Can Laws of Emergence Settle the Question of Whether Artificial Intelligence Is Possible?

If you look at the history of the cosmos for the last 14-some-odd billion years, the common theme is one of ongoing emergence—of increasing novelty and complexity emerging out of materials, entities, and properties that are less complex and qualitatively distinct. For example, shortly after the Big Bang, there was a sea of energized subatomic particles too energized even to form atoms. As the universe cooled and expanded, atoms were finally able to form, making chemistry and the periodic table of elements possible. As gravity gradually allowed matter to pool and congeal into stars, and as nuclear fusion eventually became possible, the nuclear furnace inside the first generation of stars enabled the formation of the heavier elements. Later generations of stars produced even heavier elements. As planets formed around at least one of these stars—Earth in particular but likely elsewhere in the galaxy and universe as well—self replicating systems of matter, what we call “life” arose from matter that was previously not alive. These primitive life forms evolved and adapted, becoming more complex and specialized to fit various niches from generation to generation, beginning with single-cell organisms to the first multi-cellular life forms, and eventually into the thousands and thousands of species that have populated the Earth (or other life-hosting planets) to this very day. With the evolution of language and symbolic representation, our distinctly human traits, the richness of human culture became possible, manifesting itself in innumerable forms of artistic, religious, and technological expression and novelty previously unseen in any life form on the planet.

This is a familiar story to any student of cosmology or of evolutionary biology. And yet we humans sometimes view ourselves as the pinnacle of creation instead of a stepping stone along the journey of the unfolding emergence of the cosmos, which continues to this day with every invention, every evolutionary twist, every artistic creation. Presumably this ongoing emergence will continue to unfold for billions more years, with surprising twists of novelty and complexity at every turn. From the initial sea of subatomic particles following the Big Bang, no one could have predicted the richness and qualitative differences of the elements that make up the periodic table. From the periodic table alone, no one could have predicted the richness of the biological world when those elements combined in ways that made self-replication possible. From the earliest self-replicating biological systems, no one could have predicted the vast diversity of the biological world with species so varied as to fill every possible niche on the planet. Even as recently as a million years ago, no one could have predicted the richness of human culture, spirituality, and technology. At every turn, emergence is marked both by novelty and by qualitatively distinct phenomena that are not predictable from the entities and properties that existed beforehand up to that point in the history of the universe.

An interesting question in the field of emergence theory is whether there are laws of emergence. Even if each emergent phenomenon or entity is genuinely unpredictable, are there laws of nature that govern which higher-level emergent properties are instantiated by which types of physical systems, even if those emergent properties had been previously unseen in the history of the cosmos? For example, the laws of nature that enable a modern smartphone to work as it does were arguably written into the fabric of the cosmos long before there ever was any such thing as an actual smartphone. This says to me that there is some form of deterministic creative potential in the laws of nature and in the nature of matter itself, some untapped sphere of possibilities that merely require the right environment, the right arrangements, and the right background conditions to reveal themselves in actuality in the unfolding evolution of the cosmos. A smartphone was always a possibility, even from the earliest moments following the Big Bang in which there were, as yet, not even stable atoms, much less molecules and the necessary materials and elements to build a smartphone. Yet, the potential for a smartphone—and for human beings, and for culture, and for all the technological inventions with their yet-to-be-discovered properties and phenomena—was there from the very beginning of the physical universe.

While the laws that govern emergent properties may seem mysterious prior to their discovery and instantiation in the unfolding emergent complexity of the universe, which, again, is still ongoing, in hindsight these laws of emergence do seem to be discoverable after the fact. The periodic table, for example, and the way it arises from the Pauli Exclusion Principle are well understood by physicists and chemists at this point. Similarly, at least some of the mechanisms of evolutionary biology—genetic mutations, natural selection, the origin of species, and so on—are also understood. While each of these mechanisms, and all of them together, was responsible for the emergence of all of the biological (and eventually cultural) diversity seen on the Earth today, and while each leap in novelty and complexity gave rise to then-new and previously unseen properties and phenomena, it does seem possible to discern something akin to genuine laws of emergence and laws of complexification in retrospect after each evolutionary leap of emergence in the history of the natural world.

The laws of physics, chemistry, and biology are fairly well, if still imperfectly, understood. In contrast, if there are analogous laws of the higher-level sciences (the higher-level sciences themselves having been made possible by the evolution and emergence of new complexity and new phenomena in the history of the cosmos), they are seemingly less well understood. If there are laws of emergence of consciousness, of psychology, of culture, of sociology, of technology, and so on, they are still in the process of being discovered, in part because the phenomena in question have not fully emerged or developed—the richness of human cultural and technological innovation has only just begun, a blip of time in the overall future history of the cosmos. Presumably, however, such laws of emergence do exist, and all of the evolutionary and emergent human potential was already written into the fabric of the cosmos, not just in the laws of nature studied by physicists but in the laws of emergence and complexity that do and should count as laws of nature, as much as any law of nature studied by physicists.

One day, perhaps far in the future, the laws of emergence governing how exactly consciousness, for example, arises from the bits of grey matter and from various arrangements of neurons in the human brain may very well be discovered. The emergence of consciousness is still a mystery to contemporary neuroscience, despite the fact that neuroscience has come a long way in revealing which portions of the brain are responsible for which cognitive activities; the mechanisms and laws of emergence governing exactly how and why consciousness arises are still unknown. Presumably, however, there are indeed laws of emergence—laws of nature—that govern how and why consciousness arises from the brain, laws which may be discovered and known at some point in the future.

Should laws of emergence become better known and understood than they are today, specifically about the emergence of consciousness, it may very well be possible to answer some of the fundamental questions about the possibility of artificial intelligence. At this moment, because we do not yet understand the laws of emergence that govern the emergence of consciousness, even of our own human consciousness, we do not yet know whether it is possible for other types of physical systems (silicon-based computational systems, for example) also to instantiate the emergent property of consciousness, in some form, however similarly or distinctly to human consciousness. As laws of the emergence of consciousness become better understood, however, it may at some point be possible to know for certain that various arrangements of matter (gray matter or silicon or otherwise) give rise to various types of emergent properties like consciousness.

At present, because these presumed laws of emergence governing the instantiation of consciousness by physical systems like our brains (or by artificially intelligent computational systems) are still unknown, we must resort to cruder methods for discovering whether a computational system has genuine artificial intelligence (i.e., whether it is really consciousness), such as the well-known Turing Test for artificial intelligence. The idea behind the Turing Test is as follows: if a computer system’s responses are indistinguishable from those of a human being, over time, then we should conclude that the computer system is conscious—or, at least, we would have the same reasons for thinking the computer system is conscious as we do for thinking any other human being beside oneself is conscious. Should the laws of emergence governing the emergence and instantiation of consciousness become known and understood, however, it may very well become possible to know for certain that any particular physical system, such as an artificially intelligent computer, is instantiating consciousness, and perhaps even which particular mental states are being instantiated, and, more importantly, why those mental states are being instantiated. While the Turing Test may not settle the question of artificial intelligence with certainty—as the genuine consciousness of a computer system that passes the Turing Test would still be known only by inference and by analogy to human consciousness, should laws of emergence of conscious be discovered and understood, it may very well be possible to know with certainty which types of physical systems can and cannot produce the emergent property and phenomenon of consciousness (self-awareness, intentionality, perception, emotion, etc.).

Perhaps our understanding of laws of emergence will never actually keep up with the actual unfolding of emergence, novelty, and complexity as the universe and things it contains continue to evolve and produce all the future phenomena, life forms, technologies, and cultural developments yet to come over the next several billion years. Presumably, however, those laws of emergence exist and are discoverable in principle, although perhaps only in retrospect after the relevant evolutionary leaps of emergence have occurred. This is an inherent tension in the discovery of laws of emergence. On the one hand, laws of emergence do seem to be deterministic. Fundamental laws of physics seem to give rise to laws of chemistry, which in turn give rise to laws of biochemistry, which give rise to laws of neuroscience, which in turn give rise to human consciousness and all of its cultural and technological richness. On the other hand, each new emergent phenomenon was previously unseen and unknown in the entire history of the cosmos before it actually evolved or emerged. Human consciousness, technology, and culture—including the aforementioned invention of the smartphone—were unpredictable before they actually occurred in the unfolding history and emergence of the cosmos. This tells me that there may be something inherently unpredictable about future leaps of evolution and emergence—the exact phenomena and properties instantiated in the future might be genuinely novel, never before having been seen in the history of the cosmos until they actually emerge (an ongoing process continuing into the future), even if there are deterministic laws of emergence. In principal, however, there should be laws of nature that govern which types of emergent properties are instantiated by which types of lower-level systems, even despite the novelty and surprise inherent to the emergence of any new phenomenon at a particular point in time or in the unfolding evolution of the universe.

To me, the most fruitful future area of inquiry for the natural sciences would be to focus on drawing deterministic causal connections between lower-level entities and phenomena (like brains) and higher-level entities and phenomena (like consciousness) in an attempt to discover and to genuinely understand the laws of emergence that govern why qualitatively different phenomena seem to be instantiated with every leap of evolution and emergence. Only by discovering and genuinely understanding these yet-to-be-discovered laws of emergence can we settle some of the stickiest and most persistent philosophical questions in the history of philosophy: the nature of consciousness, the relationship between higher- and lower-level properties and entities, the possibility or impossibility of artificially intelligent (genuinely conscious) computer systems, and so on. It is not enough to recognize that emergence occurs (itself an undeniable fact from even a cursory look at the history of the cosmos, or of the history of our own planet). Only be discovering the laws of emergence that, in principle, reveal how and why various emergent properties are instantiated by lower-level physical systems (such as the way consciousness arises from the arrangement and activity of neurons in our brains) will we ever truly understand the qualitative richness and diversity of the entities and phenomena that make up the still-unfolding evolutionary history of the cosmos as a whole—of the life forms and culture and technology it contains, those in existence today (having already emerged) and of those still to come as the universe continues to evolve, and as new entities and phenomena and complexity continue to emerge in all the wondrous and unpredictable ways yet to be seen.