**3. A brief history of understanding**

Notions addressed in this part were developed elsewhere [10–13, 27–30] and will be summarized briefly here. A preview will help putting the notions together: Environment is in flux, survival depends on an organism's ability to adapt to the changing environment. Adaptation makes the world livable while understanding makes it intelligible, that is, amenable to prediction and explanation (i.e., connecting likely future events to their plausible causes in the past and present). Mechanisms of understanding complete the transformation of sensory streams into world models that generate such predictions and explanations. The transformation starts with mechanisms of sensation and perception that are available, in different forms, in other species, and culminates in the mechanism of understanding unique

**79**

**Figure 3.**

*(anticipatory planning).*

*Brain Functional Architecture and Human Understanding*

period of roughly 150,000–60,000 years ago [35–37].

to humans. Learning response-reward (response-punishment) patterns increases reward chances and decreases punishment risks when conditions recur. A repertoire of such learned patterns constitutes a model of the environment instantiated by pattern matching. Understanding is an advanced adaptive mechanism serving to overcome the inertia of prior learning and optimize responses when conditions are novel or violate the previously acquired conditions-response associations in a consequential manner (e.g., learned responses cease to be rewarding) [10–13]. This characterization is consistent with definitions of intelligence in the literature ("fluid intelligence" [18, 31–33]) establishing understanding capacity as the central,

Complex life forms have been developing on Earth at an accelerating pace: From the emergence of unicellular organisms some 3.7 billion years ago, to (the emergence of) multicellular animals 900 million years ago, to vertebrate 530 million years ago, to primates about 70 million years ago, to the detachment of the human branch from the chimpanzee/bonobos primate branch 6 million years ago to, finally, the emergence of anatomically modern Sapiens [34] at the time period of 200,000–100,000 years ago (the emergence of language is attributed to the time

Recent findings indicated genealogical continuity in Sapience in the last 28,000 years, i.e. from Upper Paleolithic to modern times [38]. During the same period, the size of the braincase has been decreasing, having lost more than 10% of its peak value [39]), after a preceding period of about 6 million years during which the size almost tripled [40]. Recent analysis comparing the results of electrophysiological, anatomical and fMRI studies in humans and non-human primates associated development of intelligence primarily with reorganization of brain mechanisms [41]. These findings seem to indicate that reorganizations entailed higher efficiency so that progressively more complex tasks could be carried out

*Gap X denotes discontinuity in the development of cognitive capacities. Simple organisms interact with substances located on their 'blankets', more complex organisms can move towards and reach for target objects (denoted by black circles) in close proximity to their blankets (e.g. salamanders shoot their tongues to catch insects), and advanced animals (apes, some avians) can use a few supplementary objects (denoted by shaded and white circles) to act on the target (e.g. chimpanzees can connect sticks and pile up boxes in order to reach a hanging fruit). Humans are discontinuous with the other species in that they can form coordinated structures (designs) comprising indefinitely large sets of supplementary objects giving access to indefinitely distant targets, with the possibility of postponing acting on such targets until some indefinitely remote future moments* 

*DOI: http://dx.doi.org/10.5772/intechopen.95594*

defining feature of human intellect.

**3.1 Evolutionary precursors**

*Brain Functional Architecture and Human Understanding DOI: http://dx.doi.org/10.5772/intechopen.95594*

to humans. Learning response-reward (response-punishment) patterns increases reward chances and decreases punishment risks when conditions recur. A repertoire of such learned patterns constitutes a model of the environment instantiated by pattern matching. Understanding is an advanced adaptive mechanism serving to overcome the inertia of prior learning and optimize responses when conditions are novel or violate the previously acquired conditions-response associations in a consequential manner (e.g., learned responses cease to be rewarding) [10–13]. This characterization is consistent with definitions of intelligence in the literature ("fluid intelligence" [18, 31–33]) establishing understanding capacity as the central, defining feature of human intellect.

## **3.1 Evolutionary precursors**

*Connectivity and Functional Specialization in the Brain*

novices waste effort in considering illegal moves [26].

define the main challenges facing a theory of understanding:

negligible duration (on the evolutionary time scale).

The next part focuses on the emergence of understanding.

the understanding capacity,

**3. A brief history of understanding**

b.how such mechanisms could emerge and

in chess have been always appreciated, there were no satisfactory methods for quantifying them until the era of chess computers. Chess algorithms required hardware with operating speed at or above <sup>8</sup> 10 position evaluations per second in order to compete with expert players capable of carrying out at most one or two position evaluations per second. Understanding the game compensates for the 1: <sup>8</sup> 10 disadvantage in speed: expert players perceive configurations of pieces as compositions of "complexes", deriving game plans from apprehending coordinations between the "complexes" [25]. Findings in [25] suggest that expert game models take the form of simultaneous structures, not unlike the matrix in **Figure 1**. A novice's perception is limited to a few adjacent cells in the matrix (2–3 moves look-ahead involving 2–3 pieces) while expert models can include a hierarchy of matrices encompassing the entire configuration and extending to 10–15 moves look-ahead Position analysis involves envisioning variations for some of the moves, constrained by the entire web of coordinations across the matrix. As a result, experts are not distracted into considering spurious (weak) moves, no more than

To summarize, the previous section associated understanding with the development of mental models representing entities, their behavior and different forms of behavior coordination in the form of simultaneous memory structures. It was suggested that simultaneous coordination suppresses combinatorial explosion, confining the process to an infinitesimally small volume in the vast combinatorial space (considering possible move combinations in chess, similar to considering possible letter combinations in playwriting, quickly brings one to the realm of counting protons in multiple universes). Prediction, explanation and planning are enabled by mental modeling. This section reviewed extreme cases when modeling processes failed to establish coordination between a few directly observable and persistent entities and succeeded in quickly coordinating multiple, transient and/or unobservable ones. Summarily, suggestions and observations in Section 2

a.what neuronal mechanisms can account for the successes and shortcomings of

c.how could they develop in the human species within the time period of

Notions addressed in this part were developed elsewhere [10–13, 27–30] and will be summarized briefly here. A preview will help putting the notions together: Environment is in flux, survival depends on an organism's ability to adapt to the changing environment. Adaptation makes the world livable while understanding makes it intelligible, that is, amenable to prediction and explanation (i.e., connecting likely future events to their plausible causes in the past and present). Mechanisms of understanding complete the transformation of sensory streams into world models that generate such predictions and explanations. The transformation starts with mechanisms of sensation and perception that are available, in different forms, in other species, and culminates in the mechanism of understanding unique

**78**

Complex life forms have been developing on Earth at an accelerating pace: From the emergence of unicellular organisms some 3.7 billion years ago, to (the emergence of) multicellular animals 900 million years ago, to vertebrate 530 million years ago, to primates about 70 million years ago, to the detachment of the human branch from the chimpanzee/bonobos primate branch 6 million years ago to, finally, the emergence of anatomically modern Sapiens [34] at the time period of 200,000–100,000 years ago (the emergence of language is attributed to the time period of roughly 150,000–60,000 years ago [35–37].

Recent findings indicated genealogical continuity in Sapience in the last 28,000 years, i.e. from Upper Paleolithic to modern times [38]. During the same period, the size of the braincase has been decreasing, having lost more than 10% of its peak value [39]), after a preceding period of about 6 million years during which the size almost tripled [40]. Recent analysis comparing the results of electrophysiological, anatomical and fMRI studies in humans and non-human primates associated development of intelligence primarily with reorganization of brain mechanisms [41]. These findings seem to indicate that reorganizations entailed higher efficiency so that progressively more complex tasks could be carried out

#### **Figure 3.**

*Gap X denotes discontinuity in the development of cognitive capacities. Simple organisms interact with substances located on their 'blankets', more complex organisms can move towards and reach for target objects (denoted by black circles) in close proximity to their blankets (e.g. salamanders shoot their tongues to catch insects), and advanced animals (apes, some avians) can use a few supplementary objects (denoted by shaded and white circles) to act on the target (e.g. chimpanzees can connect sticks and pile up boxes in order to reach a hanging fruit). Humans are discontinuous with the other species in that they can form coordinated structures (designs) comprising indefinitely large sets of supplementary objects giving access to indefinitely distant targets, with the possibility of postponing acting on such targets until some indefinitely remote future moments (anticipatory planning).*

without increasing the size of the neuronal pool. Section 4 will suggest the type of reorganization that could produce such revolutionary improvements.

Comparing modes of interaction between the organism and environment across the spectrum of life forms reveals discontinuities between Sapiens and other species, as shown in **Figure 3**. The term 'Markov blanket' [13, 14] denotes an enclosing boundary (e.g., membrane) separating organism from the environment (the notion will be defined more precisely in the next section).

Differences between Sapience and other species are qualitative: they lie not in the increased quantity of supplementary objects but in the drive to keep extending the reach of action (action envelope) and to form progressively more complex designs comprising growing numbers of objects of increasing variety. Stated differently, animal envelopes are limited to the immediate proximity of their Markov blankets while human envelopes undergo indefinite expansion. Amplifying Shakespeare's insight (expressed succinctly in the epigraph), it can be suggested that animals seek biological equilibrium with their environment (i.e., maintaining inflows of energy and nutrients at life-sustaining levels) while humans seek cognitive equilibrium entailing demands not reducible to those associated with sustaining life. Hence, gap X. What is the nature of that gap?

#### *3.1.1 Learning and pattern recognition*

Consider challenges facing organisms in a changing environment. Assume first that the varying flow of conditions (stimuli) includes some recurring patterns. Since finding successful responses consumes time and effort, recognizing such patterns and re-using the responses saves both. The strategy works best when patterns comprise a few contiguous stimuli that trigger a small repertoire of fixed responses. However, even this simple strategy working under favorable circumstances can become self-defeating when the circumstance change, as illustrated in the following example.

Salamanders shoot their tongues at objects (insects) whose size, speed and distance from the animal fall within some fixed ranges, which requires anticipatory response control (early activation of the projector muscle relative to the tongue launch) to improve the chances of successful intercepts. The shooting mechanism was fine-tuned by evolution (developing spring-loaded type of tongue ejection yielding high energy output), making the animal a successful predator [42]. Consider a hypothetical scenario when the advantages are turned into detriments. The shooting mechanism is thermally sensitive: the speed of tongue retraction increases with temperature [42] which can be used, potentially, to increase the amount of prey intake per unit time. Assume that the animal can learn the 'higher temperature – higher intake' association, compelling it to seek high temperature spots. Such learning will keep paying off for as long as the prey cooperates: if the insects start moving faster in the vicinity of hot spots (or avoid them, etc.), the intercept success rate will decline. However, the animal will be bound to continue the heat-seeking behavior until the association decays, which might cause it to die from hunger and/or exhaustion (missing targets decreases food intake but not the costs). The point is that the ability to suppress learned behavior can yield quantum leap improvements in adaptive robustness, by reducing the probability of 'blind persistence' types of error inherent in recognition-centered strategy, and/or reducing the severity of the consequences. In general, the strategy works if short contiguous patterns (compact patterns) recur with frequency sufficient for satisfying the organism's survival needs. Assume that the requirement is not met, forcing the animal to seek strategies applicable in more complex stimuli configurations**.**

**81**

**Figure 4.**

*Brain Functional Architecture and Human Understanding*

successions of episodes, as shown in **Figure 5**.

Removing (or relaxing) the contiguity requirement changes an animal's view of the environment: form a noisy stream of compact patterns to a stream of uncertain structure where patterns can no longer be readily discerned. Stated differently, in streams of non-contiguous patterns (dispersed patterns) stimuli groupings in one pattern can be interspersed irregularly with groupings belonging to other patterns, thus allowing extending patterns over indefinitely long stimuli sequences and time periods. Dispersed patterns place organisms at the horns of a dilemma, as shown

Pattern composition in **Figure 4(2)** is inherently uncertain, gradual reduction of the uncertainty proceeds reversibly through the stages of a) defining entities (as compositions of states), b) defining behaviors (as patterns of state transition) c) defining relations (as forms of behavior coordination), resulting in the construction of simultaneous structures representing interactions between entities in

Strategies in **Figure 4(1)** and **(2)** reside at the opposite sides of gap X: cognitive operations underlying the former are exogenously driven, i.e., triggered by the environment and carried out under feedback control, while operations underlying the latter are endogenously-driven, i.e. decoupled from the sensory inflows. Rudimentary forms of such decoupling manifest in animal behavior, e.g., dogs following a prey that disappears behind an obstacle might not chase it around the corner but run to intercept at the opposite corner. On the human side of the gap, reversible operations become available gradually as the person matures, causing characteristic errors (e.g., young children fail in the "toy has moved" task requiring that association (toy, cover1, spot1) is followed by dissociation (toy, cover1, spot1) ➔ (cover 1, spot 1) ➔ (toy, cover2, spot2), see section I.2.a. A different form of dissociation deficit manifests in older children when they fail to dissociate container from the contents: a child watching liquid being poured from one container to another can believe that the amount changes with the size of the

Note that entity construction principles in **Figure 4(2)** and **5** express an implicit assumption that entity's identity can be preserved in different manifestations in non-contiguous episodes, that is, the same entity can have different (non-overlapping) manifestations and, vice versa, different entities can have identical manifestations (e.g., in Greek mythology, enterprising Zeus was appearing to mortal women in the form of a swan, a bull, or even a shower. On one occasion, Zeus presented himself to a lady in a form that was identical to her husband (Amphitryon) in every

*Transition from compact to dispersed patterns inside gap X. 1) contiguous stimuli grouping ABC recurs at irregular noisy intervals, response strategy consists in finding activities rewarded by ABC and emitting them whenever the pattern is recognized. 2) removing the contiguity requirement changes the strategy from pattern recognition to pattern construction. Here is the dilemma: stimuli A, B, C can be manifestations of either different entities requiring different responses or different states of the same entity requiring the same response (possibly, with modifications). Whatever the resolution, it might change at some later point in time, e.g. XYB and XYC can be determined to be the states of some entity Z, causing a to recede into the background noise, etc.*

*DOI: http://dx.doi.org/10.5772/intechopen.95594*

*3.1.2 Gap X*

below.

container [43].

*Brain Functional Architecture and Human Understanding DOI: http://dx.doi.org/10.5772/intechopen.95594*

#### *3.1.2 Gap X*

*Connectivity and Functional Specialization in the Brain*

will be defined more precisely in the next section).

life. Hence, gap X. What is the nature of that gap?

*3.1.1 Learning and pattern recognition*

without increasing the size of the neuronal pool. Section 4 will suggest the type of

Comparing modes of interaction between the organism and environment across the spectrum of life forms reveals discontinuities between Sapiens and other species, as shown in **Figure 3**. The term 'Markov blanket' [13, 14] denotes an enclosing boundary (e.g., membrane) separating organism from the environment (the notion

Differences between Sapience and other species are qualitative: they lie not in the increased quantity of supplementary objects but in the drive to keep extending the reach of action (action envelope) and to form progressively more complex designs comprising growing numbers of objects of increasing variety. Stated differently, animal envelopes are limited to the immediate proximity of their Markov blankets while human envelopes undergo indefinite expansion. Amplifying Shakespeare's insight (expressed succinctly in the epigraph), it can be suggested that animals seek biological equilibrium with their environment (i.e., maintaining inflows of energy and nutrients at life-sustaining levels) while humans seek cognitive equilibrium entailing demands not reducible to those associated with sustaining

Consider challenges facing organisms in a changing environment. Assume first that the varying flow of conditions (stimuli) includes some recurring patterns. Since finding successful responses consumes time and effort, recognizing such patterns and re-using the responses saves both. The strategy works best when patterns comprise a few contiguous stimuli that trigger a small repertoire of fixed responses. However, even this simple strategy working under favorable circumstances can become self-defeating when the circumstance change, as illustrated in the following

Salamanders shoot their tongues at objects (insects) whose size, speed and distance from the animal fall within some fixed ranges, which requires anticipatory response control (early activation of the projector muscle relative to the tongue launch) to improve the chances of successful intercepts. The shooting mechanism was fine-tuned by evolution (developing spring-loaded type of tongue ejection yielding high energy output), making the animal a successful predator [42]. Consider a hypothetical scenario when the advantages are turned into detriments. The shooting mechanism is thermally sensitive: the speed of tongue retraction increases with temperature [42] which can be used, potentially, to increase the amount of prey intake per unit time. Assume that the animal can learn the 'higher temperature – higher intake' association, compelling it to seek high temperature spots. Such learning will keep paying off for as long as the prey cooperates: if the insects start moving faster in the vicinity of hot spots (or avoid them, etc.), the intercept success rate will decline. However, the animal will be bound to continue the heat-seeking behavior until the association decays, which might cause it to die from hunger and/or exhaustion (missing targets decreases food intake but not the costs). The point is that the ability to suppress learned behavior can yield quantum leap improvements in adaptive robustness, by reducing the probability of 'blind persistence' types of error inherent in recognition-centered strategy, and/or reducing the severity of the consequences. In general, the strategy works if short contiguous patterns (compact patterns) recur with frequency sufficient for satisfying the organism's survival needs. Assume that the requirement is not met, forcing the animal to seek strategies applicable in more complex stimuli configurations**.**

reorganization that could produce such revolutionary improvements.

**80**

example.

Removing (or relaxing) the contiguity requirement changes an animal's view of the environment: form a noisy stream of compact patterns to a stream of uncertain structure where patterns can no longer be readily discerned. Stated differently, in streams of non-contiguous patterns (dispersed patterns) stimuli groupings in one pattern can be interspersed irregularly with groupings belonging to other patterns, thus allowing extending patterns over indefinitely long stimuli sequences and time periods. Dispersed patterns place organisms at the horns of a dilemma, as shown below.

Pattern composition in **Figure 4(2)** is inherently uncertain, gradual reduction of the uncertainty proceeds reversibly through the stages of a) defining entities (as compositions of states), b) defining behaviors (as patterns of state transition) c) defining relations (as forms of behavior coordination), resulting in the construction of simultaneous structures representing interactions between entities in successions of episodes, as shown in **Figure 5**.

Strategies in **Figure 4(1)** and **(2)** reside at the opposite sides of gap X: cognitive operations underlying the former are exogenously driven, i.e., triggered by the environment and carried out under feedback control, while operations underlying the latter are endogenously-driven, i.e. decoupled from the sensory inflows. Rudimentary forms of such decoupling manifest in animal behavior, e.g., dogs following a prey that disappears behind an obstacle might not chase it around the corner but run to intercept at the opposite corner. On the human side of the gap, reversible operations become available gradually as the person matures, causing characteristic errors (e.g., young children fail in the "toy has moved" task requiring that association (toy, cover1, spot1) is followed by dissociation (toy, cover1, spot1) ➔ (cover 1, spot 1) ➔ (toy, cover2, spot2), see section I.2.a. A different form of dissociation deficit manifests in older children when they fail to dissociate container from the contents: a child watching liquid being poured from one container to another can believe that the amount changes with the size of the container [43].

Note that entity construction principles in **Figure 4(2)** and **5** express an implicit assumption that entity's identity can be preserved in different manifestations in non-contiguous episodes, that is, the same entity can have different (non-overlapping) manifestations and, vice versa, different entities can have identical manifestations (e.g., in Greek mythology, enterprising Zeus was appearing to mortal women in the form of a swan, a bull, or even a shower. On one occasion, Zeus presented himself to a lady in a form that was identical to her husband (Amphitryon) in every

#### **Figure 4.**

*Transition from compact to dispersed patterns inside gap X. 1) contiguous stimuli grouping ABC recurs at irregular noisy intervals, response strategy consists in finding activities rewarded by ABC and emitting them whenever the pattern is recognized. 2) removing the contiguity requirement changes the strategy from pattern recognition to pattern construction. Here is the dilemma: stimuli A, B, C can be manifestations of either different entities requiring different responses or different states of the same entity requiring the same response (possibly, with modifications). Whatever the resolution, it might change at some later point in time, e.g. XYB and XYC can be determined to be the states of some entity Z, causing a to recede into the background noise, etc.*

#### **Figure 5.**

*An irreversible stimuli stream is transformed into a simultaneous record, cycles of reversible operations on the record (select/deselect, etc.) produce simultaneous structures comprising various entities interacting in series of episodes (see Figure 1).*

detail but was not her husband – Amphitryon was quite sure of that). Implicit explorations of logic in Greek mythology were made explicit by Aristotle in the Laws of Thought, including the Law of Identity.

#### **3.2 Crossing gap X**

According to an appealing hypothesis [44], the earliest steps in the expansion of the human envelope were associated with predation by throwing projectiles (stones). Accurate aiming requires precise coordination of several variables including launch angle, velocity, weight and size of the stone, distance to the prey and its size, and release time, with the width of the release time window limited to a few milliseconds (e.g., 11 milliseconds for a rabbit-size stationary target located 4 meters away, these results will be re-visited in the next section). Analysis based on experimental findings (narrowing the time window involves synchronization in neuronal clusters of growing size) demonstrated that increasing distance to targets while maintaining the hit rate requires explosive growth in the number of neurons responsible for precise timing (64-fold and 729-fold increase in the number of neurons to double and triple the distance, correspondingly).

Anatomical limitations imposed on the volume of cranial cavity appeared to exclude the possibility that a growing variety of high-precision activities (e.g. splitting stones for different tasks) could be obtained by developing narrowly specialized neuronal modules. Anatomical limitations enforce other trade-offs having impact on cognitive performance, e.g., increasing the speed of pulse conduction would require increasing the thickness of myelin wrappings, which would decrease the number of neurons the cranial geometry can accommodate [40, 44] .

In addition to constraints in brain size and conduction speed, another physical factor having decisive impact on brain processes is limited supply of energy for powering them. Since physical constraints on brain processes are non-negotiable, the only avenue for obtaining quantum advancements in cognitive performance depicted in **Figure 2** appears to be dynamic optimization in their deployment, which boils down to global coordination via the mechanisms of mental modeling.

**83**

**Figure 6.**

*Symbol* ⋈ *denotes coordination.*

*Brain Functional Architecture and Human Understanding*

These notions will be addressed in the theory of understanding in the next part, following another example of mental modeling in the closing of this part.

involved in constructing and operating catapults are suggested in **Figure 6**.

For the sake of argument, assume that advancing the predation-by-throwingprojectiles strategy involved invention of catapults, in the simplest form of a board (B1) balanced on a base, or fulcrum (B2). Note that neither component, if considered individually, betrays any hint as to its potential usefulness for projectile throwing. Moreover, when considered jointly, these components afford numerous arrangements that are all useless (e.g., the base on top of the board, etc.), with only one particular form of base-board position coordination yielding the benefit. Operations

Note that the product of modeling is a new entity (a weapon) that has properties unavailable in the components and expands the activity envelope (larger distances, heavier projectiles). Running the model yields understanding, i.e., informs operation and aiming procedures. For example, envisioning one side of the board going up brings to mind the image of the other side going down, envisioning increasing the distance to the target brings to mind the image of increasing the length of the shoulder (shifting the projectile away from the base), etc. That same process underlies prediction (e.g. hit probability) and explanation (why hitting that target

It is interesting to note that children up to a certain age, when learning to operate

toy catapults, are often incapable of forming proper models and keep shifting projectiles in the direction of the target as it moves away (shortening the shoulder), even after having watched the proper operation multiple times [43]. The instinctual tendency to grasp receding objects by extending arms and moving after the objects

Recent theories concerning the origins of language placed the capacity to perform reversible juxtaposition (operation Merge) at the foundation on which all other language mechanisms have been built (B2 B1) ➔ C (operation Merge com-

nism that makes possible constructing responses to indefinitely large patterns comprised of non-contiguous stimuli groupings (dispersed patterns). Construction proceeds through identifying entities, their properties and behavior and the forms of inter-entity behavior coordination, culminating in the production of simultaneous, tightly coordinated structures comprising multiple entities (mental models). Models are amenable to manipulations, giving rise to the dual capacity for predicting likely events (changes in the entities) and identifying their causes in the past or present (explanations). In general, any organism can be viewed as a cast molded by the environmental niche it occupies, e.g., salamander is a 'cast mold' of environment where particular (edible) insects having size and speed within some fixed ranges are flowing into a volume in space reachable by the animal in unit time in quantities sufficient for the animal's survival. The total model includes biophysical component

*Modeling starts with selecting entities (objects) and juxtaposing them as separate (independent) entities, followed by associating them in a composite structure allowing inter-dependence, followed by coordinating the entities to form a model (note that juxtaposition brings components together in an arbitrary order while association imposes order, setting the stage for establishing a higher degree of order in the model).* 

To summarize, this part defined understanding as an advanced adaptive mecha-

resists learning. Young children cannot understand catapults.

bines syntactic objects in an arbitrary order [45]).

*DOI: http://dx.doi.org/10.5772/intechopen.95594*

over there is unlikely?).

#### *Brain Functional Architecture and Human Understanding DOI: http://dx.doi.org/10.5772/intechopen.95594*

*Connectivity and Functional Specialization in the Brain*

Laws of Thought, including the Law of Identity.

**3.2 Crossing gap X**

*episodes (see Figure 1).*

**Figure 5.**

detail but was not her husband – Amphitryon was quite sure of that). Implicit explorations of logic in Greek mythology were made explicit by Aristotle in the

*An irreversible stimuli stream is transformed into a simultaneous record, cycles of reversible operations on the record (select/deselect, etc.) produce simultaneous structures comprising various entities interacting in series of* 

number of neurons to double and triple the distance, correspondingly).

the number of neurons the cranial geometry can accommodate [40, 44] .

Anatomical limitations imposed on the volume of cranial cavity appeared to exclude the possibility that a growing variety of high-precision activities (e.g. splitting stones for different tasks) could be obtained by developing narrowly specialized neuronal modules. Anatomical limitations enforce other trade-offs having impact on cognitive performance, e.g., increasing the speed of pulse conduction would require increasing the thickness of myelin wrappings, which would decrease

In addition to constraints in brain size and conduction speed, another physical factor having decisive impact on brain processes is limited supply of energy for powering them. Since physical constraints on brain processes are non-negotiable, the only avenue for obtaining quantum advancements in cognitive performance depicted in **Figure 2** appears to be dynamic optimization in their deployment, which boils down to global coordination via the mechanisms of mental modeling.

According to an appealing hypothesis [44], the earliest steps in the expansion of the human envelope were associated with predation by throwing projectiles (stones). Accurate aiming requires precise coordination of several variables including launch angle, velocity, weight and size of the stone, distance to the prey and its size, and release time, with the width of the release time window limited to a few milliseconds (e.g., 11 milliseconds for a rabbit-size stationary target located 4 meters away, these results will be re-visited in the next section). Analysis based on experimental findings (narrowing the time window involves synchronization in neuronal clusters of growing size) demonstrated that increasing distance to targets while maintaining the hit rate requires explosive growth in the number of neurons responsible for precise timing (64-fold and 729-fold increase in the

**82**

These notions will be addressed in the theory of understanding in the next part, following another example of mental modeling in the closing of this part.

For the sake of argument, assume that advancing the predation-by-throwingprojectiles strategy involved invention of catapults, in the simplest form of a board (B1) balanced on a base, or fulcrum (B2). Note that neither component, if considered individually, betrays any hint as to its potential usefulness for projectile throwing. Moreover, when considered jointly, these components afford numerous arrangements that are all useless (e.g., the base on top of the board, etc.), with only one particular form of base-board position coordination yielding the benefit. Operations involved in constructing and operating catapults are suggested in **Figure 6**.

Note that the product of modeling is a new entity (a weapon) that has properties unavailable in the components and expands the activity envelope (larger distances, heavier projectiles). Running the model yields understanding, i.e., informs operation and aiming procedures. For example, envisioning one side of the board going up brings to mind the image of the other side going down, envisioning increasing the distance to the target brings to mind the image of increasing the length of the shoulder (shifting the projectile away from the base), etc. That same process underlies prediction (e.g. hit probability) and explanation (why hitting that target over there is unlikely?).

It is interesting to note that children up to a certain age, when learning to operate toy catapults, are often incapable of forming proper models and keep shifting projectiles in the direction of the target as it moves away (shortening the shoulder), even after having watched the proper operation multiple times [43]. The instinctual tendency to grasp receding objects by extending arms and moving after the objects resists learning. Young children cannot understand catapults.

Recent theories concerning the origins of language placed the capacity to perform reversible juxtaposition (operation Merge) at the foundation on which all other language mechanisms have been built (B2 B1) ➔ C (operation Merge combines syntactic objects in an arbitrary order [45]).

To summarize, this part defined understanding as an advanced adaptive mechanism that makes possible constructing responses to indefinitely large patterns comprised of non-contiguous stimuli groupings (dispersed patterns). Construction proceeds through identifying entities, their properties and behavior and the forms of inter-entity behavior coordination, culminating in the production of simultaneous, tightly coordinated structures comprising multiple entities (mental models). Models are amenable to manipulations, giving rise to the dual capacity for predicting likely events (changes in the entities) and identifying their causes in the past or present (explanations). In general, any organism can be viewed as a cast molded by the environmental niche it occupies, e.g., salamander is a 'cast mold' of environment where particular (edible) insects having size and speed within some fixed ranges are flowing into a volume in space reachable by the animal in unit time in quantities sufficient for the animal's survival. The total model includes biophysical component

#### **Figure 6.**

*Modeling starts with selecting entities (objects) and juxtaposing them as separate (independent) entities, followed by associating them in a composite structure allowing inter-dependence, followed by coordinating the entities to form a model (note that juxtaposition brings components together in an arbitrary order while association imposes order, setting the stage for establishing a higher degree of order in the model). Symbol* ⋈ *denotes coordination.*

(body and the sensory-motor periphery, e.g. the tongue-ejecting mechanism) and regulatory component orchestrating activities within the body and at the periphery (i.e., animal's behavior in the environment). Both components undergo evolutionary development in the species while behavior regulation is amenable to adaptive changes in individuals during their lifetime (learning). In animals, learning is restricted to condition-driven variations within narrow envelopes of geneticallyfixed condition-response patterns and propensities. Condition-driven learning extrapolates from past precedents while mental modeling enables prediction and response construction under conditions having no such precedents. More precisely, models integrate past history within cross-coordinated structures so predictions produced by operations on the structure can be made consistent with (plausible under the entire past history) without repeating any of its elements. Moreover, models allow reproductive construction without replication, e.g., coordinations in the basic catapult were reproduced in numerous designs.

As observed by Jean Piaget [46].

*"…mental coordinations succeed in combining all the multifarious data and successive data into an overall, simultaneous picture, which vastly multiplies their powers of spatio-temporal extension, and of deducing possible developments" ([46] p. 218).*

Summarily, it has been suggested that a) the protohuman-to-human transition was associated with the emergent capacity to construct responses to dispersed stimuli patterns and b) the capacity is rooted in the mechanisms of mental modeling that represents such patterns as coordinated structures that suppress combinatorial explosion inherent in the construction process and reduce the number of response compositions to a few plausible alternatives.
