ink transition
play buttonpause icon
view nft
view history

How Conscious Thinking & AI Differ:
Gödel-Completeness

Philosophy of Consciousness: Dennett & Chalmers
Daniel Dennett and David Chalmers are famous philosophers of mind who are well-known for asking hard questions. In particular, the hard problem of consciousness, as Chalmers put it, is that we don’t know what causes qualia, or experienced subjective feelings. This problem comes out of understanding to some extent what consciousness is, but not understanding all the way. In this Spark, we’ll glance at the widely-accepted view, refine it with work by Antonio Damasio, and finally get around to exploring the importance of the work of Kurt Gödel in pursuit of our ultimate prize: a novel differentiator capable of providing clear demarcation between conscious thinking intelligences that can self-justify and self-motivate on the one hand, and non-conscious repositories of stored prior knowledge that are strangely developing something like agency in contemporary computer systems, on the other. 

We have a process we can experience called consciousness that goes on in our biology, but, as Chalmers likes to ask, how does it develop feelings? Dennett likes to say that the stuff Chalmers labeled “easy” was in fact easy, but that the hard part isn’t that we don’t know where qualia come from, it’s more that we don’t know how the internal components of consciousness operate; i.e., suppose someone become conscious of something - alright, what then? This seems like a reasonable critique but Dennett himself seems to take Chalmers at his word with respect to a good deal of the groundwork here that Damasio would like to correct us on.

In the view of modern philosophy, espoused primarily by Dennett and Chalmers, the buck stops here, at the intersection of feelings and conscious thinking. We are forced to admit that we don’t know what consciousness is or how it works at the most granular level if we accept this view. To their credit, the philosophers have in a way shunted us directly into the current we need to be in, here - the only logical next step is to take a quick inventory and find out what else we don’t know. That’s why Damasio is so critical to any contemporary study of philosophy of mind and/or consciousness.

Where Damasio and Gödel meet, regarding consciousness, is a most remarkable intersection. Gödel, with his Incompleteness Theorems, achieved a sort of jailbreak of linguistic thinking by proving in language that an arbitrary theorem cannot prove anything with respect to the world outside itself. In other words, Gödel has taught us that language is incomplete because it cannot justify itself. Damasio has been pushing the envelope from the other direction, studying brains and minds and people in his medical practice, then taking his observations and building a philosophy of mind that still holds unexplained components. These areas of unexplained territory, for Damasio, are always shrinking - just as the mind/body problem has retreated under his leadership from full-blown dualism to an embodiment/enactive complex system (Damasio, 1994).

Antonio Damasio: Re-scoping the Problem of Consciousness
Hanna Damasio was the mind behind the reconstruction via computer simulation of the brain of Phineas Gage, the rail worker who famously had a tamping rod blasted through the part of his brain known as the prefrontal cortex. His personality changed; for the first time science had a strong and well-documented case of physical changes to the brain resulting in the permanent alteration of personality! Both Damasios, then, work in neuroscience, and at least on occasion in problems dealing with the prefrontal cortex.

In 1994, Antonio Damasio successfully put forth the hypothesis that rationality itself was guided by emotion. This is easy to take as self-evident, today, with studies upon subjects such as the metabolism of the attentional guidance network showing relationships between our feelings and even basic cognitive processes such as what we’re aware of, but in 1994 the field was different. 

Back then, Cartesian Dualism was seen as a default view of consciousness - the rational mind segregated from the flippant whims of the irrational flesh. Descartes’ Error won wide acclaim and the Somatic Marker Hypothesis was on its way to reshaping the modern understanding of what a brain does and what reason is. Today every neuroscience student is taught both Cartesian Dualism and the embodiment model that has begun to supersede it in the theoretical understanding of cognitive neuroscience.

Antonio Damasio’s work involves a great deal of philosophy as well as meaningful empirical investigations into neurophysiology and neuroanatomy. He describes consciousness as originating in primarily three centers of the brain, though it is largely a global phenomenon in his view. These key regions are the brain stem, the thalamus, and the prefrontal cortex. Consciousness is made of feelings.

In a nutshell, we have a choice to make - do we want to trust the philosophers that certain problems are intractable? Or would it be better to join Damasio and empirically investigate the metabolic underpinnings of conscious thinking to see for ourselves? 

In many cases, the best thing to do is to investigate and then evaluate results as objectively as possible. We’ll go far deeper into this subject in the full-length book, but for the purposes of this Spark our main goal is to lay out the land and open the field up to discussion. However, as the metabolic theory of consciousness takes shape, additional domains become available to it. 

Just as the Somatic Marker Hypothesis Damasio developed in the 90s is now featured in neuroscience textbooks, science moves forward in the direction of utility, regardless of the approval of philosophers. The theoretical side of the conversation is yet relevant, of course, but being able to measure empirical phenomena will revolutionize things in a way the theorists haven’t quite been expecting. From wanting & liking, to consciousness itself, to conscious subprocesses such as subjectivity & the feelings it owns, the ingenuity of interdisciplinary researchers pushes the investigation forward again and again.

“Subjectivity is a process, of course, not a thing, and that process relies on two critical ingredients: the building of a perspective for the images in mind and the accompaniment of the images by feelings.” (Damasio, 2018, 149.)

To really understand what’s going on in the process we call consciousness, we need to do more than decode signals. We need to understand what is happening where, and why. Medicine and science are converging on a view much like Damasio’s, but there are problems. We can measure the electrical fields and signals of the brain, but we need a robust theory of what the cells are doing to understand the role the fields play. 

In the first book by Antonio Damasio that I read, Looking for Spinoza, the biological force of conatus is forwarded as a sort of backstop, a general story we can tell about what cells want and thus why behavior emerges (Damasio, 2003). Over time, it seems that the level of detail that can be invoked in the description of the behavior of conatus, in Damasio’s view, is increasing as our investigations of neurological phenomena continue to provide more detail. 

Though the heart of the problem of the origin of conatus remains unresolved, the push - driven by thinkers in the neurosciences - to access the contents of conscious thinking through material reductionist language goes on. 

Thinkers including Roger Penrose are working to understand the electromagnetic phenomena & quantum properties of the brain just as others map functionality to region. But there is one thinker who stands above the rest, whose insight serves as a guide to the hardest part of the problem: Kurt Gödel. 

The Magic of Kurt Gödel’s Incompleteness Theorems
A hundred years ago, a logician who took walks with Albert Einstein at the Princeton Institute For Advanced Study ventured a critique of mathematics that would resound for decades to come. Claude Shannon, the godfather of information theory, joked about having to apply the Gödel Incompleteness Theorems during the course of his work at Bell Labs, even. The reason Gödel’s thinking was such a fascination for so many brilliant minds even during his lifetime has its roots in the common questions we all have around language. What are the limits? Is anything we can say actually objective? What can we really know?

These questions and more are profoundly engaged by Gödel’s conclusion in “On Formally Undecidable Propositions of Principia Mathematica and Related System,” that “the statement ‘c is inconsistent’ is not c-provable” (Gödel, 1930). A system’s inability to reach outside of itself is relevant to contemporary discussions regarding AI and doom for obvious reasons, namely that the whole problem goes away if we can figure out how to put it on rails that keep it from spinning out of control. 

Incredibly, the incompleteness of systems in Gödel’s work provides a great deal of support for the ethics of ambiguity later forwarded by Simone De Beauvoir! It also tracks quite nicely through Kant’s Critique of Pure Reason, undergirding the importance of the good will for Kant. Beauvoir’s primary goal is to explicate the ways in which wills can go wrong while reifying the core premise of modern liberal individualism: there is no one way to do it, no system to be made that could would bring peace or order to humanity by codifying and enshrining a set of linguistic theorems. 

Postmodernism is the realization that efficient social systems must be open systems that enable people to bring their own interpretations and applications. Postmodernists embrace the view that there could be a single source of truth. Indeed, such conceptions collapse under the weight of the realization that we cannot linguistically code our way to a perfect description of the world, which falls directly out from Gödel’s work.

Concepts like the critique of Holocaust-era metaphysics espoused by Theodor Adorno in Negative Dialectics are thoroughly empowered here as well, by the basic assumption set. In a nutshell, if all of the Germans had been aware that a perfect system was impossible, perhaps they’d have put more effort into resisting the simple-minded ideologies of the Third Reich.

Popular philosophy in contemporary times has been hamstrung by its trouble in assimilating the idea that the systems of power that fund it are deeply flawed, which in some sense is no surprise. Gödel’s theorems, taken all the way to the bank and completely cashed out in terms of rational minds, agree with Damasio’s position that reason cannot escape its emotional component. In the end, language is a tool that minds can use to computationally model the world. Language will never be able to directly access the world in itself. 

We see limitations in philosophy related to higher education funding, we see minds that, whatever their intelligence level, cannot know things they do not care about, and we see language systems that conform to the desires, drives, and tendencies of their human users as naturally as those human beings use their language to describe their surroundings. This is the path from incomprehension to deep understanding of what language is and how it works. 

What is Gödel-Completeness?
To move forward with our understanding of consciousness and its differences from the sort of phenomena we associate with artificial intelligence, we need to be sure to stipulate adequate definitions of our terms. We should start by evaluating the concept of Gödel-completeness in light of the contemporary state of philosophy of mind.

Gödel-complete systems are systems which do not suffer from the weakness in formal systems of language that Gödel described as incomplete. In essence, a Gödel-complete system is a system which is able to self-actualize, to take action. These systems are not composed of language, but rather exist beyond the bounds of description. Gödel-completeness is a helpful concept because it relates a formal constraint of systems of language to establish a comparison between conscious and non-conscious living things.

Philosophy of consciousness has suffered many difficulties. These range from Cartesian dualism’s rise and decline to the phenomenological enterprise that eventually blended with cognitive psychology in the latter half of the twentieth century. Mind and brain are increasingly thought of as related mechanical reductive systems, but the difficulties with problems such as the so-called “hard problem of consciousness,” first forwarded in the 1990s by David Chalmers, are nonetheless regarded as unresolved. In general, the mind is thought to emerge from the activity of the individual cells of the brain, but difficulties arise when someone attempts to give a complete account of the way in which this happens.

Gödel-Completeness and Consciousness
What are the differences between actual intelligence and artificial intelligence? What can machine learning algorithms achieve that human minds cannot? Are there still tasks only human brains can complete? These and more questions are on the minds of thousands of people, perhaps even millions or billions, around the world. 

In short, the most captivating property of AI is that, when it uses language, people seem to want to infer that it is conscious. It is, of course, not conscious, at least not in the way people are. But it does serve to extend consciousness. So while the fears in the AI Panic Letter are almost laughably inaccurate, AI is still a powerful technology even without consciousness. The more reasonable approach recently put forward by IBM is valuable as we begin to think a bit more about the machine learning side of the consciousness question (Rossi, 2023).

The simplest response to the questions that surround ChatGPT and consciousness, is that certain human intelligence components, including the will, the self, and the worldview, are not likely to be replicated in machines anytime soon. 

One argument to keep in mind is that humans are animals, which have a remarkable capacity to perform cognitive computations that execute in the real world, whereas digitized algorithms that run on machines are not Gödel-complete and can only perform digital computations. Gödel-completeness is a property that is only found in conjunction with worldview; it is not something that can be replicated in incomplete non-living machines. Piccinini views the brain’s computational exercise as a form of analog computation, which in general is a bit slower and less abstract than digital computations performed by silicon microprocessors today.

The enterprise of Worldview Ethics is the development of a specialized vocabulary to meet the needs of the discussion. I.e., we aren’t simply taking sides in an arbitrary debate, our goal is instead to build a new way of speaking about the subject that satisfies objections from all parties to the debate. Worldview Ethics is a subdomain of moral philosophy designed to dialectically develop a lexicon in the space between the rapidly expanding universe of observable phenomena and popular culture such that a layperson’s understanding of the matters being discussed can be rapidly improved. 

In some ways, the job at hand is the same as the one performed by science communicators for decades now. Part of moral philosophy, in many popular works today, is already the interpretation of laboratory experiments to be analyzed for their content with respect to observable mores of the context in which they are found. Behavioral psychology and dialectical behavior therapy already accomplish one of the goals of moral philosophy when they are used to improve the outlook of a patient by applying a study of observable phenomena to the goal of making a mind better. 

The primary goal of Worldview Ethics is to improve the way we talk about the things that minds do and what minds are, as well as what machines are and what computation is. If we can establish something like a shared mutual understanding of the basics of these phenomena, we’ll be empowered to take charge of our own lives along dimensions we were barely aware of previously.

Conscious Intelligence is a Gödel-complete System
My argument in this Spark is that we can benefit from a frame that describes the sort of intelligence that human minds have, which we can refer to as the Gödel-complete type of system. Other systems of intelligence that take language and process it aren’t Gödel-complete because they have no independent existence, but they can still be very powerful in their own right. In fact, with regard to math and now language more broadly, it is becoming both more accurate and more efficient to use computer programs to chart and plot courses through complex layers of abstraction. In this way, the computer is like a front end loader if the brain is the construction worker and the idea being created is a job site.

Gödel-incomplete systems extend human capabilities and enable people to have higher level interactions with language because they are essentially language coprocessors. It’s worth noting that the opponents of these technologies are voicing the same sorts of anxieties that have been around since at least the eighties, and that the objections are not completely unfounded, except when they conflate Gödel-complete intelligence & willpower with Gödel-incomplete systems that are made of language.  

The difference between Gödel-complete and Gödel-incomplete systems is in self-activation, and though experimental laboratories will almost certainly try to build Gödel-complete systems at some point, success in these endeavors is not guaranteed. By turning our gaze toward Gödel-incomplete systems and developing our capabilities to do linguistic tasks with their assistance, we extend our intelligence into fields where it did not reach before. We could learn things that contradict our views or we could see just how right we have been and develop our understanding down to minute detail. 

It is possible that the mysteries of conatus will develop new descriptions and perhaps equally so that our understanding of it will advance, enabling us to increasingly give deeper and more meaningful accounts of these phenomena. In either case, the advancements in technology may baffle, excite, and overwhelm us. The reason ethics is worth the intellectual investment it requires is that it alone can provide us with the theoretical tools we need to coordinate our behavior and move forward to solutions even when things are most uncertain. 

References
1. Association for the Advancement of Artificial Intelligence. (Franscesva Rossi & co.,  April 5, 2023.). Working together on our future with AI. AAAI.
https://aaai.org/working-together-on-our-future-with-ai/

2. Damasio, A. (1994). Descartes' error : emotion, reason, and the human brain. New York : G.P. Putnam.

3. Damasio, A. (2003). Looking for Spinoza : joy, sorrow, and the feeling brain. Orlando, Fla. : Harcourt.

4. Damasio, A. (2010). Self comes to mind: Constructing the conscious brain. Pantheon/Random House.

5. Damasio, A. (2018). The strange order of things. Pantheon/Random House.

6. Daniel, T. Dylan (2018). Formal Dialectics. Cambridge Scholars Publishing.

7. Dennett, Daniel C. (2018). Facing up to the hard question of consciousness. Phil. Trans. R. Soc. B 3732017034220170342 http://doi.org/10.1098/rstb.2017.0342

8. Gödel, Kurt. (1931). On Formally Undecidable Propositions of Principia Mathematica and Related Systems. New York, NY, USA: Basic Books.

9. Maley, C. J. (forthcoming(b)). Analog computation and representation. The British Journal for the Philosophy of Science doi:10.1086/715031
metamask logo
connect with metamask