Second Order Consciousness, Sapience, and Wisdom
George Mobus
[Longish!]
On and off I have played around with the nature of human consciousness and its relationship with sapience as I have been developing it. Specifically in the working paper "The Components of Sapience", in a section titled "Consciousness and the Mind Architecture", I hint at the relationship but admit there that I could not give a more full accounting of what that relationship was, or how it might work.
In this post I want to share with you some of my further speculations on the matter and tell you what I think that relationship is, and, to some degree, how it works. This may seem obvious to my more sapient readers! But it is something of a breakthrough for me, so bear with me if you can.
Before I get to human/sapient consciousness I need to address the more, if you will, ordinary kind of consciousness, the kind we share with all mammals and possibly birds. This is just 'plain' consciousness or being aware of the world and ourselves in the here and now, the processing of 'images' in working memory and the role this form of consciousness plays in mammalian behavior. It may surprise you to learn that what you are basically conscious of (the world in front of you and your decisions for action in that world) are actually after the fact of perception and action selection! That is, ordinary consciousness, or what I will call First Order Consciousness (FOC) is just a reflection upon what the rest of your brain (intelligence + creativity + affect) decided to do before your phenomenal experience even became aware of it.
But it does serve a purpose.
First Order Consciousness — FOC
Is there anyone here who doubts that their pet dog or cat is conscious? Most mammalian pet owners have little doubts that their pets have memory encodings (representations encoded in neural circuits in association cortex, e.g. anterior parietal lobes) of them, their selves, and meaningful cues such as when the owner gets a can of food out to open - the cue that dinner is served. Most of the encoding is accomplished by basic Pavlovian and operant conditioning. But the owner also knows that the pet has desires and even intentions (as when your dog brings his favorite toy and begs you to play). That is, the pet is aware of the situation in the environment, has some inner desire, formulates a course of action, executes that action, and responds to what the environment (esp. the pet owner) does in response to that action. This is FOC.
Now it is true that we cannot literally experience any phenomenal percepts (mental experiences) that the animal has. We cannot say what it experiences directly, or has been popularly said, know what it is like to be another kind of animal. But few who have any experience with pets doubt that they are aware of and interact intelligently with their worlds in essentially the same way we do when were not thinking about being aware (like when you arrive at work and realize you can't recall thinking about driving as you did it). The latter invokes a notion of Second Order Consciousness (SOC) which is what I will be focusing on a bit later. But we cannot grasp what that means until we get a working view of FOC and what is going on in the brain to produce it.
In consciousness studies, everything from neuroscience to philosophy, a recurring question arises in the evolutionary-inspired form of "What is consciousness for?" Some thinkers have come to the conclusion that consciousness is just an epiphenomenon, that is something like an accidental but meaningless result of brain functioning. In other words, in the extreme version of this, consciousness serves no real purpose, it is just a sign of what is going on under the hood. Now there is some justification for this when we think only about the real-time version of conscious awareness.
The reason some people come to this conclusion is that there is a growing body of evidence showing that conscious awareness of our own decisions comes some time after (maybe several milliseconds) our pre-motor and motor cortices have taken a decision and initiated motor outputs! In other words, our conscious processing parts of the brain had nothing to do with making a 'conscious' decision, but only acts to display that decision (for some inexplicable reason) in working memory. In yet other words, free will is an illusion.
Well, yes and no. The interpretation of the evidence of a delay between the actual initiation of action and the conscious awareness of it is still open. We know that when we repeat certain behaviors, particularly motor skills, we form habits or non-conscious responses to the appropriate situations. This is a means of developing a more efficient reaction to those situations. Rather than having to think through what needs to be done each time the situation arises, we simply learn to do the right thing and then it becomes automatic after sufficient reinforcement. This capability for learning new behaviors, practicing and then having them become automatic seems to be handled in the brain by the lateralization of the hemispheres. According to Elkhonon Goldberg's theory (The New Executive Brain, Oxford University Press, 2009), the right hemisphere is devoted to novel situation/action processing, where one has to be actively conscious of the situation and 'consciously' choose the right action (as well as consciously monitor and control the follow through making small adjustments as needed). After some number of times doing the same things the neural pathways and motor sequence programs are strengthened to a point where they can operate without 'conscious control'. What may be happening is that the right hemisphere is actively engaged in constructing and monitoring the progress of the new behavior while the left hemisphere is mirroring the results and doing the actual control output to the motor cortex. With repetition the left hemisphere strengthens the synaptic pathways that perform the 'correct' or successful (hence rewarded) actions. At some point the right hemisphere's guidance is no longer needed. The left hemisphere takes over and the memory patterns (engrams) are hardened to long-term memory traces. So, eventually, the pattern of situation-response or behavior becomes automatic and not under conscious control.
This is an intriguing theory and though it is not as thoroughly developed as it might be, it does help explain the results of the time delay between action selection and conscious awareness. If the behavior has become routine or a habit, then the left hemisphere has simply automatically done its job and the need for intervention is not there. However, the fact that the action took place might be useful information for another purpose and so the fact of the action is recorded in working memory and accessible to monitoring by the areas of the prefrontal cortex involved in executive control. The reason is that in real life things change and if all behaviors became permanently automatic this could have dire consequences for the organism. If the situation response were to fail to provide a good outcome then the response would have to be altered or even possibly abandoned. Thus, if needed, the right hemisphere would need to be called in again to reconstruct the appropriate behavior and start the cycle of learning and engraining over again.
Functionally it might look like Figure 1.
Figure 1. A functional view of conscious control over action selection mapping in the mammalian brain.The perceptual system (the neocortical regions) and the limbic system (e.g. amygdala) receive information about the immediate world from the senses (time t1). The limbic system also functions as an evaluator of the goodness or badness of the environment and can compare the old state with the new state (time t2) to determine if the action taken made things better or worse (or, of course, neutral). Under normal conditions the perceptions of the situation in the immediate world result in activation of the action selection map, which then activates a motor output program as dictated by the situation; one that had been learned previously. The program should benefit the organism by putting it in a better situation, or preventing a worse situation from obtaining.
Figure 2 is a very simplified diagram of the concept of an action selection map. The 'map' refers to a large array of initiating points for motor programs. Each element in the array initiates a specific behavior sequence. During development the organism the map constructed the translation from a perceptual analysis, called a situation analysis, to an appropriate behavioral output. in vertebrates prior to mammals a great portion of this map was 'hard wired' under genetic control, with very little adaptability (learning). Evolution had been the determinant of what behaviors suited what situations for animals that had relatively simple and stable environments. Simple, means not much in the way of food choices and not many smart predators to worry about. Think of a frog sitting on a lily pad. Fly goes by and flash goes the tongue. Snake happens to be nearby then jump into the water and swim into the mud. Pretty simple and automata-like (more on instinctive behavior below). To be sure, some learning, in the form of short-term habituation or sensitization, or slightly longer-term behavioral shaping, were present in these animals, but by and large their behavioral repertoire was fixed.
For animals like mammals that were evolving to fit more complicated niches something new was needed. Food opportunities were more varied and shifted with seasons and longer-term climate shifts. Predators were other mammals and getting smarter. What was needed was and adaptable mapping system, one in which the appropriate response could be learned and nuances of behavioral sequences could be adjusted as needed. Enter the adaptive map.
Figure 2. The action-selection map is essentially an array of possible initiating points for a combinatorial set of motor programs. An analysis of the situation from the perceptual system activates the appropriate initialization point in the early motor cortex. Thereafter the activation chain follows a pre-learned sequence of specific muscle (group) stimulations that result in specific behavior. All of the arrows in this figure are assumed to have been learned from actual experience (as in Figure 1) over time. Note that the role of the lower brain and brain stem have been ignored in this diagram. Two, of many possible initiation points are shown in the map. The situation detected by the perceptual analysis system maps to the specific one (in gray) that sets off a learned or instinctive sequence of motor stimulations. The sequence determines which muscle groups are activated in sequence. Not shown is the role of the cerebellum or the motor relay nuclei in the lower brain stem. Motor sequence options run from left to right. The arrows show the flow of activation that produces the final behavior response. These kinds of circuits have been successfully demonstrated in more complex robots!
This same causal loop (perception, action selection, motor output or action taken, change in the world, new perception) was at work in instinctive behaviors in earlier animals (reptiles and amphibians). The difference is that instincts were hard wired action selection maps whereas with later animals, mammals in particular, the map is subject to modification with experience by virtue of learning capacity in a new cortical structure, the cerebellar neocortex. The role of the monitor circuits, shown in Fig. 1, the new prefrontal cortex, is to keep track of performance and activate learning (and also overriding previous learning) as needed to assure longer-term good decisions. Mammals gained adaptive control, or, in other words, adaptive behavior that could adjust complex behaviors over longer time scales as an individual's environment changed in unpredictable ways. This was quite an evolutionary innovation, even though it took advantage of previously evolved structural architectures and cellular/circuit functions (I will explain below how evolution of the brain has been a process of reusing old structures in new ways).
This higher-order function, of monitoring and adjusting to maximize long-term quality of the action-selection map function, is tactical control. It is monitoring the interactions of the organism with its environment and acting to assure that the animal learns (and re-learns when needed) the optimal set of actions to take in the array of situations the animal might find itself in. As animals evolved more elaborate and flexible circuits (another subject some day) this permitted them to adapt their behavior to a wider array of possible situations, i.e., a more complex environment.
Here we have a distinction between automatic reactions handled by the action-selection map when the synaptic links between perceptual inputs (situation analysis) and motor program outputs have been sufficiently strengthened. Until these links are strengthened it is up to the monitor to guide the process based on the evaluative feedback gained from actual experience.
There is, as you would guess, far more complexity to it than captured in these figures. For one thing there appears to be some kind of mechanism within the monitor that focuses attention on the particulars of the environmental situation and on the actions just taken. This is held in what is called working memory by means not yet understood. Possibly this is what allows the prefrontal cortex of the right hemisphere to 'play' around with options and then guide the formation of new synaptic connections as it attempts to better shape the situation-action selection mapping. This attention focusing mechanism is what gives rise to the phenomenological experience — being aware of the world and its effects on our organism.
This is purely speculative, of course, but there is one more piece to the story. It has to do with a very ancient part of the brain, in the brain stem, that seems to be involved in keeping track of what the world does to us versus what we do to our selves. You know that when you pinch yourself it doesn't seem to hurt, at least not as much as when someone else does it. That is because this very ancient neural circuit registers the difference between self and other. When you do something it recognizes this and inhibits any untoward response your body might have to a stimulus. Otherwise you would jump every time you inadvertently touched yourself. You would think something out there in the world had touched you and you would react. This same region of the brain seems to be involved in general stimulation of the prefrontal cortex (this is the bit that seems to be stimulated by caffeine), the wakefulness stimulus. I suspect that it is also the key in understanding the phenomenological experience of selfhood. This function (along with pain sensation) is tied up with the fact that an animal can't act as if it were just an observer of things happening to its body. It has to have a sense of things happening to its body. In other words, it has to have a sense of ownership, of personal integrity that must be maintained.
First order consciousness, then, is primarily responsible for the real-time tactical management of the body in its relationship with the environment. It is constantly monitoring the goodness of results from lower level action decisions (made subconsciously) and when the results were not as favorable as they should have been, it intervenes to modify the automatic response and learn a new situation-action selection mapping. This is the basic awareness all animals have. Of course in humans it is greatly elaborated by having much greater memory storage capacity, more novelty processing capabilities, etc. But essentially, humans share this basic consciousness with our mammalian cousins. This is not what makes humans different. Evolution, however, had learned a neat trick in pushing the brain into new territory. The development of a new circuit to handle this interesting tactical control, the neocortex and prefrontal areas in particular, was, as it turns out, just an elaboration on an already existing architecture. The monitor pictured in Fig. 1 is actually just another map process accreted onto the existing action selection map by virtue of some interesting mutations and selection.
One mechanism that seems to provide a lot of the raw material for natural selection to work on is to have an 'accidental' replication of an existing structure, like the paleocortex, in a redundant copy. At first the copy probably only serves as a backup and probably takes up resources without any immediate payback. But over time, if it doesn't prove outright handicapping it might start to diverge in function from the original structure and if the new functionality proves useful... voila, something new is born.
Imagine, roughly speaking, the action selection oval above being replicated in some kind of mutation event. At first it is just a redundant action selection processor with the same inputs and outputs. It doesn't actually add anything to the animal's capabilities. But then imagine that over some length of time other mutations start to creep in and instead of getting its inputs from the perceptual system directly it gets them from the original action selection processor. Instead of outputs to the motor system it sends its outputs back to the action selection processor. Over enough time it takes on a new ability to implement a new kind of action selection mapping, not from external stimuli, but from stimuli from the old action selection map. It then produces a 'motor' output that, instead of activating a motor program, activates a learning process in the action selection map. In time, the new piece of hardware (or actually wetware) evolves into the tactical controller because having adaptive tactical management proves to be highly successful and gives the possessing animals great advantages in terms of exploiting more complex environments.
This 'trick' might have come in handy a second time. As mammals gave rise to primates, and primates became ever more clever in their tactical living, especially in social groups, they gave rise to hominins who had a spark of something even more interesting than mere tactical control. Something new, but derived from something very old, had emerged. A new patch of tissue at the very tip end of the forebrain started to elaborate and expand. And the later hominids, especially Homo, started showing evidence of something like tactical control but with much larger time and space dimensions — strategic control for the management of self and group.
Second Order Consciousness
So if all mammals have some degree of FOC, is human consciousness really different or is it just more of the same only 'bigger'?
I believe it is actually both the same and different! We do share the basic FOC with our mammalian predecessors and co-inhabitants of the Ecos. But something new has been added on top of the FOC. A new level of consciousness that performs a new set of functions and produces all that is unique in human behavior, such as true symbolic language and, as I will argue, sapience.
Imagine that mutation pulled the same kind of trick used to generate the new tactical control function of FOC. Figure 3 shows a functional diagram after this trick was pulled and a long time for divergent evolution between the components had passed. The new component is an elaboration of the previous monitor processor, only now it monitors much, much more.
Figure 3. A functional view of second order conscious (strategic) control over the tactical control system (first order consciousness). The new upper layer of extended working and tacit memory and the new kind of 'monitor', form a strategic management function that is able to model the world from historical experience and project possible future worlds based on reasonable evolutions of the present world.
What has happened here is that the entire complex of the tactical controller (action selection monitor), working memory, and tacit knowledge data base memory (I have not shown explicit memory in these diagrams since it is part of the perceptual and action selection map systems) have been replicated and expanded. They took on expanded roles to handle complex social relations as well as temporally extended engrams. The hominid brain developed the ability to have memories based on longer histories but also memories of possible futures. The brain developed an ability to think about the future. This gave it a new capability in terms of monitoring and controlling the tactical level. It could formulate more elaborate plans that took on a strategic nature.
The brain did not just create a new cortex, as happened with the neocortex deriving from the paleocortex. The constraints of the skull overall volume might have negated this possibility. Or it might be that there simply was no more benefit to be gained from an entirely new cortex overlying the neocortex. Instead what seems to have happened is the frontal lobes, where the original action selection map was located, produced additional patches of neocortex that became what we now call the prefrontal cortex (indeed the precursors for PFC can be seen in some of the smarter mammal species like dogs and elephants). The neocortex had evolved to handle something like modular functionality in patches within a single planar structure. If you needed new functions you just replicate an existing patch that has a related function and then let it evolve divergently to perfect the new one. The prefrontal patches appear to have originated and evolved in this manner. The skull shape had to evolve to accommodate this but since simians and some apes had taken to arboreal lifestyles and needed eyes in front of the head for binocular vision, the availability of the space above the eye sockets became available. The forehead had room to protrude outward and forward without seriously risking birthing problems (until we get to humans who seem to have pushed the envelope).
The PFC in humans has exploded, in evolutionary terms. It now comprises about half of the area of the entire frontal lobes (merging into the pre-motor cortex in the latter). During hominid evolution we can see this structure rapidly growing relative to the rest of the brain even as the whole brain got larger. Then with the emergence of later species in the genus Homo we see it develop further still until in humans it occupies a significant fraction of the whole neocortex. And with its explosion, behaviors in the hominids started to change. Homo erectus may have been the first hominid to migrate out of Africa (apparently sapiens did so later in several waves). Clearly hominids had developed considerable capacity to learn new behaviors, but more importantly, they had started to think about the future and consider alternative futures and their consequences. Most animals that end up leaving their normal habitat are forced out by some calamity. Early humans appear to have been wondering what was on the other side of that next range of mountains.
They also started developing more complex tool making. This involved intentionally sharpening spear points and chipping rocks to create cutting tools. They intentionally took control of fire. All of these behaviors required some forethought about what the future uses would be rather than just breaking off a stick for immediate use when you saw a prey animal. Foresight is involved in yet a newer kind of thinking — strategic thinking. Strategic thinking involves thinking about the possible futures and setting plans to guide tactical actions intended to bring about the most desired future. Thus, the more elaborate PFC patches operated to perform executive functions that provided this facility.
FOC started out as a monitor, evaluator, and longer term adjustor for the operational control system of the action selection map. It guided the learning of appropriate tactical behaviors to benefit the organism in a dynamic, complex environment. It evolved into a system that maintains awareness of the environment through attention focus and a kind of decision processing, not unlike the action selection mapping process, that guided learning. This was the evolutionarily nascent prefrontal cortex. As it evolved in higher mammals and especially primates, it expanded its temporal and spatial scope with expanded memory systems to keep track of the immediate future and the immediate past. There was nothing like language, but if there had been it might have sentences like this: "I just ate. I'm tired. I'll take a nap." Upon waking: "I just slept. I'm hungry again. I'm going to get something to eat." Not much depth in space or time.
Then something marvelous happened. One little patch of the PFC, the one directly behind the eyebrows underwent a further development; in essence it provided a wholly new kind of map. It provided the ultimate in integration from all other functions of the brain. And it stimulated the growth and development of the brain as a whole, but particularly the neocortex. That patch is the frontopolar or Brodmann area 10 (BA10), which I have written about before. In that post I mentioned the diagrams which I've decided to post here.
Since I have described the relation between BA10 evolution and sapience I won't go into that here. Rather I would like to focus on BA10's role in second order consciousness which is the subject of this posting.
I have to admit I have been tempted to invoke the notion of third order consciousness to distinguish between early hominid mentation and that of sapiens'. But upon reflection I decided that that would be overkill. Perhaps the development of sapience and SOC should be called second and a half consciousness (S½OC!) Strategic thinking is just one component of sapience and by itself isn't what full human consciousness is about. Nevertheless I have some reasons for thinking that the vast majority of modern people have very weak strategic thinking capacity, perhaps extending over only a small number of people in their social circle and a short period of time (like one week at best). It seems rare to encounter truly long-term, wide boundary strategic thinking in the population. Nevertheless what even a little bit of strategic capacity initiates is the need to communicate much more complex ideas between individuals in a social group. Lions, wolves, dolphins, and especially chimpanzees (social animals) can communicate tactical commands when conducting group hunts. The extent of their strategic thinking is laying plans to conduct the hunt itself. Humans, on the other hand, plan days or weeks in advance. Those plans can include more complex information about where the game will be at various times during a season. Such planning requires careful preparation which means the need for more elaborate communications.
Hence the advent of language. The FOC system developed a precursor to language when it developed the ability to construct sequences of actions in the action selection and motor output regions of the brain. Such a sequence might be thought of as a list of things to do next: Contract muscle A, then contract muscle B while relaxing A, etc. At a little more abstract level this could give rise to the initiation point in the map coming to represent the whole sequence and being given a representation as a 'move', stored in a separate memory area. Then a sequence: Move arm down; pick up object; move arm up... etc. would start to look like what we would call a narrative. Note that going from the details of each muscle contraction to the abstract sequence of moves requires the exact same sequential processing but with the added sequence control over the map itself. That means another, higher level, map! There are a lot of details I am glossing over here, but I am asking you to accept that the sequencing of abstractions is no different in principle from the sequencing of muscle motions. But it does require an additional form of memory related to working memory that allows the temporary work space for producing a construction of this type and holding the representations of the abstractions. The working memory is used to activate each representation in the sequence and keep smaller sequences (phrases) active for future back references.
With abstractions — names of things and actions — and with motor control of tongue and glottis, we come to that new capability of language. We can construct narratives of what we are doing in an abstract form — words and sentences — and convey these to our fellow human beings. The PFC gave us the machinery to represent and manipulate abstractions in narratives. These are available not only to others through speech, but to something else inside our heads; the 2½ mechanism that arose from the workings of BA10. Just as the PFC monitored the longer-term performance of the tactical control system, FOC, and provided the beginnings of strategic thinking, so the BA10 patch stimulated the development in a number of other PFC patches to monitor the narrative production of the FOC system. We literally started to listen to our own sentences. And that is when it became necessary to exercise quality control over those sentences in the sense of thinking ahead and reflecting on our own pasts to make adjustments in our thinking. In other words we had to learn from our mistakes and successes by monitoring our own actions (as narratives), including our speech acts, and eventually our own thoughts (in the FOC sense). Second order consciousness now came into full being with the advent of an expanded BA10 patch as the ultimate convergence point for all activities in all other systems and the capacity to integrate those activities into an individually consistent on-going narrative over one's lifetime. Our sociality required integration of our moral sentiments along with our desires and other limbic responses. The system provided an ultimate action selection map in the form of directing thoughts and inhibiting improper actions in the other systems that were merely able to respond to the immediate situation.
We humans can ask ourselves questions like: Why did I do that?, Why did I say that? Or even more subtle kinds of questions like: Who am I?, Why am I here? Where do I fit into the scheme of things?
"I" as the actor is not really the real-time actor at all. It is my BA10 (and its integration with the rest of the PFC) in an on-going monitoring of my organism's decisions and action results over long time periods and using the information to make adjustments. I am my strategic thinking acting on my tactical thinking acting on my operational controls so that my organism might succeed and thrive over the course of my lifetime. Moreover, my success can translate into the success of my children and their children if what I have learned is worthy.
Our sapient brains have the ability to monitor our own actions and thoughts, and even our own plans, aspirations, and beliefs. We can do this because there is a higher-level action selection function that monitors our long-term being and is set to generate such (existential) questions as a prod back to FOC. The latter then acts on our behavior — for example we may be driven to seek answers or to learn more about the wider world. We humans are the product of evolution continuing to elaborate a simple theme: map situations to actions and then monitor the results for goodness or badness, first in real-time, then in short-period time (tactical) where the situation is the product of the real-time system, and then in longer-period time (strategic) where the situation is the long-term performance of the tactical system. The latter come in the form of narratives or sequences of abstract representations.
Many such narratives taken together form stories. And many stories accumulate over ones lifetime. More importantly all stories share some common attributes such as plots and themes. These latter considerations are yet higher level abstractions. And it is these abstractions that are tucked away in tacit knowledge, ready to be called upon to use for guidance in making decisions in real-life stories that unfold before us. We come at last to wisdom. All of the abstract story lines capture all of the essential elements of how to live, what happens when such-and-such takes place, who (what character type like protagonist) should do what and when. The wise (sapient) person has aggregated many of life's stories in their abstract form and has access to their internal logic as part of that tacit knowledge that builds up over a lifetime of experiencing the specific stories embroiling specific real people.
The wise person is forever evaluating their own thoughts, indeed their whole mind. Often this is done subconsciously (the FOC system can do what it has to do according to what is happening in the immediate environment but the SOC is free to wander). But just as often the sapient being is observing their own thoughts consciously and attempting to evaluate the goodness or badness as part of a life-long attempt to produce moral thinking for the biggest story of all.
Thus, I think sapience and second order consciousness go hand-in-hand. They are just different aspects of the same mental phenomenon. I would expect, too, that the level of sapience has a great deal to do with the 'level' of consciousness. Wiser people are more conscious of both the world and themselves. And people who have the mental competence to be more conscious of themselves and the world as they mature become wiser with age.
One day I shall have to write this all up in greater detail.
[EDIT: 6/29/2010. I thought I would add this diagram as sort of another view of the brain's architecture with respect to operational, tactical (and logistical), and strategic control. The "eye balls" (kind of hoaky I know) represent the monitoring function of the next layer in the hierarchy. Think of each higher layer as an "observer" of the one below. The strategic layer is the last one on the right. This is primarily the job of the prefrontal cortex with the BA10 patch acting to "organize" and integrate everything. The black thick arrows providing recurrent feedback to prior levels (motor program, sequences, and narrative) are the higher level abstractions mentioned above. The narrative is the language stream of what is happening and other "thoughts" that loops back through the strategic controller's observation of the tactical level to produce the stream of consciousness (phenomenological experience). The aggregate of narratives form abstract representations we call stories as part of tacit knowledge (strategic models).
George ]