Colin, yes you answered my questions about consciousness. To summarize, by consciousness you mean qualia, that which makes you different than a philosophical zombie. Since a zombie is by definition behaviorally identical to a human, there is no test for consciousness and no capability that depends on it. You could simply declare that neurons are conscious and transistors are not, even if they implement identical functions, and nobody could argue otherwise.
Do I have this correct? Or do you mean something else? Anyway, I understand your confusion. It happens to a lot of smart people like Penrose, Chalmers, Pierce, and others who talk about AI but are not building it. When brains think about themselves, they bump into Wolpert's lemma, which says that no computer can model itself (a consequence of his theorem that two computers cannot model each other). If you believe your brain has a soul or a homunculus or little person inside, then that is the error that Wolpert proved must exist. Those actually building AI (Google, Facebook, Amazon, etc) don't care about consciousness because (by definition) it has no effect on behavior. If your machine passes the Turing test, it is conscious as far as anyone can tell. But maybe I can ease your confusion before you start writing your paper. Qualia is what perception feels like. Free will is what decision making feels like. Feelings are mental states that produce reinforcement signals. A positive signal modifies your behavior to repeat actions that preceded it. Negative (pain) does the opposite. Perception, thinking, and action feel good most of the time so that you will act in a way that ensures they will continue by not dying. Fear of death leads to more offspring to propagate your DNA. You could reproduce all of this in a computer, or let it evolve. Either way, you will end up with programs that claim to be conscious, but unable to prove it, just like you. On Mon, Jul 1, 2019, 6:48 PM Colin Hales <[email protected]> wrote: > > Next installment. > > Matt? This is where the questions you asked about consciousness get my > answer, except for my specific proposal for the brain biophysics that > originates consciousness. That discussion is implicit to the chip design. > > It's the last thing to do before the silicon replacement argument gets its > 21st century upgrade and then I can deliver the chip design for an > artificial brain/mind that does not use computers (in the mainstream sense > of the word). > > This needs editing, but it's basically OK. Please refer to the earlier > science framework figure including (a)...(e) and the more recent figure > that has only (e). > > cheers > colin > > CONSCIOUSNESS > > An explicit, formal neuroscience of consciousness, with its own literature > and conferences, a dedicated community and an explicit presence in teaching > curricula, is roughly 30 years old. The word consciousness has been used, > so far, without technical specificity. The community that carries out the > science seeks a scientific account of the origins of the first-person, > experiential life we all have. When we use the phrase ‘mental life’ or > ‘mind’ we are referring to the collection of unified 1st-person > experiences that we each have. That is what the word consciousness means to > the community that does the science. Let’s call this the 1PP for ‘1st-person > perspective’. We describe the different modalities of experience with names > like: ‘vision’, ‘audition’, ‘olfaction’, ‘gustation’, ‘touch’, situational > emotions like fear, disgust, anger, happiness, ‘primordial’ emotions <ref > denton> associated with basic homeostasis like hunger, thirst, orgasm and > pain. Note that this use of the word consciousness does not refer to > physiological states described with words like ‘awareness’, ‘alertness’, > ‘awake’, ‘asleep’, ‘vegetative’ or ‘coma’. These words refer to a brain’s > ‘operational state’ that determines the degree and scope of consciousness > (1PP), but does not originate it. If a person is asleep and dreaming, there > is a 1PP and by the definition of the word as used in the science, the > dreamer is conscious to that extent. > > Practitioners in the science of consciousness seek a scientific account of > (i) what originates the very existence of a 1PP and (ii) what properties of > its origins give rise to the distinct characters of different modalities > such as the ‘redness of red’ or the ‘pain-ness of pain’. Practitioners of > the science refer to the science activity as an account of “*what it is > like to be*” an agent with a brain. They do this because one of the few > certainties we have in the science is that it originates in the *cranial* > central nervous system. No other part of the nervous system generates the > 1PP. > > > > As an explanandum in science, consciousness is unique. Its uniqueness > arises under (e) in the science framework. Having a nascent science of > consciousness means that (e) in the framework is the place where the 1PP > entered science for the first time as an explanandum, in the second half of > the 20th century. Comparing and contrasting “*What it is like to be > (a)…(d) LEFT, (a)…(d) MIDDLE and (a)…(d) RIGHT*” does not currently > appear in any of the sciences (a)…(d). But in neuroscience, comparing and > contrasting (e) LEFT/MIDDLE/RIGHT is a valid science activity. The addition > of the 1PP to science represents a major structural shift in the conduct of > science itself. Therefore, we must be extremely attentive to how the > science is conducted. This is the motivation for the detailed attention to > the conduct of science presented above. We can now see how consideration of > artificial consciousness and its role in artificial intelligence, is an > intrinsic and mandated part of a complete neuroscience of the natural brain > as the originator of all behaviour. > > > > Under (e), a scientific account of consciousness is literally an account > of what originates ‘observation’, including ‘scientific observation’ (in > context). Where ‘measurement’ refers to specific contents of consciousness > (in a science context), ‘observation’ occurs when the scientist experiences > (1PP) the measurement. That is, in accounting for the 1PP of a natural > cognitive agent, we are also explaining an ability for scientists to > observe nature at all. The fundamental problem with the science is that we > are using observation to ‘explain’ an ability to observe. This is a > troublesome logical state for the science to inhabit. Scientists under (e) > are literally self-describing/explaining how it is we can do science, and > furthermore, we are creating, in principle, an artificial scientist. This > level of troubled self-referentiality is deeply intrinsic to the science. > While this fact gets reported sometimes <Edelman>, it is perhaps the most > under-appreciated part of the science of consciousness. Under the framework > (e), this situation can now be properly recognised. It is the only place > where science is required to account for how science can happen at all. > > > > The final important fact of consciousness is the ‘cranial central nervous > system’ specificity of the origins of a 1PP. This means that ‘being’ (some > subset of) the physics inside the cranium, a*nd only that physics*, is > the singular sure fact we have acquired about the origins of consciousness. > Identifying the specific subset of the cellular biophysics responsible for > a 1PP is the goal of the science. In the previous discussion of (a)..(d), > it was observed that each had a mandated ‘essential physics’ upon which > function was critically dependent. In (e) we are able to examine if there > is ‘essential physics’ in the brain. We observed that the brain must be > very special if it is the only place in science lacking essential physics. > Identifying that physics (or its lack) is a top priority. In consideration > of consciousness, however, we have a double obligation to do the science > correctly. It may be that the essential physics of the brain is the same > physics that originates consciousness. The full integration of all of these > issues under (e) needs very careful management if the science is to be > conclusive. > > > > Shortly we are revisiting the ‘silicon replacement’ thought experiment to > better tease out how it involves itself in the science of identifying this > potentially ‘essential physics’. When the physics of the brain is replaced > with something else, it possible that it is how it impacts consciousness > and its visible consequential effects on behaviour, that is central to the > science. It is certainly a viable way of approaching the ‘essential > physics’ issue. We must be able to compare/contrast the same brain function > before and after the physics is replaced. We can do that under (e). > Ultimately, “*What is the function of consciousness in intelligent > behaviour?*” is the big question. To do that science properly we have to > hypothesise which biophysics is thought responsible for a 1PP (which is > literally to propose a ‘theory of consciousness’) and then compare/contrast > (e)LEFT (artificial versions of it) with a lack of it (e) RIGHT and how the > resultant difference relates to the brain (e)MIDDLE and the observable > behavioural differences (say, in a capacity to learn). Regardless of any > claimed connection between ‘essential brain physics’ and consciousness, we > cannot do any of the science conclusively when (e)LEFT is missing. > > > > *Artificial General Intelligence List <https://agi.topicbox.com/latest>* > / AGI / see discussions <https://agi.topicbox.com/groups/agi> + > participants <https://agi.topicbox.com/groups/agi/members> + delivery > options <https://agi.topicbox.com/groups/agi/subscription> Permalink > <https://agi.topicbox.com/groups/agi/T87761d322a3126b1-Mba552a7cc5fff5747db2c0ee> > ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T87761d322a3126b1-Mb0f221b040d881ab349d1d17 Delivery options: https://agi.topicbox.com/groups/agi/subscription
