From HIS meditations on consciousness, Ram recently asked: "WHERE is the self in all of this?" He cited hundred of concepts of the self (see below). This is actually a useful entry point to issues of importance in designing advanced computer networks and real mathematical understanding of conscious or intelligent systems.
I did not respond immediately to Ram;s question, because a certain level of mathematics would be threatening to people who depend on thinking in words.
But my wife showed me a video and book by a prominent Buddhist thinker, teaching at Princeton, Robert Wright. (
https://www.amazon.com/Evolution-God-Robert-Wright-ebook/dp/B002AKPEHW/).
He says: Buddhism says that there IS no self, that it is totally illusory. He backs this up with lots of work in neuroscience, pushing a "modular theory of the brain/ Prior to deep meditation, are we just a COLLECTION of very cacophonous little selves?
In fact, large parts of computational neuroscience take that viewpoint, consciously or unconsciously. Long before I read Wright;s book, I was well aware of Grossbergs seminal attenmpts to map out module after module in the adult brain, circuits to detect edhes, circuits to do X y and z. A vast collection of particular things. The question then becomes: how are these things INTEGRATED? As Walter Freeman asked: what is the underlying principle of MASS ACTION, which applies learning in a way which really integrates these pieces into an effective, unified whole?
Kozma, Davis and myself have written many dozens of papers on the mathematics and principles of how mundane brains actually do that, but this month I see a bit further into now that mathematics actually works, not just in tiny brains but in the general case. The details are important, but there are a few very basic principles which are very powerful if fully understood and implemented.
In just a few cryptic words, I would refer to CONDITIONAL CONGRUENCE of active modules, interfaced by stable but growing PARTIAL GATING. And I see how this connects both to noosphere mind dynamics, possibilities for more advanced internet systems (which we are not yet ready for, really), and for understanding a decently approximating the grand canonical Boltzmann density operator for our entire cosmos (which the mathematics of intelligent systems can be applied to).
Perhaps I will finish writing a partial draft which will really explain part of this. But for now...
somehow I feel called to mention a practical little example, on the outskirts of intelligent systems, which was crucial to me in really seeing how partial gating works. It seems like a simple piece of engineering, but again and again I have been amazed at how simple pieces of GOOD engineering can manifest very important general principles.
The example comes from the field of energy system modeling. ONE of the primary (mass action) roles of cerebral cortex, after all, is to predict or model the environment of the organism as it appears on the "movie screen of the brain," the thalamus. Prediction or modeling is a very fundamental component or aspect of any intelligent system.
From 1979 to 1989, I worked in a place charged with developing and using the main official models of the US and world energy economy in the US government, EIA/DOE. We spent lots of energy studying and advancing the art of how to build good predictive models of any system described by streams of time-series data. But we ALSO spent time on INTEGRATION, how to build a system which could COMBINE multiple models of multiple subsystems, to work together effectively. (
https://www.osti.gov/biblio/6288212/ Many of the techniques we studied are actually more powerful and general than any of the data analytics and policy models we ead of today. It is really sad to see models of covid and of coming eocnomic changes which fall far short of the quality we once had.)
Back in those days,the Electric Power Research Institute (EPRI) created a vision of a new, future type of modeling system, which would be far more powerful and agile than what we had in those days, In a way, it entailed developing a whole ECOLOGY of modules.
WHY spend hours of (then expensive) computer time running the most precise, detailed model possible of X, when, for a given application, it would be good enough to use reduced form mirror of that best model? PARTIAL GATING, in essence, is the relation between the big X and the reduced form, to be connected of course with modules which integrate through multiple gates in a larger system of integrated modules.
It is ever so simple, in a way.
The mathematics of deciding whether two modules are CONGRUENT with each other, where one module can effectively approximate another, is very fundamental. For many years, people working in practical neural network engineering (back in the days when computer scientists mostly felt there is no such thing as a general learning neural network) paid special attention to the approximation theorems of Andrew Barron of Yale, showing how ANT smooth function can be approximated at bounded cost by a certain simple type of neural network, That type was not powerful enough to approximate the kinds of mappings or relations we encounter in making decisions in life, like how to navigate a cluttered space, but with Kozma and Ilin I later showed how to train a more powerful type of neural network. At
www.werbos.com/Erdos.pdf,
I eventually provided a whole LADDER of worlds or primary functions we kight try to approximate, and more complex types of neural network capable of approximating them.
And so... in an ideal world, I feel we may be ready to take that further, There are more general approximation theorems possible, which could unify this whole area, and also incorporate simultaneity ACROSS time, as in the kind of equilibrium between past and future which the Everett/Wheeler/Deutsch formulation of modern quantum field theory (QFT) assumes.
Some folks on this list have been raised to believe in many fundamentalist religions, such as the belief in cartesian dualism and "god in a [pair of sunglasses (yje Copenhagne model of measurement from the 1920s which remians very influential in tribes which have yet to learn modern mathematics let alone mathematical physics or engineering). In truth, I do not really believe that the modern "Schrodinger equation ( psi dot = i H psi) is the total dynamic law of entire cosmos or multiverse, but it is the best basis we have right now to make better real sense of the emergent phenomena we have around us, on all scales of human experience. To ignore what it can tell us is just neurotic avoidance behavior. (Yes, we see a lot of that kind of behavior lately in very high places, but that is no justification for imitating it.)
The more general approximation theorems with partial gating would not only fit the kind of "matching which we see in the mundane brain (beautifully described in Pribrams bopk Brain and Perceptio) but also the capabilities of fifth generation analog quantum computing learning systems (previewed in the paper by Werbos and Dolmatova in Quantum Information Processing). Ordinary Darwinian evolution is basically a slow, feedforward process to reach an equilibrium thermodynamic distribution, but quantum evolution is the time-symmetric equivalent, much faster, and one way to try to understand not only the evolution of noospheres but of the patterns in the larger canonical Boltzmann distribution which is full of its own type of "matching" or "reflection."
It is not just a poetic accidnet that the words "mirror" or "reflection occure over and over again in the deepest mystical literature as well. I cannot help recaling "the mirror or Ameraterasu" for which I posted a photo link at
werbos.com/religions.htm.
Of course, the degree of congruence is always approximate in brains (and even in themrodynmaics at a cosmos level). Great work on neural networks and on consciousness osmetimes seems like a hall of mirrors. The less evolved system of policy intelligence we see even in the best of governments and human organizations is sometimes more like a hall of fun house mirrors, where the imperfections can be devastating. A lot of the "fake news paranoia left and right has that kind of flavor, and is truly dangerous to the survival of all of us. Perhpas we need more of a a different kind of mass action, more like Freeman's than like the left or the right or other such dogmas negligent of the need for truth, science ande yes, mathematics. I really applaud the person Robert and I know, who put a huge statue of Von Neumann in the center of her living room (and did friend me on Facebook).
Best of luck. We do need it.
Hi all,
From various disciplines/investigations, so far, there are 48 facets/sub-aspects/notions of the self, which are listed below.
As per (Strawson, 2000, p. 39) “[1] the conceptual self, [2] the contextualized self, [3] the core self [(Damasio, 2010)], [4] the dialogic self, [5] the ecological self, [6] the embodied self, [7] the emergent self, [8] the empirical self, [9] the existential self, [10] the extended self [(Kim and Johnson, 2014)], [11] the fictional self [no-self (Metzinger, 2003) because the self is a construct of the brain that can be easily modified], [12] the full-grown self, [13] the interpersonal self, [14] the material self [objects and cultural artifacts: (Woźniak, 2018)], [15] the narrative self, [16] the philosophical self, [17] the physical self, [18] the private self [and [19] public self], [20] the representational self, [21] the rock bottom essential self, [22] the semiotic self, [23] the social self [human beings: (Woźniak, 2018), (Moutoussis et al., 2014; Friston and Frith, 2015)], [24] the transparent self, and [25] the verbal self (cf. e.g., James, 1890; Stern, 1985; Dennett, 1991; Gibson, 1993; Neisser, 1994; Cole, 1997; Butterworth, 1998; Gazzaniga, 1998; Legerstee, 1998; Gallagher and Marcel, 1999; Pickering, 1999; Sheets-Johnstone, 1999) [Strawson (1999b)]”.
In addition, [26] proto self and [27] autobiographical self: (Damasio, 2010) or continuous self (Newen, 2018)], [28] “the self that enjoys self-consciousness, which includes the sense of body ownership (‘this is my arm’), the sense of agency (‘this is my action’), the sense of authorship of thoughts (‘this is my thought’)” (Newen, 2018), sense of self in agency and perception (Hohwy, 2007) [29] “working self [=contextually relevant part of the self-model] must still involve explicit as well as implicit information if we want to describe the embodied self adequately concerning all its behavior, its behavioral dispositions, and its explicit autobiography (as expressed in narratives in a situation)”.
[30] “the phenomenal self (reflecting self-related content of consciousness)” (Woźniak, 2018) [the self-as-object (me-ness) or objective self?: “My arm is broken”, “I have a bump on my forehead”: (Wittgenstein, 1958)]
[31] “by limiting ourselves to discussing conscious content representing one’s body one can speak about the bodily self” (Woźniak, 2018), self as sense of bodily ownership (Apps and Tsakiris, 2014) [32] “by imposing limits to conscious experience of one’s possessions one can speak about one’s extended self” (Woźniak, 2018), , [33] perceived self, and [34] free-energy self.
As per (Gallagher, 2013) on <A pattern theory of self>, “we may want to add [35] “the neural self,” [36] “the synaptic self” (LeDoux, 2002); or what we might call [37] “the midline self” [in reference to self-referential processes in the cortical midline structures (CMS) (Northoff and Bermpohl, 2004); “the midline theory of self (or for short, the midline self)” (Gallagher, 2013)]. […] Many of these concepts of self were developed in the plural. James (1890), for example, distinguished between the physical self, the social self, and the private self. Neisser (1988) discussed five types of self-knowledge corresponding to [6] the ecological self, [38] the interpersonal self, [2] the conceptual self, the extended self, and the private self. Despite the terminology suggesting a plurality of selves, however, Neisser (1991) carefully refers to them as aspects of self – e.g., the ecological aspect of self.” [39] Experiential aspects of self (Gallagher, 2013), [40] affective aspects of self (Gallagher, 2013), [41] situated aspects aspects of self (Gallagher, 2013). [42] spiritual self (Poll and Smith, 2003), “self-as-object” as mental processes and content (Woźniak, 2018)).
[43] cognitive self (Strawson, 2000, p. 39), [44] the self-aware, independent, eternal/immortal, passive invariant self (PIS) is from the non-interactive dualism-based Sankhya, [45] transcendental self (Chris Nunn), and [46] “Pure Consciousness” or “Pure Self-Observing System” as in Advaita Vedanta (Maharishi Mahesh Yogi, Alex Hankey, and (Boyer, 2018), [47] In the IDAM (inseparable dual-aspect monism), the “self” is the experiencer/cognizer/actor as a non-physical (np) aspect of a self-as-subject-related state of a subject’s midbrain system with respective neural-physical basis (NPB) as the inseparable physical (p) aspect. Let us call it <the IDAM-based individual dual-aspect-self>. [48] Metaphysical self ((Searle, 2005), (Woźniak, 2018)).
We can categorize all notions of the self in two groups:
(I) James’ “Me” (self-as-object in the sense of Phenomenology (experience of self, phenomenal self): Wittgenstein’s “I” (‘I see me in the mirror’) and/or Wittgenstein’s “Me” (‘I see me in the mirror’); this includes the notions/meanings/definitions/sub-aspects/facets of self as in [1]-[42]. In other words, “Me” (self-as-object) is the SE of self, i.e., self-experience/consciousness/awareness, which can include Damasio’s core self (np aspect) and the self (np aspect) that has neural-physical basis (NBP) = Northoff et al’s CSMS-NN and its activities = the p-aspect of a self-related state of a mindbrain system in the IDAM framework.
(II) James’ “I” (self-as-subject, metaphysical self: “The self as a metaphysical fact that consciousness is subjective: ‘the Thinker that does the thinking’) in the sense of metaphysics (existence of self); this includes the notions/meanings/definitions/sub-aspects/facets of the self: [43]-[48]. In other words, “I” (self-as-subject) is experiencer/cognizer/thinker/actor; it is NOT the SE of self. In the IDAM, it is the experiencer/cognizer/actor(performer of actions) as a non-physical (np) aspect of a self-as-subject-related state of a subject’s mindbrain system with respective neural-physical basis (NPB) as the inseparable physical (p) aspect.