The Indian tradition is replete with modeling of human experience. In the previous post I discussed how Sāṃkhya’s model of mind, body, and consciousness might be used to rethink the conceptualization of evolutionary theory and neuroscience. I argued that the ontology of Sāṃkhya would suggest that the interactions between mind and body are entirely within the scope of “nature” and that “consciousness” is outside this natural system, yet it gives the appearance of consciousness within nature. How might this ontology look at AI?
There are two forms of AI that exist right now of which I am familiar; this discussion excludes the types of AI one might envision we shall develop in the future. The first is neuralinks that would connect the human brain to external computers, phones, machines, bodily limbs, etc.; and the second is independent robots that can execute complex actions and use machine learning. Will we ever see the types of robots in science fiction movies? I do not have a strong view on that, but I do believe that various forms of AI will be used for travel, sales, agriculture, and human services will be used more and more in the near future.
From the perspective of Sāṃkhya, all forms of AI would be considered “natural” in the sense that they part of nature or prakṛti; even human artifacts are still natural in the broad and general sense that Sāṃkhya might be thought about in English. The word “artificial” as it is used in AI often seems to be used in the sense of a human creation that is distinct from the biological structures created by “natural” processes like natural selection.
But I have two specific thoughts on AI. The first is that Sāṃkhya has a robust concept of “intelligence” and in their view the intelligence functions as an entirely natural and yet unconscious manner. The intellect (e.g. buddhi) requires the consciousness to function, but consciousness exists independently from the intellect (the innate features of the consciousness is a matter of discussion among different scholars). Generally speaking, intellect in Sāṃkhya does not imply an independent or self-directed will; its being is activated if and when it is connected to a particular consciousness, but the consciousness is the being that provides intellect with the ability to function in the various ways that it can function. The intellect could not develop a self-reflexive state of awareness, since that portion of cognition is provided by the consciousness which uses or observes the functions of intellect when it is in connection with a natural mind-body complex.
The notion that intelligence functions when in connection with a consciousness brings me to the second thought I have on Sāṃkhya. The ego is what connects the nonphysical and eternal being of the consciousness to the natural and ever changing mind-body complex. The birth of a human or animal in Sāṃkhya is essentially the attachment of a particular consciousness to a particular mind-body complex through the power of the ego. Thus, for Sāṃkhya, the creation of a robot that has AI would require the ego to attach to said robot. The question, then, for Sāṃkhya is: is it possible for humans to create a machine to which a consciousness can attach via the ego?
Readers of this blog, however, must also be familiar with the Buddhist theory of pratītyasaṃutpāda, or the co-dependent origination; the Nyāya theory of the padārtha, or that reality is composed of distinct categories that correspond to specific words; the Vaiśeṣika atomic theory; and there are other theories in Jainism and a wide range of distinctions within those mentioned. Would these conceptual schemes look at AI differently? Would they raise different problems? My goal here was not to go into these theories in any depth, but ask a more general question about the term “artificial” and how that would be thought of in a conceptual scheme that see all forms of intellect as natural; and to think about the problem of the ego as it pertains to the creation of AI machines or robots.
Jonathan,
your question now verges on the ‘ hard,’ range. where answers elude us. You can say it’s directly as hard as comprehending the Tattvacintamani of Gangesa, with its four books ranging over sensation, perception and inference. And the scholars of our time have tackled only sections of that problem. Jonardon Ganeri speaks of Gangesa developing a technical language for Nyaya, and that one can compare to a programming language, on the specific model of Steve Wolfram’s Mathematica. Right now at Wolfram they are asking in public, how does their language come to represent causality as understood in physics, and that is the core challenge in programming a robot!
Back in Indian philosophy, Gangesa touches on vision and the line of sight, which may be broken, as in an eclipse. The same theme is briefly stated in the Yoga-sutra, but there one might think the text scrambled. In the background, eclipses were studied for millennia in Babylonia, and from there the Paulisa Siddhanta joined the Sanskrit tradition. So the problem of causality can be mapped from Gangesa to the Yoga-sutra, and that interests me, along with the debate at Wolfram.
These are very complicated questions, which engage whole teams of people over many years. And here pose a large challenge to comparative philosophy. From the Yoga-sutra one might track back to Samkhya through the Samkhya-Patanjali cosmology, but that route is scarcely known today, except as Jyotish astrology, or Maya weaving illusions, which sounds like the hard problem of the philosophy of consciousness.
Thank you for an illuminating reply.
I was not aware that Ganeri and compared Gangesha and Wolfram; I respect the work of each of them and shall look for that immediately. If you would be so kind as to shove me in the right direction, that would be appreciated. You may be aware Wolfram recently did an interview with Lex Fridman on AI: https://www.youtube.com/watch?v=ez773teNFYA&t=1880s
I am curious how you connect Gangesha and Yogasutra on the matter of causality if you care to say more.
I understand : Sankhya philosophy views consciousness and nature are two eternally and independently existing entities.Intellect is a product of nature and in some mental states it is in proximity with consciousness.When they are in proximity, a reflex relationship happens between them.It means consciousness appears to be experiencing what ever is happening in the intellect, like knowing.enjoying and being an agent and intellect appears to be having consciousness.Ego is the result of this relationship.
As per this understanding,consciousness has no role in inner workings of intellect and hence natural intelligence,like artificial intelligence does not require
consciousness.
I agree. That says much more clearly what I am thinking about this.
Let me ask you this:
In classical Samkhya (e.g. the Karika, the Sutras, the Samkhya-yoga commentators) can buddhi (whether natural or artificial) function as an agent within or a distinct feature within the purusha’s experience of prakrti if the buddhi is not connected by the ahamkara to a particular purusha?
By being in reflex relationship ( which is described as happening between said entities in proximity ) the activities within intellect are experienced in consciousness and intellect appears to be having consciousness.The intellect is doing the work of an agent
and agency is felt in consciousness.
This reflex relationship has a fascinating philosophical history. Socrates in the Symposium of Plato speaks of what he learned from one Diotima, a priestess of Mantinaea, a city allied to Sparta. She described philosophy as something between matter and God, hence of a spiritual nature: and compared it to the mediating ratio (extreme-and-mean-ratio), which appears again in the Divided Line image in Plato’s Republic.
Soraj Hongladrom at Chulalongkorn University in Thailand found it again in Kant’s Practical Reason, linking freedom and morality! Moral action leaves others free to be, so it’s reciprocal: and Claude Levi-Strauss traced that principle through language and trade to kinship in anthropology.
Science follows the same elusive logic: when we test a concept of natural law, it must appear again in the experiment as a logical form of the process, as now traced in quantum theoretical analysis. And that form-in-process one can now take to programming.
I’m very pleased to find that this thread of truth ran clear past the whole sorry, muddled and homesick colonial history, thru independent Thailand! And let’s us now pick up the threads of freedom…
That’s richly painted story of philosophy. Thank you. You wrote: And that form-in-process one can now take to programming. Does that mean for you that the subject of a scientific (e.g. a specific bit of AI tech, like machine learning computer program) might one take make itself as an object of study. I’m thinking here of Kant’s argument that for the subject to know itself it must make itself an object of experience, and in doing so it presses it into the forms reason like time and space. Do you think AI can make itself an object in its own processes? Can AI study itself the way we study ourselves?
Hi Jonathan and Orwin,
Fascinating discussion. I am very interested in this topic.
I had a quick question that seems to me to be relevant to the topic of AI and Conversation in a variety of Indian traditions.
It is common to distinguish between phenomenal consciousness and intentional consciousness, and also between phenomenal and access consciousness in contemporary philosophy of mind.
In most discussions of AI I have found that this distinction is often overlooked and not considered when bringing Indian philosophical traditions into contact with AI.
I am curious about Samkhya and whether it draws the distinction between intentional, directed consciousness and phenomenal, what is like, consciousness.
Some current theories might hold that the two always come together, but I think cases of blindsight suggest otherwise.
When we talk of consciousness within Samkhya are we mostly talking about phenomenal consciousness?
Thanks
Hi Johnathan,
My first answer got lost -probably timed out, – so this is a quick answer. All the basics of computer systems are self-directed, from the power-on-self-test (POST) thru the command-line interpreter, to compilers which generate aps. And the sense of causal necessity acting thru space-time comes up in relational databases, which constitute their own transactional histories. AI is very largely *dependent on looking in to “virtual reality,” and in that way very different to natural intelligence. All the a priori type reasoning which zKant rejected from psychology and theology comes alive in artificial rules- based systems.
Hi Anand!
Your distinction goes back to Plato responding to Lycophron. in an erotic wrangle about being and predication. Lycophron argued for a pure process language, with no use of the copula in predication. Plato answered that we use the copula in two senses, one descriptive or phenomenal. for states of being; and another for what we call intentional actions.
For Samkhya I suspend judgement for now, having noted the pradhana: it’s a triadic semantics, which may not translate easily. But John Dewey in his book on logic, in Part 2, where his reform of logic gets under way, captures something of the sense of purpose or teleology that emerges in Samkhya, but more especially the theistic Samkhya-karika.
Hi Orwin
I am familiar with Dewey’s work. I will look into it to see what I can get. I am generally trying to find answers in Indian philosophy of mind to the distinction between phenomenal states and intentional states and how that is handled in various traditions and where it can be used to make an in road into issues in AI and Consciousness.
Thanks for your help!
Returning to your original question. Jonathan , about that nature of thought or intelligence, the mechanics of computing take off from the clock. So the thread to watch at Wolfram is on the Charge Parity Time theorem in physics, which reaches to fundamentals. And computer science took off from Alfred Korzybski on time bindings.
Following your article on self-organization, which I found very refreshing, morphology today finds cells moving in waves to lay out the body plan, as if by time binding! So it’s the process of growth that should run like clockwork. And our mental capacity grows out of that. . .through the senses and their objects.
But there I find an almighty academic wrangle over Caraka, Ayurveda and the ancient Samkhya, reaching back to Kapila. Coming from morphology and thus Panini, Im interested in all that, but the heavy hitters simply want to stamp it out, as rhetoric tied to the very successful Ayurvedic industries.
Briefly, I have a hard result in psychophysics, a true Shannon measure that covers all five sense objects in terms of free energy densities. It’s a phonon heat capacity model, and fits Caraka’s account very well. And gives a generic measure of sensory information. But to publish it as such looks harder now. As for new science ground under dogma, history is littered with it. . .a large thread of pain answered cryotically by Buddhism.
Wolfram and Samkhya go together nicely. I just talked to him about it in March!
Can you say more on that?