionism, both early and later, as can be seen in Candrakīrti’s critiques of Dińńaga.
Reductionism is an approach that is of immense value in explaining the structures and minute interactions between phenomena, it seems to me. But both in the classical Indian and modern Western contexts, I’m not convinced the approach serves as a supervening explanation of everything. On top of that, I still see no compelling reason to believe that, just because compound things can be analysed into elemental parts, those compounded things, once constituted, should be considered less real than the parts. Baseballs might be made of atoms, but only a certain configuration of atoms can make a baseball, and once made, the baseball is a thing to be reckoned with–as most windows would attest.
These are, at any rate, fascinating debates. And the perspectives offered on them from many traditions are uniquely informative.
Hi Dr Berger,
Thank you for the fascinating posts and insightful discussions! Reading through the question by Boram Lee I was struck by a question in the same vein:
Specifically, this has to do with the term “अनन्यथासिद्धि” in the definition of causality found in some Nyaya texts.
For example, the phrase “अतिरिक्तमथापि यद्भवेत्” (कारिकावलि v.20) seems to capture the idea of causal exclusion fairly closely. Ie, if the relations between the parts are sufficient to account for the putative effects of the composite, it seems like the composite, being in addition to an already detailed sufficient cause would be “अतिरिक्तम्” and therefore causally redundant–अन्यथासिद्ध.
Moreover, the second category, given be the phrase “कारणमादाय यस्य” (कारिकावलि v.20), also seems to preclude a composite from being a cause of effects that are already accounted for by the actions of its parts. After all, the composite is a necessary antecedent to the effects of the parts when arranged in particular relations only because it is a necessary effect of the parts so arranged. In this sense. it seems like it would fulfill the category of being “आदायैवान्वयव्यतिरेकौ”–its necessary concomitance with its effect being merely secondary to its relation to its parts.
This leaves me with two questions, one comparative and the other internal to Nyaya. 1) Would you consider the notion of अनन्यथासिद्धि as being comparable to the idea behind the causal exclusion argument and 2) do Naiyayikas think composites can be causally efficacious and if so how do they think composites fulfill the अनन्यथासिद्धि constraint?
Thank you for this detailed and fascinating question.
I am at something of a twofold disadvantage here, given that 1.) practically all my books are in my office but I am stuck at home during the COVID-19 mitigation measures and don’t have access to them, and 2.) I am not as well-versed in the Kārikāvali as I am in other texts. So I would want to have a more detailed look at these passages before I give a conclusive response.
By way of a brief reply to your questions, however, I will say the following for the moment.
To your first question, it is certainly the case that Nyāya philosophers had a robust conception of anyathāsiddhi, and were quite keen on flagging cases in which mere antecedents to an effect were wrongly cited as the causes of an effect. So, yes of course there is a conception of a variety of “casual exclusion” in Nyāya thought.
To your second question, this problem of what actually causes our immediate perceptual awareness came up repeatedly in the part-whole debates Nyāya philosophers were having with Buddhists. The Naiyāyikas had to concede the point that the causes of our direct sensory awareness of phenomena were the parts of a phenomenon. We do not see every side or feature of a tree at once, but see only parts of its trunk, branches and leaves that meet our eyes, in this case. So the whole of the tree, the composite entity, does not cause our perceptual awareness, even though it is an antecedent condition of our perception. But this admission did not lead classical Naiyāyikas to abandon the reality of the whole, or the composite substance, because branches and leaves could only be branches and leaves to the extent that they inhere in the tree. Dharma-s are, for them, always located in the dharmin. Now, it is certainly possible to launch serious critiques against the Naiyāyiaks on this matter–one could go on insisting, as the Buddhists did, that this admission that wholes do not immediately cause perceptual cognitions calls their entire commitment to wholes into fundamental question. I doubt I could resolve all of these complex and contested issues to satisfactory detail in a blog post reply. I merely bring it up to point out that the anyathāsiddhi problem as it pertains to our perceptual awareness would not have compelled that Naiyāyikas to abandon their otherwise thorough commitment to the robust reality of composite phenomena.
In any case, once I have full access to my library and have had a chance to mull over the specifics of your question, I may reply further. Thanks much for the fascinating questions.
Thanks so much for the response!
As you suggested, one of the main concerns for me when thinking through this was to what extent Nyaya would be able to respond to the sort of attack Dignaga pushes in the Alambana Pariksha. Much of the secondary literature on the subject seems to focus on the Sauntrantaka and Vaibhashika as Dignaga’s main targets but it seemed to me that he had a general audience in mind.
It’s easy to get the impression, though, from the focus on the Buddhist opponents, that this argument only bites against reductionists. Hence my question. Since if, as you suggest, Nyaya concedes that composites are not the direct causes of sense experience, it does not seem, after all, like they have any advantage over reductionists like the Sauntrantakas and Vaibhashikas with regards to this line of attack. I suppose they would have to push back against the premise that the perceptual object must be a cause of the perception. But, not only does the causal condition on the perceptual object seem fairly compelling in its own right, but this would also seem to wreck havoc on the traditional Nyaya account of pratyaksha as इन्द्रियार्थसन्निकर्षोत्पन्नम्.
In any case, this is probably not the sort of thing that can be simply resolved, but thanks for all the explanations and thought-food!
I wonder, though, if the argument about what directly causes sensation is dispositive of the entire issue of the existence of composite things. Just because only certain features of a thing are in contact with the sense organs at any one time, does that warrant the conclusion that only those features of a thing exist? There may, for example, be antecedent conditions that do not directly contribute to a certain causal process, but that does not entail that the antecedent conditions do not exist at all. One might, for example, extend this consideration to the issue of partially-locatable features. Does the fact that only parts of an animal’s body are in contact with only parts of a tree entail that only those parts of each phenomena that are in direct contact exist, so that there is really no animal and no tree? The Naiyāyikas appear to have been quite comfortable conceding that direct sensory contact usually only casually involve certain features of a thing at a time without believing this fact had such devastating entailments for the existence of composite things. The Buddhists, on the other hand, rule out any other considerations of actual existence by definition, through their emphasis on direct causal conditions alone.
Major and ongoing debates, it seems to me, usually only start with the drawing of important distinctions, rather than end there. Just as the Naiyāyika would concede the point under discussion, for example, the Dińńaga and Dharmakīrti Buddhists would concede that we have cognitive experiences distinguishable as varieties of pratyakṣa which were structure-less and structured. But it is only then that the debate starts, and does not end, regarding how to account for the processes of and evaluate the reliability of structured perceptual experience. And it is precisely because of the fact these debates are only deepened and not decided by important distinctions that the Indian philosophical tradition provides us with such endless resources for thinking about these issues.
Thanks for your contributions here, Doug. I am in much sympathy with your own story, and attitudes toward qualified realism. If I may make a quick addition to the discussion with Pramod: Uddyotakara explicitly argues that composites exist over and above their parts precisely because they have causal capacities that the mere parts do not.
“Yarn is different from the cloth made from it, since the two have different causal capacities (sāmarthya), like poison and antidotes.” Under NyS 2.1.36 (A. Thakur ed. 236.18-21)
This is a more general statement of the view that composites are more than their parts, since they can act in a coherent way, being (for example), held and pulled merely by holding or pulling a part of the whole (NyS 2.1.35).
That entire section reflects on the question of perception and composites being a cause of perception, and Uddyotakara spends much effort in responding to arguments advanced by Dignāga and other critics of composite wholes. One crucial part, not always highlighted in this regard is his defense of conjunction as a real property.
“You (a Buddhist reductionist) suggest that things are nothing more than a particular arrangement of their parts. But is this particular arrangement itself something distinct or not? If you say “yes” than there has to be something other than the parts themselves, and you must tell us what it is. If you say “no” then there is no other thing, and your talking about a particular arrangement is meaningless babble. For us, in contrast, conjunction accounts for this particular arrangement, and is itself something else.” (under NyS 2.1.33, A. Thakur ed. 210.1-4)
Hello Matthew, this is extremely interesting. Where should I look in Uddyotakara for these arguments, in the Nyayavarttika?
As for the quoted argument, I am a bit underwhelmed by it, to be honest. Since samyoga is a quality, Buddhist reductionists do not have to admit to there being anything (= any dravya) other than the parts. Samyoga is simply the quality of the parts themselves. (Unless, of course, the Buddhists in question deny that these things have qualities).
(I am assuming that a whole that inheres in its parts is a dravya. As I understand it, the atoms are nitya-dravya, and complex entities composed of atoms are anitya-dravya.)
The entire passage of Nyāyasūtra 2.1.33-36 focuses on this issue. You can find translations of it in the first volume of Ganganatha Jha’s translation, along with Uddyotakara and Vātsyāyana’s comments. If you can find it, a currently out-of-print translation by Mrinalkanthi Gangopadhyaya has all of Vātsyāyana. Stephen Phillips and my recent select translation of the Nyāyasūtras discusses these in chapter 5, Substance and Causation. We have tried to select the most important parts of their works, and some snippets from Vācaspatimiśra.
If you read Sanskrit, you can see the page number and line numbers in the editions by Anantalal Thakur in parentheses above.
FWIW, I think the argument is stronger than you do. If one tries to reduce a composite by saying it is just a certain arrangement of parts, then that arrangement is itself doing some real philosophical work. If one admits that it is, then one is close to allowing properties, and probably relations, into their ontology.
Thank you very much for pointing me towards the sources. I have been meaning to get your book. Jha’s translation with Uddyotakara’s commentary cannot be found online, same with Gangopadhyay.
Anyway, I certainly think that claiming wholes have causal properties that their parts do not is the way to go. The puzzle is how to do this without endorsing a kind of downward causation that violates physical closure, i.e., without introducing new forces above and beyond the fundamental forces of nature that account for all physical event. I hope to find some promising directions for solving this problem in the Nyaya texts.
Hi Pramod, thank you for drawing attention to Kaarikaavali. I downloaded the text and tried to understand the relevant portions (it is difficult for me!).
Question (2) in your first comment is important and I am very much interested in it. Anyhow, I will try to answer your Question (1).
Ananyathaasiddhi comes close to the idea of “redundant causation” used in causal exclusion arguments, but the Nyaaya concept is broader, e.g., it includes the cause of the cause of an effect (the father of a potter who makes a jar, which is admitted as the cause of the potter but excluded from being the cause of the jar). But redundant causation (or overdetermination) cases typically discussed in the West are simultaneous events each of which can apparently bring about the effect without the other.
Also, a causal exclusion argument seems to me to have the following three- or four-part structure.
(1) Redundant Causation: C1 in some supervenient domain D1 is the putative cause of E1, and C2 in some subvenient domain D2 is also the putative cause of E1 by being the cause of E2 (which subvenes E1).
(2) Causal Closure: Given the laws governing D2 and the principle of supervenience, the occurrence of C2 is sufficient for the occurrence of E1 and E2.
(3) Causal Exclusion: Because C1 is redundant as a cause of E2, C1 is not a cause of E2.
(4) Alexander’s Dictum: To be is to have causal powers, therefore C1 does not exist.
Here, ananyathaasiddhi looks partially relevant to (1) and (3), but we will need to establish other correspondences to be able to reconstruct the entire argument in Nyaaya terminology. E.g., instead of “supervenience”, we could use “inherence” (samavaaya). And at Step (3), we could appeal to parsimony considerations (laaghavatva).
It will be interesting to see if the causal exclusion argument can survive this translation into Nyaya terminology and modes of argumentation. (For my part, I am struggling through the causation and inference sections in Tarkasangraha, so I am far off from being able to provide even a decent reconstruction.)
(3) Causal Exclusion: Because C1 is redundant as a cause of E1, C1 is not a cause of E1.
It also occurs to me that many analytic philosophers would regard causation as transitive: a cause of a cause of an effect is also a cause of that effect. But the father of the potter who makes a jar according to Karikavali is ananyathasiddhi and therefore not a cause of the effect. It seems to me that the Nyaya position here is far more sensible, but I am veering off topic so will stop.
A fascinating post, thank you Douglas, and my thanks for the resources that this blog provides.
If I might just mention, with regard to Schopenhauer and Vivekananda, you say: “…a move that, as Paul Hacker showed, had much influence on Vivekanada through Paul Deussen.”
I would just point out that it has been demonstrated by both Gwilym Beckerlegge and Thomas J. Green that Vivekananda formulated this ethic before he knew of Paul Deussen. Its a minor footnote to this interesting post, but I would encourage readers to finally move past the tropes of Paul Hacker’s work in this area.
Thank you for the references, Mark. I will look them up.
Douglas, I very much appreciate your detailed response, thank you!
I myself am a nonreductionist about ordinary mid-sized objects and the natural kinds around us. Especially the natural kinds, which I take to have their own nature (svabhava) or proper function determined by nomological necessity or evolutionary history. Due to the combined influence of Putnam, Kripke, Millikan, Aristotle, etc., I guess.
Nonetheless, I feel that there are several very good philosophical arguments in favor of reductionism or eliminativism. The most challenging two IMHO are the causal exclusion and sorites arguments. I can’t figure out exactly how these arguments fail to work, and no amount of table thumping, foot stomping, and stone kicking can make them disappear. Which is one reason why I’m hoping to learn more about what the Naiyayikas have to say.
I do not expect the beliefs of the scientists to give us solutions to the philosophical problems that vex us. Scientists specialize in this or that domain, but when they step outside of it to tell us how one domain relates to the other, they are not much better than the rest of us amateurs. How one domain of expertise relates to another is the philosopher’s niche, partly because it lies in the gap between domains of expertise and no one is a qualified expert in it, and partly because it is the philosopher’s job to show how all the domains hang together as a whole. That’s what the Nyaya-Vaisheshika system seems to be doing with the 9 padarthas.
As for the ultimate constituents making up the physical world, your point is very much valid. However, Merricks easily sidesteps this issue by stipulating that “atoms” in his argument is just a placeholder for whatever the final physics of the future says about it. For all we know, there are no ultimate constituents and we can keep splitting down ad infinitum. Even then, it appears that his argument will work.
Thank you for your explanation of Nagarjuna. Aside from the skandha-s, there is also the chapter on fire and fuel, which according to Garfield contains Nagarjuna’s criticism of one-way dependence espoused by some Buddhists.
My worry here is that Nagarjuna often uses correlative concepts or terms that depend on each other conceptually or by definition. But the fact that we cannot understand or talk about the one without the other only means that there is conceptual or terminological dependence between them. It does not mean that one cannot exist without the other.
To use a toy example for illustration, assume that husbands can exist without wives. But that’s absurd. If there are no wives, there can be no husbands, and vice versa. Therefore, husbands depend for their existence on wives, and vice versa.
All that this toy argument establishes is conceptual or terminological interdependence of “husband” and “wife”. The men and women marked out by these terms can certainly exist without the other. How does Nagarjuna make the leap from conceptual/terminological interdependence to ontological interdependence?
Now of course this is an extremely uncharitable parody of Nagarjuna. But it might apply to at least some of his arguments. E.g., if the Fire and Fuel chapter used the word “wood” instead of “fuel”, I suspect some arguments may not work (e.g., fuel cannot be fuel without fire, but wood can certainly be wood).
There is no doubt a rich variety of arguments in the Mulamadhyamakakarika, and the correct reading must be extraordinarily complex. So I would appreciate any recommendation on which translation or commentary (I’ve read Garfield) is the most helpful in removing the kind of worry I have.
Hi again, Boram. Thank you for all the fascinating posts.
I hardly think all of Nāgārjuna’s arguments, whether in the MMK or Vv, are convincing. His arguments in the MMK often address different issues; some of them are about conceptual correlativity, some about causal powers, and some about identity and difference.
In terms of the chapter you talk about above, indhana is actually “wood,” but more specifically in this chapter it is treated as “kindling.” In this chapter, Nāgārjuna is primarily concerned with whether or not fire and wood-as-kindling are the same or different from one another (satattvaṃ ye bhāvānāṃ ca pṛthak pṛthak). Most of the chapter’s argument, however, is focused on establishing that fire cannot be fire without kindling, since then there would be no causal account of fire being ignited. The arguments about kindling are not so much arguments entailing that wood could not exist without fire, but that, in order for wood to be kindling, it must be ignited by some causal factors. To give an account of burning kindling, then, an analysis of the sameness or difference of fire and wood will not do, because for that account we need to explain their causal interaction (prāpsyate). He actually alludes in this same chapter to a man and a woman, but in Vv. his preferred example is father and son, and in both cases, if we want to explain these pairs *as* “husband and wife” or “father and son,” we need a casual story that accounts for the mutuality in the relation. These kinds of causal accounts will not be simple, linear ones–they will not be clarified at all by mere reference to svabhāva or parabhāva–, but instead will be complex. One might object that these relations are about conceptual correlatives, but that seems to me to be part of the point, namely that our relationships and the activities we undertake in those relationships do involve social norms. Our lives are not just about physical interaction, but are also about interactions within the purview of vyavahāra. The need for causal accounts that explain mutual arising is why, at the end of the Vv, Nāgārjuna feels justified in pointing out that his epistemological critiques in that work are not intended to annihilate the notions of pramāṇa and prameya entirely, but to elucidate the fact that a specific “pramāṇa-prameya” complex arises from causal interaction and cannot be accounted for through inherent sameness or difference. Now, whether one finds Nāgārjuna’s arguments here compelling–or even pertinent–or not, I think his intention almost all of the time is to undermine causal accounts that base their validity on the supposed svabhāva of individual phenomena.
As far as translations are concerned, as someone who has studied Nāgārjuna a lot, I am probably not completely satisfied with any one translation. If I had to pick one English translation to prefer, I suppose it would be the one by Siderits and Katsura. So far, the piece of secondary literature on Nāgārjuna I like the most is Joseph Walser’s Nāgārjuna in Context, which for the most part is excellent, though I think at times he places some of Nāgārjuna’s arguments far closer to those of the Pudgalavādins than I would.
Thank you very much for this exceedingly clear explanation! The “father and son” analogy from Vv is much better than the one I made up about husbands and wives, since, as you point out, there is a causal story to account for the terminological correlation. Likewise for “indhana and fire”.
And thank you for the recommendation of Siderits and Katsuna. I will have to look very carefully at how Nagarjuna handles examples like “father and son” as I read through their translation. (Sometime in the future, since I am now focusing on Nyaya and Vyakarana).
Some musings below on the examples, which you can skip because they are largely speculation at this point. I will certainly have to wrestle with the book itself.
In each these relational pairs (father-son, indhana-fire), one of the pair seems to me to have a stronger claim to being causally independent of the other. The man who happens to be a father produces the son, but could have existed without producing a son (causal independence). Of course, the conventional designation of the man as “father” depends on the existence of the son (conceptual/terminological dependence), by vyavahaara as you say.
Another possible option is that the man could not be the very same man, with all his particular features, experiences, and relationships, if the son did not exist. Here we are individuating the person so finely that any change in his features, experiences, or relationships would have produced a different person. A fragile entity, we might say.
Nicholaos Jones (“Nyaya-Vaisheshika Inherence…”, 2010) adopts a similar approach in trying to come to terms with the Huayan idea that the entire house depends for its existence on any one of its parts, for instance on the rafter. He adopts mereological essentialism, so that a change in any one of the parts destroys the whole and creates a new one. Adopting fragile entities here might be a more general approach. Then it can be shown why the rafter depends on the house as well as other parts of the house. (Being a naturalist myself, I do not like admitting such fragile entities, since nature does not seem to be carved out that way, especially at the biological level, where there are enduring hylomorphic wholes that can survive inessential changes.)
Thank you all so much for the detailed and enlightening responses. To simplify things, I’ll just respond to everyone here–and I apologize in advance for rambling on
I definitely take your point that the distinctions and arguments drawn here are only the start of the conversation. And I agree that the argument in the Alambana Pariksha does not establish that wholes do not exist. But I also don’t think that this is the goal of the text in the first place. As I understood it, at least, it seemed like what Dignaga wanted to do in the AP is merely to explore certain constraints on what perception can be about and what knowledge it can generate. And I feel Nyaya, precisely because of its especially strong conception of perception and the substantive role that this plays in preventing their empiricism from collapsing into skepticism, is particularly sensitive to any constraints that are placed on perceptual knowledge–so I worry that even if the AP’s conclusions have no direct metaphysical import whatsoever, its epistemological conclusions could still be quite devastating to the Nyaya project if left unanswered.
Thanks a lot for this. I will admit that the first time I read Uddyotakara’s discussion of samyoga and his causal argument for the distinction between composites and their parts (which was actually in the awesome Nyaya Sutra translations by you and Phillips, if I recall!) I was not entirely sure what to make of it.
I still feel that way, I think, just because I’m not so sure it works well as a response to Buddhists who don’t seem to believe that the cloth and the thread are the same thing anyway (recall, for ex the passage in Milinda that denies the identity of the chariot with its parts). Moreover, the claim that samyoga must exist as a distinct entity or we wouldn’t be able to use it to “explain away” composites, while an interesting point, won’t worry Dignaga so much, I think, since he has apoha to explain how referentially empty expressions can still be meaningful.
Still, I suspect Uddyotakara hasn’t spoken his last about this and I’ll still have to think over it, I’m sure.
Thanks for the clarification about the structure of the causal exclusion argument. I think you’re right that the notion of anyathAsiddi is broader than the principle of causal exclusion and is meant to account for all cases where necessary antecedence != causality.
As for the specific argument structure you give in outline. I think that Nyaya might actually reject (1). Specifically, they might argue that C2 is actually anyathasiddha with regards to E1 because it is the cause of E2 directly and only the cause of E1 via its relation to E2, which would fall under category (2) in the kArikAvali:
अन्यं प्रति पूर्वभावे ज्ञाते यत्पूर्वभावविज्ञानम् (K v. 19)
Which in the svavritti is further explained as:
अन्यं प्रति पूर्ववृत्तित्वम् ग्रिहीत्वा एव यस्य यत् कार्यं प्रति पूर्ववृत्तित्वं गृह्यते तस्य तत् कार्यं प्रति अन्यथासिद्धत्वम् |
In other words, since C2’s necessary antecedence to E1 is only established on account of its necessary antecedence to E2 on which E1 supervenes, it is anyathAsiddha with respect to E1. This, in effect, reverses the direction of the redundancy in (3)–securing the distinctive causal role of composites.
This ties into Matthew’s very helpful citation where Uddyotakara speaks of the cloth having distinct causal capacities from the thread it inheres in or–even more evocatively–that poison has distinct causal capacities from its molecular constituents.
To draw this out more–consider the case of the poison again. When a person injests poison and gets sick–we can say, on the on hand, that the poison caused their illness. On the other hand, we can say that the molecular components of the poison caused a drangement in the patient’s biochemistry. Moreover, we can say that their illness supervenes on the deranged bichemistry (or, alternatively though perhaps not equivalently–the one is a samavAyi-kAranam of the other).
The poison is a nimitta-kAranam of the illness. The deranged biochemistry is a samavAyi-kAranam of the illness. The molecular constituents of the poison in causing the biochemical derangement form the nimitta-kAranam of the illness’ samavAyi-kAranam. But does that make the molecules of the poison a nimitta-kAranam of the illness as well? Perhaps not.
What we have here are two separate causal stories, one at the macro level and another at the micro level. Each is internally coherent and relates to the other via the supervenence-subvenance relation. But there is some danger in collapsing the two narratives and allowing for causal language to cross the boundaries between levels. I’m thinking here, ironically enough, about Mark Siderits’ discussions on Buddhist paleo-compatabilism. Whatever the precise textual-historical merits of this idea are, I think he makes an interesting case arguing that when we ascribe (instrumental) casual roles to objects at the ultimate level for events occurring at the conventional level, we run the risk of illegally crossing explanatory levels. The only difference between this Buddhist account and one that may be more appealing to the Naiyayika is that, for the Buddhist, the language of composites is merely conventional while for the Naiyayika it is just as real as the talk of the parts.
Now, is this a compelling account of the causality of composites? I’m not sure. I think it does suffer from issues of economy, at least, and is probably counterintuitive as well. It may also still be susceptible to Dignaga’s epistemological attacks in the AP among other things.
Pramod, thank you for this detailed discussion, it is very helpful.
It now seems to me that the causal exclusion argument can be set up in the Nyaya system as well, given its account of causation.
I was mistaken in thinking that supervenience would correspond to the Nyaya concept of inherence (samavaaya). Supervenience is meant to be an asymmetric dependence relation that entails bottom-up determination (though the technical formulations for it could never capture the intended dependence/determining relation), and in the Nyaya system the concepts of inherent and non-inherent causation (samavaayi- & asamavaayi-kaarana) play that role. And this makes the causal competition at issue in Step (1) of the causal exclusion argument even clearer!
By way of illustration, let a red jar be the effect, the potter’s wheel and kiln on which it is produced nimitta-causes, and the atoms composing these things the samavaayi-causes.
By way of illustration, let a baked red jar be the effect, the fire in a kiln the nimitta-cause of that effect, and the earth-atoms and fire-atoms composing these be the samavaayi-causes.
Then we have two apparent causal chains leading from the these causes to the effect. First, we have the causal chain at the macrophysical level connecting the fire to the baked red jar. Second, we have a complex causal chain at the at the microphysical level: the tejas-atoms coming into contact with earth-atoms and rearranging them as the samavaayi-cause of the baked jar, and changing their color to red, this becoming the asamavaayi-cause of the red color of the baked jar.
So, we have the two causal chains leading to the same effect, one from macro-cause to macro-effect, and the other from micro-cause to micro-effect that in turn causes the same macro-effect. Thus, redundant causation.
Now we can ask, does the macro-cause really bring out the macro-effect all on its own? Or does it do so only via the underlying micro-causes? This is enough to get the causal exclusion argument underway, along the lines indicated by Kim and Merricks.
I think you are right that the causal exclusion problem can be expressed with the conceptual apparatus the Naiyayakas had available to them, but I do not think the problem itself actually arises in their system. This is because the way they understand anyathasiddhi actually blocks objects at the micro-level from being nimitta-causes of macro-level events–preventing them from competing with macro-level objects as causes of macro-level effects.
Consider your example of the micro-level description of a clay pot being baked into terracotta. You express the causal exclusion as it arises in this case with:
> So, we have the two causal chains leading to the same effect, one from macro-cause to macro-effect, and the other from micro-cause to micro-effect that in turn causes the same macro-effect.
However, there is a problem in the second causal chain from “micro-cause to micro-effect that in turn causes the same macro-effect”. Namely, the micro-effect and macro-effect compete as effects for the same cause. This is anyathasiddhi.
There are a couple noteworthy points here. The first–which I only just realized as I was revisiting some of the literature on anyathasiddi–is that a crucial difference between anyathasiddhi and the causal exclusion principle is that whereas the latter concerns competition between causes, the former generally involves competition between effects. This is even reflected in the name “anyathasiddhi” which refers to a cause that is “already accounted for”. Take this particularly lucid description in Keshava Mishra’s Tarka Bhasha:
> तन्तुरूपस्य तु नियतः पूर्वभावोस्त्येव, किन्त्वन्यथासिद्धः, पटुरूपजननोपक्षीणत्वात्. पटम् प्रत्यपि कारणत्वे कल्पनागौरवप्रसङ्गः
> The color of the thread, though, is an invariable antecedent [to the cloth]. However, it is anyathasiddha because it[‘s causal role] is exhausted in causing the color of the cloth. If we consider it to be a cause of the cloth, too, then it would result in a multiplication of theoretical posits.
The Karikavali, as usual, gives a more detailed analysis of precisely when multiple effects compete (versus a genuine case of multiple effects) and how to distinguish genuine effects from cases of anyathasiddhi. In this specific case, the relevant category is given (as in my earlier comment) by:
> अन्यं प्रति पूर्वभावे ज्ञाते यत्पूर्वभावविज्ञानम् (K v. 19)
Namely, If the causal relation between C1 and E2 can only be apprehended in dependence on the apprehension of a separate causal relationship between C1 and E1, then C1 is anyathasiddha with respect to E2. In other words, only an immediate effect is a true effect, mediated effects are blocked by anyathasiddhi. So, if we have three objects C1, E1, and E2–where C1 causes E1 and E1 causes E2–then C1’s relationship to E2 is not independent but mediated by E1, causing anyathasiddhi.
In our case, the causal relationship between the tejus-atoms and the terracotta pot is mediated by the causal relationship between the tejus-atoms and the earth-atoms. Here, again, we have two competing effects for the same cause. The causal role of the tejus-atoms in this story is, to use Keshava Mishra’s language, exhausted by their causing the rearrangement of the earth-atoms, the proximal cause. We cannot, for this reason, extend the causal role of the tejus-atoms to the more distal terracotta pot as well, because they are anyathasiddha. So, the tejus-atoms are not causes of the terracotta pot, per Nyaya theory, and cannot compete for this role with the macro-level fire. As such, the causal exclusion problem does not arise. Note, this doesn’t interfere with causal closure because closure still obtains at each level respectively; however, anyathasiddhi insulates the two levels so that there can only be samavayi and asamavayi causal interactions but no nimitta causal interactions that cross explanatory levels.
To put it otherwise, anyathasiddhi ensures that we not skip over the crucial mediating role inherence relations play in the causal story connecting the micro and macro levels. Each individual element in the microscopic state only interacts with other individual elements, never the complex as a whole. So we cannot ascribe any direct causal role to any individual element regarding the whole–we can only do this at the macroscopic level by considering the one subcomplex of the whole causal complex (ie the fire) to be interacting with another subcomplex (the clay). The principle of anyathasiddhi serves to preserve this distinction.
As an aside, there is another response available to the Naiyayika regarding the causal exclusion argument as it pertains specifically to the mereological context (though not in the case of mental causation) which relates closely to Uddyotakara’s comments on NyS 2.1.33 (quoted previously by Matthew).
Causal exclusion can only occur between sufficient causes. However, no individual element at the microscopic level–the individual tejus atoms, for example–act as sufficient causes for the macroscopic object–the terracotta pot. Because, the other atoms etc in the microscopic story are also necessary for the final effect to manifest. The sufficient cause, at the microscopic level, can only be ascribed to the microscopic state or arrangement as a whole.
You concede the same in your earlier comment about my net example, when you state (on behalf on the reductionist): ““fish-containment” is the effect of the fibers arranged net-wise”. If any single individual net-wise arranged fiber were a sufficient cause of the fish containment, then it would compete with and exclude the other net-wise arranged fibers in the net! So, the fibers can act as sufficient causes only when taken collectively. But, in order for the collection of fibers to compete, causally, with the macroscopic net, the reductionist must ascribe genuine causal powers to the collection. But this is precisely what the reductionist wishes to deny via this argument! Presupposing what is to be disproven–this is pakshābhasa and is a point of defeat.
Last point, which concerns mental causation as well as mereology, is that Nyaya denies Alexander’s dictum (which they are aware of via Dharmakirti). I don’t have a specific reference at hand; however, a common formulation that we see repeated among Naiyayikas concerning the notion of existence is the following: “अस्तित्वं ज्ञेयत्वमभिधेयत्वम् “, which asserts the three-way equivalence between existence, knowability, and expressibility. Thus, even if the causal exclusion argument ends up being successful, it would likely not lead Naiyayikas to reject the existence of mental phenomena insofar as they would consider mental phenomena to be independently apprehendable via anuvyavasaya and expressible in language (there is, of course, no private language in Nyaya).
Pramod and Boram, a text came to mind that might interest you: Arun Ranjan Mishra’s *Nyāya Concept of Cause and Effect*. It is a historical survey from early Nyāya to later Navya Nyāya thought that may prove useful to your research.
To Pramod and Matthew.
First off, thank you for the reference Matthew.
As usual, thanks for your thoughtful and instructive reply. Much appreciated!
I will first note a confusion that we should be careful to avoid. The causal exclusion argument does NOT take microphysical events/entities/features as nimitta-causes of macrophysical events/entities/features.
Imagine, if you will, a path diagram, with nodes representing dependent or independent variables, and unidirectional arrows representing cause-to-effect or other one-way dependence relationships. So, for instance, we can have two mental events connected to each other by a horizontal arrow. Below that, we can have two nodes representing neurophysiological events, also connected to each other by a horizontal arrow. Then we can have two vertical arrows running from the neurophysiological events to the mental events to indicate that the mental events are realized by the neurophysiological events. This is enough to get Kim’s causal exclusion argument going. At no point does it require a diagonal arrow pointing from the first neurophysical event to the second mental event. Likewise for Merricks’s version.
This being the case, the causal exclusion argument likewise does not require a diagonal arrow directly connecting the tejas-atoms to the baked red jar. And nothing in my description of this particular example suggested (I hope!) that tejas-atoms should be a nimitta-cause of the jar itself. To repeat, the fire in a kiln is the macrophysical nimitta-cause of the baked red jar. The tejas-atoms that compose into the fire are the nimitta-cause of the rearrangement and recoloring of the earth-atoms that compose into the baked red jar. In the latter case, the rearranged earth-atoms conjoined are the samavayi-cause of the jar, and the red color of the earth-atoms are the asamavayi-cause of the red color of the jar (this is of course an abbreviated description, the actual causing from the atomic level to jar level will proceed level by level, through dyads and so forth). And I think this will meet the ananyathasiddha condition. After all, I am sticking very close to the three kinds of admissible causes and examples for them, and the usual description given for the baking (paaka) process.
That’s why I spoke of competing causal chains rather than causes when discussing the Nyaya example. A causal chain from C2 to E1 does not require a direct causal link between C2 to E1, only a series of direct causal links that have C2 and E1 at either end.
Below, please find my comments on other specific points that you make (put in quotation marks).
(1) Pramod: “[A] crucial difference between anyathasiddhi and the causal exclusion principle is that whereas the latter concerns competition between causes, the former generally involves competition between effects.”
This is a very interesting proposal, I will have to think more about it. My own impression is that anyathasiddhi is about screening off concomitants of the cause that are causally irrelevant to the production of some given effect. As I understand it, it does not rule out competition as such, whether between causes or effects. Thus I am inclined to think that the causal exhaustion principle mentioned by Keshava Mishra is a different and additional principle used to support anyathasiddhatva reasoning.
But let’s consider how the causal exhaustion principle might apply to our example. The tejas-atoms composing the fire do so as samavayi-cause of the fire, but their causal role is not exhausted by this effect. By coming into contact with earth-atoms, all the while preserving their macrophysical appearance as fire, they serve as nimitta-cause of the rearrangement and recoloring of earth-atoms. To generalize, samavayi-causes can also be nimitta-causes at the same time, even though their effects are different.
If Naiyayikas like Keshava Mishra wish to preserve the causal exhaustion principle, then this leads to two interesting possibilities. First, they could insist that tejas-atoms are a cause of the fire, or a cause of the earth-atom-rearrangement, but not both (this would lead to a different kind of causal exclusion argument involving competition between effects!). Second, they could provide a more fine-grained analysis of the causal exhaustion principle. so that tejas-atoms’ composing into a fire does not prevent their producing a change in the arrangement of earth-atoms, and their producing a change in the arrangement of earth-atoms does not prevent their producing a change in their color.
(2) Pramod: “Note, this doesn’t interfere with causal closure because closure still obtains at each level respectively.”
I would say that causal closure certainly does not apply at the macrophysical level, and this is especially true in the Nyaya-Vaisheshika view. Naiyayikas are atomists after all, and you can’t have macrophysical phenomena without atoms causing them. A constant refrain in the dravya portion of Tarkasangra is that eternal substances are the atoms, and non-eternal substances are their effects. So, compound phenomena cannot have just compound phenomena as their causes, but also require atomic causes. This shows that causal closure does not apply to compound phenomena.
That leaves us with causal closure of uncompounded phenomena. You can either accept it or deny it. If you accept it, then you are accepting that uncompounded phenomena cannot be caused by compound phenomena, and that brings you very close to accepting the epiphenomenal status of compound entities. If you deny it, then you have to show that downward causation is possible without violating the universal laws governing uncompounded phenomena. In terms of modern physics, you have to show how there can be downward causation without introducing new forces of nature aside from the four fundamental forces. This is what is involved, I believe, in showing that wholes have genuine causal powers above and beyond the causal powers had by their parts.
As a non-reductionist, I accept the second horn of the dilemma, but the daunting task of showing exactly how remains.
(3) Pramod: “[I]n order for the collection of fibers to compete, causally, with the macroscopic net, the reductionist must ascribe genuine causal powers to the collection. But this is precisely what the reductionist wishes to deny via this argument!”
When you arrange Lego-bricks house-wise, is there anything more than the bricks and the contact between them? Being in contact and having a house-like structure are all properties of the bricks themselves.
Atoms can come into contact, and through a recursive process starting from the base level of atoms, get compounded into larger structures. This seems to be an additive process. I have suggested in a previous comment that we cannot get emergent phenomena and emergent causality through an additive process. We only get more of the same. We need an account of how something qualitatively different can emerge. Perhaps the Naiyayikas have an account, but if not, my recommendation here is to accept modern Western science, especially chemistry, the molecular level at which qualitative changes start to appear.
(4) Pramod: “Last point, which concerns mental causation as well as mereology, is that Nyaya denies Alexander’s dictum (which they are aware of via Dharmakirti). I don’t have a specific reference at hand; however, a common formulation that we see repeated among Naiyayikas concerning the notion of existence is the following: “अस्तित्वं ज्ञेयत्वमभिधेयत्वम् “, which asserts the three-way equivalence between existence, knowability, and expressibility.”
Alexander’s Dictum is optional for the causal exclusion argument. If I remember correctly, Kim included it in his earlier versions, but excluded it in his later writings. The dictum is required for eliminativism, but not for reductionism (mental events just are physical events, compounded entities just are uncompounded entities arranged this way or that), which can be argued for on the basis of simplicity considerations.
Likewise for Naiyayika equation involving knowability, expressibility, and existence. This principle only guarantees that there are things corresponding to our terms and cognitions, but not necessarily that these things must have separate existence. It is consistent with reductionism, and Naiyayikas do engage in reductionist exercises, e.g., when they claim that certain pramana-s accepted by other schools are reducible to the four pramana-s accepted by the Nyayayikas.
Thanks for the response, Boram. Sorry for getting back to you late. My attempt to respond to your main point is below. The following comment includes a few remarks on some of the other points you brought up.
> I will first note a confusion that we should be careful to avoid. The causal exclusion argument does NOT take microphysical events/entities/features as nimitta-causes of macrophysical events/entities/features.
Thanks for this clarification, all definitely keep this in mind! Just to be clear, though, I want to say that anyathasiddhi blocks kAranatva in general, not just nimitta-kAranatva.
>Imagine, if you will, a path diagram… At no point does it require a diagonal arrow pointing from the first neurophysical event to the second mental event.
So, if I understand correctly, the causal exclusion works here by arguing that when mental event M1 causes mental event M2, the neurophysiological processes N1 subvening M1 causes N2 which subvenes M2. Here, both M1 and N1 are determinants of the occurrence of M2, but N1 is independently sufficient for the occurrence of M2. This is what gets the exclusion going–namely that one effect has two separate determinants, with at least one of them being a sufficient determinant, both occurring at the same time without being a case of true overdetermination.
The issue here is somewhat different than the mereological case we were dealing with, first of all, because in the latter, no individual element of the microscopic causal chain is itself a sufficient determinant of the occurrence of the macroscopic effect. We have the fire’s interaction with the clay, let’s call it E1; a set of interactions between n tejus atoms and m earth atoms, E2(i,j) for each integer i in [0,n) and j in [0, m); a set of resulting rearrangements of the earth atoms, E3(i) for each integer i in [0, m); and the formation of terracotta, E4. Then, E1 causes E4 and, for each integer i in [0,n) and j in [0, m), E2(i,j) causes E3(j)–for simplicity, I’m modeling this as a 1:1 interaction. Now, where is the single event with more than one sufficient determinant at the same time?
Perhaps we could respond by saying that the various interactions of the atoms together constitute a single event. But, is an event involving a collection of interactions between a set of elements different from an event involving an interaction between a collection of the same elements ? And if we assert this equivalence, does that commit us to a reductionism concerning collections?
At least part of the issue here is that Nyaya does not actually consider causality as occurring events but between objects, so translating these intuitions into talk of events is not unproblematic. In particular, I suspect it would be difficult to meaningfully translate the distinction I just introduced above into Nyaya language if we only deal with causal relations between objects and eliminate talk of “interactions” or “events”. Be that as it may, I do think a Naiyayika would likely say no to the second question. And, although I am not sure what one would say about the first, but I do think this response is open to them.
Consider the following (admittedly imperfect analogy). When my hand touches an object, I touch that object as well—because the hand is my hand. “I” participate in at least some of the events that the parts of my body engage in. The event of my touching the object and the event of my hand touching the object are not different events. Rather, they are different descriptions of the same event. But this does not itself mean that the descriptions of the objects that feature in each case are different descriptions of the same objects. Nor does it necessarily mean that one of the descriptions involves objects that are not genuinely real. It could also simply be that in each case some aspects of the event were left out of the description. This is the fundamental difference between these sorts of cases and those of genuine overdetermination. Because, when lightning strikes a straw house and, simultaneously, an arsonist lights it on fire, the lightning striking and the arsonist’s actions are different events, different interactions so to speak. The arsonist does not participate in the lightning’s interaction and the lightning does not participate in the arsonist’s actions. They are independent and, so, overdetermine the outcome. But the relation between part and whole as between self and body is different (or so the Naiyayika may respond, at least) since one participates in the events of the other.
Perhaps more to the point, I think that the Naiyayika could and would also point out that either way we must invoke the notion of a collection to get the argument forward—ie that in both events the actors are collections and not merely individuals, such that even if one of the events competes with and excludes the other, the causal role of collections is preserved. You could respond, further, by pointing out that not even Naiyayikas think all collections are real, so the fact that a collection is involved in this case does not itself prove that the collection is real. But, the Naiyayika would respond that this is true–but, nonetheless, the fact that a collection participates in both events implies that, if the collection in the uneliminated event is real, its causal role cannot be eliminated. In other words, this example does not decide between reductionism and non-reductionism—which is really what is at issue here since the Naiyayikas are really just the defendants in this debate.
In addition to the previous comment, a couple remarks on some of the other points you brought up, which I sectioned of in a different comment from above for the sake of clarity:
> I would say that causal closure certainly does not apply at the macrophysical level, and this is especially true in the Nyaya-Vaisheshika view.
Ah, this is my bad, I should have been clearer here: I was only speaking of the closure of nimitta-kArana. All macroscopic objects have, by definition, non-macroscopic samavAyi-kAranas. Also, I do not mean to give the impression that I hold to a general closure principle for the macroscopic case. Rather, what I wanted to suggest is that something like a macroscopic closure principle obtains in a particular class of cases—specifically where the following conditions are met: (1) a macroscopic process can be represented as a network of causally interacting objects at the micro-level, (2) every object in a cause-effect pair (ie every nimitta-kArana-kArya pair) at the micro level subvenes a distinct object at the macroscopic level and (3) the supervening macro objects of such a micro causal pair themselves form a causal (nimitta-kArana-kArya) pair. In this case, every macro effect will have a macro cause and, due to anyathasiddi, macro effects will have only macro causes. That being said–thinking it over a bit–I don’t think this is actually enough to yield a closure principle: consider a projectile propelled by a chemical explosive. Here, the micro-state fixes the movement of the projectile (not considering QM for simplicity) whereas the macro state may not—ie we may not be able to predict the path of the trajectory based on just macro level properties. In this case, the specific trajectory seems like it does have some sufficient cause (since it can be determined, in principle, from the initial state) but it has no sufficient macro level cause. Perhaps, though, we can ameliorate this by arguing that it is not actually that the properties of the macro state do not determine the trajectory but merely that the properties are not known with sufficient detail.
> That leaves us with causal closure of uncompounded phenomena. You can either accept it or deny it. If you accept it, then you are accepting that uncompounded phenomena cannot be caused by compound phenomena, and that brings you very close to accepting the epiphenomenal status of compound entities.
I do hold that causal closure obtains at the uncompounded level–to deny this would break conservation laws, I think. However, I do not think this implies that compounded entities are noncausal. Because, when clusters of microscopic entities are required to cooperatively generate a single macroscopic effect, then it becomes impossible to explain how uncompounded phenomena give rise to compounded phenomena without involving talk of compounded phenomenon as an ineliminable part of the explanation.
>This seems to be an additive process. I have suggested in a previous comment that we cannot get emergent phenomena and emergent causality through an additive process. We only get more of the same. We need an account of how something qualitatively different can emerge. Perhaps the Naiyayikas have an account, but if not, my recommendation here is to accept modern Western science, especially chemistry, the molecular level at which qualitative changes start to appear.
I am not sure if I agree, here. Specifically, I do not think that the Naiyayika needs to show that compound entities must have properties that are qualitatively different or otherwise inexplicable in terms of the behaviors of microscopic phenomena to be non-reductionists. They only need to show that the explanation of macroscopic processes themselves can only function by involving compound phenomena in their description–ie that the talk of compounded phenomena is ineliminable from the language of discourse.
Moreover, I am not convinced that the relationships between such things as the properties of chemical compounds and the behaviors of their atomic components are truly different in kind from the relationships between the structure of a lego-house and the arrangement of its bricks, on the one hand, and between the arrangement of molecules in a liquid and its macroscopic properties. In your examples of “liquidity”, for instance, the reductionist would argue that liquidity is merely the conjunction of such properties as incompressibility and continuous deformability under shear stress. And these properties are, in turn, readily explicable in terms of the behaviors of microscopic particles moving through a volume via bridging devices like Newton’s Law of fluidics and the conservation laws.
So, if you say, “When you arrange Lego-bricks house-wise, is there anything more than the bricks and the contact between them?”–the same could be said of liquidity. Is liquidity anything more than the capacity for the fluid particles to move in certain ways under the influence of certain forces—ie to move continually under constant shear stress but to resist such motion under pure normal stress?
And, briefly, to the point about the collection of people example: I agree that there must be a reason why some collections can be treated as genuine composite objects while others cannot. I am not convinced, however, that we can account for this in terms of macroscopic properties that are qualitatively different from microscopic ones.
> Alexander’s Dictum is optional for the causal exclusion argument. If I remember correctly, Kim included it in his earlier versions, but excluded it in his later writings. The dictum is required for eliminativism, but not for reductionism (mental events just are physical events, compounded entities just are uncompounded entities arranged this way or that), which can be argued for on the basis of simplicity considerations.
That is an interesting clarification as well, thank you. You may be right that the reductionist can get away with a mere parsimony condition, but my gut feel is that without something stronger like Alexander’s dictum the reductionist will have a hard time forcing their conclusions on the non-reductionist. For example, the non-reductionist of mental phenomena might argue that cognitions are directly observable–and therefore cannot simply be eliminated–and, moreover, have properties that cannot be explicated in physical terms–such as unity, privacy, intentionality, etc. So, even if they turn out to be non-causal, insofar as they still play an ineliminable explanatory role, they don’t seem to be readily reducible. In actuality, most Nyaya and Vaisheshika arguments that I am familiar with for the irreducibility of the atman to the physical body follow exactly this pattern–which is made even more poignant by the fact that, as I understand it, Nyaya does not actually consider the atman to have causal powers of its own (it merely claims ownership over the actions of the body).
(5) Pramod: “If any single individual net-wise arranged fiber were a sufficient cause of the fish containment, then it would compete with and exclude the other net-wise arranged fibers in the net! So, the fibers can act as sufficient causes only when taken collectively. But, in order for the collection of fibers to compete, causally, with the macroscopic net, the reductionist must ascribe genuine causal powers to the collection.”
Consider a plurality of people. Sometimes the plurality composes a whole, say a committee. Other times they don’t, and the effects they have are the aggregated effects of individual actions. Just having a collection of people together in one space is not enough for them to constitute a whole. I’d say the same is the case for atoms conjoined together in a net-shaped or fish-shaped or baseball-shaped region of space. Something more is required for them to compose a whole that is capable of doing things that the parts jointly or separately cannot do. And then, of course, we need to explain how the whole thus constituted can do novel things without violating physical or microphysical closure in a problematic way.
On second thought, I think an example that might be more intuitive than the poison might be:
Consider the the fiber that makes up a net. When the net is used to catch fish, the containing of the fish is an effect of the net. Is it an effect of the fiber too? I think the naiyayika might say, “no.”
The fiber can only hold the fish because it forms a net, if it did not form a net it could not hold the fish. Moreover, although each fiber, individually holds up a different part of the fish, it is able to hold up the fish itself only because those parts together make up the fish, if the fish parts were not assembled into a fish, the fiber would not be able to hold up the fish.
On both counts, then, the causal capacity of the fiber to hold fish is parasitic on the fiber on the property of then net to be a cause and the property of the fish to be an effect. As such, the fiber is anyathasiddha (via अन्यं प्रति पूर्वभावे ज्ञाते etc, from above) and not a true cause.
Perhaps this sort of reasoning is how Nyaya tries to argue for the irreduciblility of genuinely emergent phenomena? So that, as Uddyotakara might say, the fiber and the net are distinct because of their distinct causal capacities?
Pramod, thank you very much for your helpful answers, I am learning a lot from them, and agree with a lot of it. I have read through small parts of Tarkasangraha and Bhashapariccheda to get a better grip on the relevant basic Nyaya concepts, and will reply to your previous comment later. Here are some quick replies to your example.
Pramod: “Consider the the fiber that makes up a net. When the net is used to catch fish, the containing of the fish is an effect of the net. Is it an effect of the fiber too? I think the naiyayika might say, “no.””
Boram: The reductionist will say, “fish-containment” is the effect of the fibers arranged net-wise. And it’s a very good thing the fish-wise arrangement comprises prithivi-atoms, for if it were jala-atoms instead, qualified by dravatva, they would have slipped through the meshes like water.
The examples I’ve seen of asamavayi-karana in Tarkasangraha consider effects produced by an additive process, e.g., whiteness of the cloth produced by the whiteness of yarn. We cannot get emergent properties this way. I may be wrong, but I get the sense that the Nyaya explanation of the liquidity of water might be in terms of the liquidity of constituent molecules, just as the red color of a jar may be regarded the effect of “baking” at the atomic level, i.e., contact of tejas with earth-atoms that make them red.
But the liquidity of water looks like an emergent property which is not found in hydrogen and oxygen molecules. The emergent property (liquidity) of the whole is here produced by a chemical reaction involving parts lacking that property (cf. John Stuart Mill on the composition of causes).
Do the Naiyayikas discuss such cases? I would like to know!