Where consciousness is concerned, the appearance is the reality.
- John Searle
Consciousness is a permanently hot topic in the philosophy of mind since Rene Descartes. Below we examine the most popular views on consciousness, encounter the famous explanatory gap that makes consciousness so elusive, find why the hard problem of consciousness is real, and point out a novel way for its solution.
Knowledge About Reality
The question of what constitutes knowledge and what is reality is not completely resolved in modern philosophy. There are plenty of views. Among the most plausible ones are:
Structural realism states that objects, properties, and relations, that constitute the world exist independently of our thoughts about them. This belief is also expanded on the unobservable entities. According to this view, science would be miraculous if our theories are not at least approximate descriptions of the real world.
Internal realism points out that we cannot talk about what is really there apart from our categories of understanding and conceptual schemas of things. For example, there may be many different views on the problems of what cats could be considered beautiful, or do numbers really exist.
Instrumentalism is a kind of anti-realism, according to which scientific theories do not reveal anything about unobservable entities. They are only good to predict the possible behavior of observables.
Knowledge is considered to be a true justified belief. This definition, in turn, raises the question of truth. There are also at least three widely accepted views:
The correspondence theory of truth, according to which it is possible to make a linguistic reference to the actual state of affairs in the real world.
The coherence theory of truth states that truth is a relation of coherence between propositions. According to it, we can only make intralinguistic references.
The pragmatic theory of truth states that any internally coherent framework may be considered true. According to this point of view, truth is just a convention.
It is from such rugged terrain we are compelled to start our journey to the ultimate questions of reality.
Reductionism
Reductionism is a way to describe complex systems as a sum of the constituting smaller parts. For us here is only important how it touches the other question of no less prominence, the question of existence. Existence is usually defined as the ability to interact with reality. Quoting Charles Peirce: "When we say that a thing exists what we mean is that it reacts upon other things". The notion of supervenience captures relations between the different levels of existence. But what levels of existence we should consider as being truly real? Are there cats or only moving blobs of atoms with pointy ears and whiskers?
Because infinite reduction seems absurd, we need to postulate the existence of some basic level of reality, atomic objects of which along with relations between them define what is real in the ontological sense. At the current state of scientific development, we perceive this level as quantum fields, although actually there may be something more profound. Probably, we will never be able to figure out what it really looks like, but we are free to reason about it. For our purposes here, let's assume that this is an interconnected grid of atomic joints with some simple properties yet to be discovered. It defines the space itself, and elementary particles are represented as relations between the joints of this grid.
If there is an ultimate basic level of reality, does it mean that all other objects could be referred to as concepts, reducible to it? It seems so. Although usually, we refer to objects as they are defined by their properties on the level of existence that is convenient for the situation. But not everything is so simple. Some properties (so-called secondary qualities), such as color, are subjective and make sense only by being considered in the context of human perception. There are also numbers that evoke the eternal question of universals, particulars, and potentiality, opening the chasm of modal logic. Some problems in modern philosophy have no universally accepted solution. Are we brave enough to turn them bottom up and examine them in a new light?
Consciousness
In our brains, we have models of reality organized as frameworks of interrelated concepts. The notion of consciousness captures our ability to meaningfully interact with reality by utilizing these models. The ability to put the world under concepts is called access consciousness. It was also noticed, that consciousness is accompanied by rich private internal experiences, that take their roots in perception. Such experiences are called phenomenal consciousness, and their very existence is found problematic, despite all common sense. In the vein of ontological monism, which is currently the most popular point of view, there is, as always, a multitude of approaches to the characterization of consciousness:
Behaviorism moves away from internal mental states and considers only the observation of external behavior as the only scientifically-plausible way to define consciousness.
Identity theories are trying to find the correspondence between the internal mental phenomena, such as the experience of color or pain, and the processes in the brain. For example, pain is considered as the stimulation of c-fibers.
Functionalism moves away from the wetware of the brain and operates by more abstract notions of functions and subsystems.
Eliminativists deny the existence of internal experiences and say that it is only sensible to discuss neurobiological processes (Matter and Consciousness of Paul Churchland is a notable example).
So, what is the problem, namely the hard problem of consciousness? It seems, that there is no plausible way to reduce to matter the qualitative internal experiences (qualia) of phenomenal consciousness. In more simple words, if you are looking at a green apple, and there is nothing green in your head, where is the green? This is called the ontological gap. The crisis is so deep, that it makes some philosophers resort to ontological dualism. A prominent example is 1996 book The Conscious Mind by David Chalmers, who states that the consciousness could not be functionally analyzed. Dualism implies that reality consists not only of matter but also of some distinct mental stuff. Such a position raises the problem of mental causation.
There is also a cohort of strong reductionists, who believe that consciousness in general is reducible to functional physical processes, and we have everything what is necessary to explain it. But they usually deny the problem of qualia, such as Daniel Dennett in his 1991 book The Consciousness Explained. Can we solve the hard problem by staying inside the bounds of ontological monism? Can we find a framework needed for the functional analysis of Qualia? Maybe it is possible. Behind the fixation on objects and properties of classical philosophy, we may just not see the right thing to reduce qualia to. But, at first, we need to make an excursion into some theory.
Automata
Finite state automata are computational models that are defined by the input/output alphabets and the state transition table. Their operation is simple: by receiving an input symbol automation produces an output symbol depending on its current state, possibly shifting to another state. The Turing machine is an example of a generalized automaton that could perform any symbolic computation. Most notably, automata can parse or generate language, which involves some meaning. A physical implementation of an automaton is also characterized by some physical process of state transitions.
Neural networks comprise a more sophisticated class of automata. Among other tasks, they are capable of pattern recognition based on the set of features fed to the network. At first, the networks are trained by encoding features as the computational weights of their synapses. Being exposed to an object, a network is able to determine does it belongs to the class of objects the network was trained on.
It is believed that the brain consists of a huge number of automata represented as neural networks. For example, the fusiform area of the brain is associated with face recognition. There are also may be even Jennifer Aniston cells, which fire when a Jennifer Aniston image is demonstrated to an observer.
A neural network tries to distinguish a duck from a rabbit.
A little philosophical vignette: it appears that all universals are instantiated at synapses in the neural networks of our brains and are a part of the more broad cognitive machinery: generalization. Their interpersonal similarity probably has something to do with the fact that all people belong to the same biological species, we all share the common body of knowledge, and the laws of physics are the same everywhere.
The Architecture of the Brain
The cortex of the human brain is a folded sheet of neurons with an area of up to two square meters. There could be up to 100 billion neurons, stacked in layers with up to 6 neurons in each. Neurons are connected through synapses between axons and dendrites. A neuron may receive around 15000 input connections.
On the basic level, neuronal ensembles perform computations by transforming input action potential frequencies to output frequencies based on the properties of their synapses and possible external modulation by other neurons or chemicals.
On the more general level, the brain is composed of the executive area located at the prefrontal cortex, and motor/sensory areas situated in the frontal, parietal and occipital lobes. Hippocampus and the temporal lobes are associated with memory.
In the sensory areas, neuronal signals from sensory receptors are transformed and integrated being passed through a number of gradual stages of processing. In the case of visual perception, there are four types of photoreceptors sensitive to the intensity of light in general (rods) or to the corresponding wavelength (RGB-cones), connected in an antagonistic manner. The intensity of each channel is modulated, most likely, as frequency. Then, is sent through the optical nerve to the occipital lobe. There it is processed by two distinct neural streams that recognize shapes and motion. Color is represented in the visual area V4.
The thalamus, a neural hub in the center of the brain, acts as a sensory gating device and is a component of many reentrant neural loops. Such loops may facilitate the integration of sensory inputs and self-amplification of unconscious brain states that eventually become available to access consciousness. In more detail this is described, for example, in the book Consciousness and the Brain by Stanislas Dehaene.
Where is Science Now?
Currently, scientists can only search for neural correlates of consciousness (NCCs) in the hope to unveil processes in the brain that are happening as some events become conscious. A very popular example could be found, for example, here.
Another approach is to measure the integrated information of the system to determine whether it possesses consciousness. The work is led by Giulio Tononi.
Predictive coding theory tries to explain how the brain builds models of reality based on statistical sampling and prediction errors.
Dual process theory aims to figure out the differences between effortful and sequential conscious processes and effortless and parallel unconscious ones.
Semantic pointer architecture provides a model of how features, images, and cognitions could be encoded and implemented on the neural level.
Neither theory, though, explains where is the green. A more philosophically-inclined answer is required.
Where is the green?
Now we know a little bit more about the brain and automata. Can we proceed with the functional analysis of qualia? Can we pinpoint the green on our hypothetical grid at the foundational level of reality? Or can we reduce it to elementary particles as we know them? Is green somewhere in between? Not exactly. If qualia would be something fundamental, we would have found them long ago. Does this mean that Daniel Dennet is right, and in the perspective of the view from nowhere qualia do not exist? Yes, he may be right! Qualia seem undescribable and could only be pointed by example exactly because we can not make a reference to them in the external reality. The only remaining options are to search somewhere else or to apply some other criteria of existence. What a proponent of monism would do? It appears that qualia may exist, but only in a personal world. They may have meaning only for the person who inhabits this world. How can we invent a distinct personal world? How can we make qualia real in it?
To solve our problem, we may introduce yet another grid of reality of the second order made of the action potential frequencies roaming through the automata of the brain. Such a "world" should have its own set of atomic entities and "physical" laws governing them. In a poetic stance, we can imagine it as an interference pattern on the fundamental grid of reality.
Although the picture looks a little bit New Agey, the connectome of the brain includes a topologically-connected space.
Our frequency-based grid looks not as firm as the fundamental one, but its existence is as conceivable as the existence of the most worthless philosophical zombie. Isn't it a kind of dualism? Definitely not. This is just another reductionist layer of existence that may be supervenient, for example, on ionic currents inside neuronal tubes. Being considered in the context of experience, it adds a whole new ontological domain. It may be populated with entities to which qualia are reducible. As in the case of the fundamental grid mentioned above, such entities may also be just relations between the "atomic" elements of our hypothetical neuronal grid. They are real in the sense that the automata of the brain can react to them.
How this grid is constructed? What entities may exist there? These are the questions for future researchers. Judging from the pattern of neuronal connections, for example, in the visual area, and from the nature of qualia, the space defined by this grid may be not just 3-dimensional. Qualia may be projections of some hyper-dimensional objects. Note also that this is not an emergentist theory. According to Kim Jaegwon such theories are useless to reductionist science. Because our second-order grid is dynamic and is supervenient on the matter, qualia may finally be reduced to basic neurobiological processes such as frequency modulation. In the sense of Whitehead's process philosophy, they do not exist but are becoming. They must have meaning only to the automata submerged into the neuronal second-order reality of a person. This explains why qualia are non-conceptual. Because they are confined to the personal reality, only this particular person could make a reference to them. To an external observer, they may look like meaninglessly waving folds of matter. Because of that, introspective methods that can interfere with these processes are required to explore them.
The existence of the process of frequency modulation captured as a time series graph.
How qualia are experienced?
Now we have a framework for understanding qualia. But the division of consciousness to access and phenomenal types may still seem problematic. We say that qualia must have some meaning. It is still unclear, how the qualitative phenomena are experienced and made sense of. There are two widely accepted views on conscious perception:
Homunculus theories - there is a homunculus that observes the conscious events, for example, the executive automata. This view is prone to infinite reduction.
Global workspace theory - there is no observer. Conscious events just "shine" in the global workspace and are available for all the interested automata constituting the brain. They may process these events in parallel.
Because our automata are citizens of both personal and external worlds, we are able to touch the external reality through our perception, the simulated fabric of which is embedded in the matter of our brains. We have automata that produce some states based on the perceptual input and automata that make sense of these states in the context of their models of reality.
Ray-tracing a Calabi-Yau manifold may be a good metaphor of a model of the perception of qualia with reentrant processes under a homunculus theory.
No comments:
Post a Comment