From: primeradiant (firstname.lastname@example.org)
Date: Wed Oct 27 2004 - 09:01:00 MDT
After some time studying the SI literature (and that of other projects/groups such as Novamente, Bitphase, Hal2004 etc.) as well as perusing trains of thought in SL4 I've finally decided to 'butt in' with some thoughts of my own - just 'thoughts' arising from my own research (I'm a psychologist). Sorry if it's a long posting, but I have the SI project very much at heart. I'd like to invite discussion, not necessarily on all points, but on those that seem most salient to any of you.
"No species truly comes of age until mind is comprehended".
Amoroso, Richard L. & Barry E, Martin. 'Modelling the Heisenberg Matrix: Quantum Coherence and Thought at the Holoscape Manifold and Deeper Complementarity'. Scale in Conscious Experience: Is the Brain Too Important To Be Left to Specialists to Study? Proceedings of the Third Appalachian Conference on Behavioural Neurodynamics, 1995 pp.351-77. Lawrence Erlbaum Associates Inc., Mahwah, NJ.
"Il y a. un problème de philosophie des sciences encore largement impensé. Tout se passe en effet comme si, loin d'être de simples gestes d'idéalisation par passage de l'approximatif au rigoureux, les actes fondateurs des grandes ruptures scientifiques étaient des actes forclusifs scotomisant une strate réelle de l'être au profit d'une élaboration symbolique.".
Jean Petitot (1978). 'Sur le Modèle Historique de Thom-Pomian' - Mathématiques et Sciences Humaines, 16e année no 64, 1978 pp. 43-70.
1.. The biases and constraints on human cognition and intelligence are fourfold: evolutionary, historical, generic and contextual.
b.. Evolutionary constraints arise from hysteresis or disjunction between rates of evolutionary change. We have by no means emerged from our evolutionary substrate as yet, as gene-culture co-evolutionary theory clearly demonstrates (Boyd & Richerson, Cavalli-Sforza, Feldman, Lumsden & Wilson etc.). This is the real meaning behind Freud's Civilization and its Discontents.
c.. Historical constraints derive from the dominant defensive constructs prevailing within a given culture at a particular time. Defensive constructs are collusional methods of intrapsychic defence erected against generic and inflicted trauma and are stimulated by anxieties arising through confrontation with the presenting environment. Our current dominant construct arises in reaction to the prevailing anxieties of our time - anxieties related to population growth, imminent resource depletion, environmental degradation and unstable multigroup interaction. Similar constructs were common to all past civilizations in their late imperial phase - as ours is now.
d.. Generic constraints arise from generic trauma. Humans are a highly neotenous species (Gould 1977) with the result that almost a third of our lives are spent in a state of dependency. Generic (or perinatal) trauma has its origins in human neotenous ontology and specifically in the trajectories of morphogenesis (growth) and the catastrophe of birth. Emotions derive from the transmarginal stresses experienced and encoded in the early (pre-natal) and perinatal trajectories of morphogenesis. Generic trauma is ahistorical - i.e it is constantly renewed at every moment, with the result that the symbolisms of uterine ecology form the Rosetta Stone of all world religion and civilisation itself becomes a fractal expansion of the fetal unconscious. This constant mass re-traumatisation precludes any attempt at universal annealment, forcing us to seek technological solutions to the crisis now facing us.
e.. Contextual constraints arise within a given group at a given moment in time in response to the stresses of the presenting environment at that moment. Psychodynamic models clearly show the nature of these stresses and the mostly dysfunctional response of particular groups to them. All futurist-oriented groups on the Web clearly show a similar range of dysfunctional symptoms. These symptoms obscure our perception of and engagement with, authentic reality. If our philosophies, mathematical systems and techniques of envisionment all suffer from this obscuration (praise Gödel), how can we possibly program a machine intelligence to transcend them?
6.. This combination of constraints arose though the development of traits that once conferred adaptive strength during our extended period on the EEA but are now proving maladaptive in the creation and management of complex technological societies and indeed now threaten to become the evolutionary nemesis of the species. These traits include neoteny induced dependence, dominance submission relations, aggression and violence, culturally-reinforced gender roles, permanent sexual arousal, dysfunctional reproductive & childrearing modes, species solipsism and species narcissism. The last two traits help us to scotomise the four constraint domains and reinforce the group fantasies of all futurist groups, the SI included.
7.. Individual and group fantasies arise from individual and group defensive constructs. Defensive constructs have nothing to do with the fact that we are 'imperfectly deceptive social organisms' (pace Nesse & Lloyd in The Adapted Mind). These constructs and fantasies form the structure of the dynamic unconscious (DU).
8.. The DU is real, and it's a great pity cognitive science continues to scotomise (ignore) this fact. The DU is entirely accessible to analysis and should be an integral component of cognitive science. The fact that it is not greatly constrains scientific advance and seriously delimits the type of future societies we can realistically envision and create.
9.. Clues to the neurogeometric origin, structure and function of the DU have been investigated by Jean Petitot , Varela and numerous others, including myself. IMHO the most realistic neural models should incorporate Edelman's neural group selection theory.
10.. The interaction of defensive constructs within various layers of the DU creates subtle twists and biases within Bayesian networks, not to mention the structuring and constraining of Bayesian priors.
11.. The paradox of generic trauma lies in the fact that while it creates constraints on human reality perception, it is also the source of motivation and creativity. The 'burning question' IMHO is how the affective sources of directed motivation can be built into Seed AI.
12.. Current CogSci should evolve more in the direction of 'embodied enactivism' (Petitot et al. Naturalising Phenomenology: Issues in Contemporary Phenomenology and Cognitive Science, Stanford University Press 1999 - esp. Ch. 1). Failure to do so will result in some embarrassing (not to say disastrous) consequences in AI and other fields (see the Petitot quote at the head of this letter). Tversky & Kahnemann simply won't do any more. I had some serious disagreements with the cryonics people over the role of the ANS in both cognition and the 'nature of the Self'. They haven't read Schore's Affect Regulation and the Origin of the Self (1994)- nor, unfortunately, do they seem to want to. Pity. All those frozen severed heads stacked up in Dewey flasks out there in the Arizona desert. how gothic!
13.. All literature, including SF, is still thoroughly grounded in EEA-derived principles (see Stanislaw Lem's Microworlds. Harcourt Brace & Co., New York 1984 - also numerous other studies by Dunbar et numerous al.). I agree with Lem that our psychology is currently unable transcend these archaic evolutionary barriers. Transhumanists and Singularitarians should take this into careful consideration when formulating their plans - after all, 'the future is a canvas on which we paint our desires' (Bene Gesserit aphorism from the Dune Encyclopedia).
n.. The Singularity faces the threat of becoming the latest is a series of perinatal fantasies of transcendence and rebirth. The converse fantasy of the 'AI as Demon' (Hollywood) is the AI as 'all protective, nurturant Mother' (e.g. Kingsley Amis' The Machine Stops - plus many more SF instances). I don't know which fantasy is worse.
o.. Nevertheless, without FAI, future projections look bleak indeed (e.g. Heibroner 1991, 1995; Wallerstein 1998).
p.. There will never be any spontaneous 'great awakening of consciousness' - no matter how much we meditate and pray, we're never going to somehow wake up some Monday morning 'all loving each other'. We need to roll up our sleeves and get to work. 'They' will never invent anything. We are 'they'.
q.. None of this necessarily impacts directly upon the future design and creation of Seed AI, but it certainly impacts on the team working towards its design and creation.
r.. Seed AI, leading to Friendly AI, if such a thing is possible, can only be achieved through a transgenerational program of hard work. Although the CMOS-nanotech interface is indeed likely to occur within 10-15 years, it will not guarantee future advance unless the human teams engaged in the construction of Seed AI is committed, motivated, and psychodynamically annealed as a maturational or advanced 'Balint-type' group. This follows from Eliezer's statement on the 'idealized altruistic human' in CFAI's Beyond the Adversarial Attitude.
s.. Without this, nothing will happen. The SIAI will run the risk of becoming what other 'future-oriented' websites are - money-spinning 'electronic tribes' who do little or no useful work whatsoever - except orbit around defensive symbolic-semantic webs woven to contain constraint-induced anxieties.
t.. The SI must therefore become polycentric and metacomplex in both though and action (see Streufert & Satish's collected papers 'Complexity Theory: Predictions Based on the Confluence of Science-Wide and Behavioural Theories'. The Journal of Applied Social Psychology 27, 23 (1997) pp. 2096-2116. This cannot happen spontaneously but requires special training.
u.. How can we presume to talk about 'collective volition' when we know so little about the psychohistorical evolution of non-Euroamerican cultures?
v.. Whatever happens, even drastic future changes will be evolutionary, i.e. the human race as 'the human family' will not be dragged some future in toto. Only some of us will change - and change we will - irrevocably Whether those of us who do so will weep or exult in our transformation will be our own affair. This is the nature of evolutionary speciation - if the constraints of a given evolutionary trajectory are transcended through speciation, other constraints will arise along the newly-chosen path (if we are thinking in terms of 'self-directed evolution'). The MOSH will remain unchanged - and this will present serious political and legal challenges.
w.. There can be no 'perfect' being in an evolving Universe. Neither can there be 'immortality' as we currently understand it. Whatever has a beginning has an end - or will change out of all recognition (the old story of the hammer - the head and the haft). IMHO this lies behind Vernon Vinge's criticism of the SETI program.
x.. A species is an energy field through which the computational power of prespace is compactified and translated 'upward' into existential reality (sorry - like all of us, I'm a prisoner of a gravity well). The 'secret' of autonomous, self-replicating and self-enhancing life lies at the boundary between prespace and neural ultrastructure. This has enormous implications in terms of 'crystalline' programming and the supportive hardware. Quantum scarring fields that have been traced and imaged so far show an uncanny and spine-chilling resemblance to primitive organic forms. Intelligence is therefore probably a functional property of matter (or, as A&M state: information of state is endemic to the very fabric of the Universe). Human intelligence is only a primitive instantiation of intelligence (as Eliezer hints at in LOGI). Nevertheless, we should not waste our time chasing GUTS. The nearest thing we have to a GUT at the moment is probably Laurence J. Gould's attempt at synthesising the formalisms of Bohm and Pribram (in Scale & Conscious Experience above) - but this does not (IMHO) clearly differentiate between individual and group consciousness.
Some of these points are elaborated further in a brief paper of mine - 'Trauma and Transcendence: Psychosocial Impacts of Cybernetics and Nanotechnology' in Proceedings of the PISTA 2003 International Conference on Informatics and Systemics (AAAS) and also available on the World Systems Analysis Archive -URL:
I would be willing to discuss any of the above points with anyone who wishes to do so - and provide fuller material (references etc.).
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:49 MDT