The great ornithologist Richard Attenborough was once asked why birds sing. After a long explanation, ending in a detailed examination of what happens during a morning chorus, he concluded, "So, if there are no predators, and it is not time to mate, why is there still a morning chorus? Well. It's too late to sleep. It's too early to catch worms. So there isn't anything else to do."

Leonardo, at MIT is one of the most Advanced Humanoid Intelligences
Leonardo, at MIT is one of the most Advanced Humanoid Intelligences


Between times when a baby is not eating and sleeping, there isn't much else to do either. If the baby is well fed, and comfortable, the mind is then available for the most important part of cognitive development: play.

What happens during play? Previous sections describe the basis of the model for this discussion:

  1. "Designing HAL (1): Hypercognition, Freud and Jung" integrated Demetriou's theory of hypercognition with theories of the unconscious from Freud and Jung.
  2. "Designing HAL (2): A Learning AI with Emotions and Dreams" outlined a framework for designing a complete computer model of cognition, including emotions and dreams.
  3. "Designing HAL (3): Developing Emotions and Self Awareness" described the basis for adding symbols to the cognitive network, explaining the indirect creation of self perception, and from that self awareness.

This section explains how the categorical domain expands and new associations formed through the activity of play. The basis of the play activity is the simple deduction:

Hide LIne Numbers
  1. If I observe A1 and A2, then usually B happens
  2. I can do A3, and A3 is similar to A1
  3. therefore if I do A1 while A2, then does B happen?
  5. If true, A3 is more similar to A1
  6. If false, either A3 is less similar to A1, 
  7.       or A1 and A2 is less likely to cause B:
  8.       If A3 and not A2, does B happen?
  9. etc...

For example:

Hide LIne Numbers
  1. when the light is on and mother is present, I usually get fed
  2. I found a button that turns another light on
  3. What if I turn the light on and mother is present, do I get fed?
  5. If true, light is more likely to cause being fed.
  6. If false, either turning the light on is less similar to the room light, 
  7.       or, light is less likely to result in me being fed 
  8.       What if I turn the light on and mother is not present?

The causal deductions themselves will be discussed in a future section on the logical processing domain. In order for the process to occur, the abstraction of A3, and its association with A1, first has to be formed in the categorical domain. This process is important in defining the perception of others, from the perception of self, so this section will discuss the nature and implications of creating new symbols in the categorical domain.

Creating Symbols in the Categorical Domain

Breazal et al (2004) describe a cognition system for a robot, capable of sophisticated interaction.

The Program architecture for the Leonardo Intelligence
The Program architecture for the Leonardo Intelligence

In this paper, it is postulated that the learning robot 'Leonardo' could understand others by imitation.

Learning by Imitation
Learning by Imitation

It was found, unsurprisingly, that the program would make some false assumptions as to the correspondence of its own motor control with the symbolic representation of the images it saw. For example, here, Leonardo considered the brows of the human face to correspond to its own ear.

Leonardo Attempting to Imitate Human Expression
Leonardo Attempting to Imitate Human Expression

In a 'blank slate' model of cognition, this kind of incorrect association must be learned, and then somehow replaced. This is the task of the categorical domain.

Creating Categorical Symbols

In the hypercognitive model, the act of introspection creates a new symbol with an association to previous symbols. The process for this was introduced under 'feelings for color' in "Designing HAL (1): Hypercognition, Freud and Jung." In this model, creating a new symbol not only a new software 'object,' but also creates an unconscious association with prior emotional states.

Types of Symbol Associations

The basis of this is very straightforward. Each symbol may be associated to another as in standard set theory, and may be either:

  • Parent: the symbol contains another.
  • Child: the symbol is contained by another.
  • Peer by Parent: two symbols with the same parent.
  • Peer by Child: two symbols with the same child.
  • Opposed by parent: two symbols may not have the same parent.
  • Opposed by child: two symbols may not have the same child.

In this model, however, the relationship does not need to be absolute. The degree by which any symbol has one of these associations is scalar. That is, the system is a modified form of the fuzzy logic of Salii (1965), whose system is a common root for linguistics (Cornelis et al., 2000), decision-making (Kuzmin, 1982) and clustering (Bezdek, 1978).

Rearranging Categorical Symbols

The symbols are rearranged in two ways:

  • Conscious Reorganization by Play- A complex process described in the future topic of causal domain processing, here mentioned by way of introduction.
  • Unconscious Reorganization by Dreams - A reorganization of accumulated symbolic data using the algorithm of simulating annealing by perturbation, described in "Designing HAL (2): A Learning AI with Emotions and Dreams."

Deleting Symbols

If one symbol can be created at intervals of the speed of a neural connection (2ms), then a neocortex containing 20 billion cells would not be exhausted of capacity for 12 years. Coincidentallly, this number is about the time that puberty commences, which will please Freudians.

Typically, a maximum of 10% of the neoctortex is active in humans, so system globals will limit the amount of processing to this level, as will be described in the control topic of this series.

It's not anticipated that a first-generation computer model based on this capacity, and emulating human cognition, will run long enough to create a problem with current limits on processing power and storage.

Developing Perception of Others

After perception of self, an individual is able to form a perception of others, by analogy. The following illustration shows how the original perception of self creates a basis for perceiving others, which through experience is linked with three fundamental symbolic complexes representing others:

  • The Peer Complex, which is directly derived from the perception of self.
  • The Maternal Complex, which is an authority figure derived from the perception of self, with an emphasis on the social 'fawn' subcomplex.
  • The Paternal Complex, which is an authority figure derived from the perception of self, with an emphasis on the social 'freeze' subcomplex.
The Creation of Symbols Representing Others
The Creation of Symbols Representing Others

After the formation of symbols representing others, and their association with others, external stimuli combine with new learning about the self to deepen and widen the understanding. Input from others then becomes a primary factory in cognitive development. Incongruity in the input from others with our own senses and memory causes a cognitive paradox, and the mechanism by which individuals respond to cognitive paradoxes is a primary factor in development.

Perception of Authority

In a healthy human, all the initial complexes are formed with strong 'like' associations to the secondary, social emotions. There is no 'enemy' complex associated with 'dislike' and the primary 'fear/anger,' or 'fight/flight' emotions. Individuals with strong 'like' characteristics are more acceptable as role models during imitation and play, and the deductions from these complexes is evaluated, during decision making, with importance even exceeding direct sensory input. This enables explanation of a large number of apparent paradoxes in adult human behavior, described below.

Authority Overrides Senses and Reason

The later needs for placing conflicting peer opinions over immediate sensory evidence, causal inference, and memory applies to an even greater extent on how adults will follow authority, and let being an authority position themselves override their normal moral beliefs.

Milgram (1961) proved that people will typically obey a leadership and even torture a person to apparent death simply by being told that they don't need to be concerned about it and should do what they are told. Moreover, they will do so even if the authority figure telling them to torture someone else is not present. Even if the person they are told to hurt is saying they have a bad heart condition and may die, they will choose to obey the authority rather than act compassionately.

The 'Stanford Prison Experiment' (1971), proved, if guards were left to their own devices in controlling prisoners, without supervision, the guards would use increasingly violent force, up to and including torture, to maintain authority.

Perception of Peers

One of the most astounding innovations in social psychology was a way to assess and define conformity pressure in decision making. This led to the formation of the 'cognitive dissonance' theory, which is central to this model because, not only does it permit integration of Jungian theories of unconscious with contemporary observations, but also it supports the strange and bizarre nature of how we respond to peer and authority judgments.

Asch (1951) was the first to suggest that people will act irrationally rather than disagree with a group. He established this to be true in a neutral situation, simply by showing three lines of obviously different lengths to a group of people, and asking them to say which is the same length as another line. One member of the group was an unknowing subject under test, and all others were secret participants in the experiment. If all the other people except the person under test answered the wrong line was the shortest, a third of the subjects agreed with the group rather than believe his own senses and say the rest of the group was wrong.

Subjects Answered Line 2 or 3 was the Same Length as the Line of the Left if All Others said so
Subjects Answered Line 2 or 3 was the Same Length as the Line of the Left if All Others said so

An interesting factor in the research was the extent to which the subjects under test would insist that they could see the wrong line as the same length. The Asch experiment has been replicated in many forms, and modern research has extended its applicability. Here are three examples.

Haidt (2000) posits "a collection of psychological mechanisms that dispose individuals selectively to credit or dismiss evidence of risk in patterns that fit values they share with others," rather than scientific fact. His experiments intend to demonstrate the truth of his hypothesis.

Prasad et al (2009) said "We form emotional attachments that get wrapped up in our personal identity and sense of morality, irrespective of the facts of the matter." This notion of ‘motivated reasoning’ is supported with experimental results in artificial settings.

Nisbet and Garrett (2010) presented subjects with a detailed rebuttal to the claim that "Feisal Abdul Rauf is a terrorist-sympathizer." Yet among those who were previously aware of the rumor and believed it, fewer than a third changed their minds.

The conventional psychological explanation of these observations is that the cognitive dissonance caused by disagreeing with a peer group is so unpleasant that individuals will alter their cognition of their perceptions rather than accept the sensory evidence as obviously true. Any model of human intelligence needs to take this important fact into account. This is why cognitive congruence/dissonance is considered essential in the executive function of this model.

Effect of Self Perception on Judgment

One issue in modeling this is how much the skill of the individual influences the judgement of self perception when it is in conflict with others.

Dunning and Kruger (1999) found that participants scoring in the bottom quartile on tests of humor, grammar, and logic grossly overestimated their test performance and ability, whereas individuals in the top quartile tend to underestimate it. Subjects failed to recognize their own lack of skill, failed to recognize whether peers or authority figures had genuine, and failed to recognize the extremity of their inadequacy.

Modeling this bias requires a modification of how people assess their self perception in relation to the number of previous times they have attempted to make a similar judgment.

Next Steps

"Designing HAL: Cognitive Architecture Overview" cites specific properties and methods to be implemented in AI software. The sixth part of this series (not published on this site) compares various software tools for writing the system, which includes neural-network modeling systems for definition of relative abstractions, and how well they can be integrated with compiled simulated annealing code for dreams and play (which is a non-trivial task).


  1. Aristotle (ca. 360BC)On the Soul
  2. Asch conformity experiments:
  3. Breazeal, C. "Sociable Machines" (ca. 2002)
  4. Breazeal, C.; Gray, J.; Berlin, M. "An Embodied Cognition Approach to Mindreading Skills for Socially Intelligent Robots"
  5. Breazeal, C.; Thomaz, A. (2005)Learning from Human Teachers with Socially Guided Exploration
  6. Cognitive Dissonance
  7. Cornelis, C.; De Cock, M.; Kerre, E. "Intuitionistic fuzzy rough sets: at the crossroads of imperfect knowledge," Expert Systems, v. 20, issue 5, pp. 260–270, 2003
  8. Demetriou, A. (2006)Neo-Plagetian Theories of Cognitive Development.
  9. Demetriou, A ; Spanoudis, G.; Mouyi, A. (2006) "A Three-level Model of the Developing Mind: Functional and Neuronal Substantiation"
  10. Dupre, J. (2002) "Making Hay with Straw Men" American Scientist
  11. Ekman, P. (1999). "Basic Emotions". In: T. Dalgleish and M. Power (Eds.). Handbook of Cognition and Emotion. John Wiley & Sons Ltd, Sussex, UK:.
  12. Festinger, L. (1957).A Theory of Cognitive Dissonance.California: Stanford University Press.
  13. Freud, S. (1933) New Introductory Lectures on Psychoanalysis, Penguin Freud Library
  14. Freud, S. (1940) An Outline of Psycho-analysis
  15. Haidt, J. (2000) "The Emotional Dog and its Rational Tail: A Social Intuitionist Approach to Moral Judgment," University of Virginia
  16. Henderson, D.; Jacobson, S.; Johnson, A. (2002) "The Theory and Practice of Simulated Annealing."
  17. Jung, C. G. (1935)Structure & Dynamics of the Psyche,The Collected Works of C. G. Jung, Volume 8. Princeton University Press
  18. Jung, C.G. (1971).Psychological Types, Collected Works, Volume 6, Princeton, N.J.: Princeton University Press
  19. Kripke, S. (1980)Naming and Necessity. Harvard University Press:
  20. Kruger and Dunning (2006) "Unskilled and Unaware of It: How Difficulties in Recognizing One's Own Incompetence Lead to Inflated Self-Assessments," Journal of Personality and Social Psychology, 77 (6): 1121–34. doi:10.1037/0022-3514.77.6.1121
  21. Kuzmin,V.B. (1982)Building Group Decisions in Spaces of Strict and Fuzzy Binary Relations, Nauka, Moscow, 1982 (in Russian)
  22. Locke, J. (ca. 1690)Essay Concerning Human Understanding, Volume 2
  23. Milgram experiment:
  24. Nisbet; E.; Garrett K. (2008) "Belief in rumors Hard to Dispel: Fact checking easily undermined by images, unrelated facts," Ohio State University
  25. Nisbett, R.E.; Wilson, T.D. (1977) "Telling more than we can know: Verbal reports on mental processes".
  26. Pinker, S. (2002)The Blank Slate: The Modern Denial of Human Nature. Continuum
  27. Plato (380 BC)The Republic, Allegory of the Cave, 514a–520a
  28. Prasad, M., Perrin, A; Bezila, K; Hoffman, S.; Kindleberge, K; Manturuk K; Powers, A. (2009) "There Must Be a Reason”: Osama, Saddam, and Inferred Justification," Sociological Inquiry, 29.2, pages 142–162
  29. Salii, V.N. (1965)Binary L-relations, Izv. Vysh. Uchebn. Zaved., Matematika, v. 44, No.1, pp. 133–145
  30. Stanford Prison Experiment:

Additional Associated References

  1. Baddeley, A; Sala, S. D. (1996) "Working Memory and Executive Control," Philosophical Transactions: Biological Sciences 351.1346, Executive and Cognitive Functions of the Prefrontal Cortex
  2. Chomsky, N. "Tool Module: Chomsky's Universal Grammar"
  3. Cowan,N. (2010) Multiple Concurrent Thoughts: The Meaning and Developmental Neuropsychology of Working Memory, Dev Neuropsychol.
  4. Cannon, W.B. (1915) "Bodily Changes in Pain, Hunger, Fear and Rage: An Account of Recent Researches into the Function of Emotional Excitement." Appleton.
  5. Davidson, D.; Suppes, P.; Siegel, S. (1957) Decision-Making: An Experimental Approach. Stanford University Press.
  6. Davidson, D (2001). Essays on Actions and Events. Oxford University Press.
  7. Maslow, A.H. (1943). A theory of human motivation. Psychological Review 50 (4) 370–96. Retrieved from
  8. Nagel, T. (1974) "What Is It Like to Be a Bat?" The Philosophical Review, Vol. 83, No. 4 (Oct., 1974), pp. 435-450
  9. Nuxoll, A. (2007) Enhancing Intelligent Agents with Episodic Memory;jsessionid=17B337D6E41731B075E9C653EA0CF1A7?sequence=2
  10. Raaijmakers, J.G (1993) "The story of the two-store model of memory: past criticisms, current status, and future directions". Attention and performance. XIV (silver jubilee volume). Cambridge, MA: MIT Press
  11. Sampson, G. (2005) Educating Eve: The 'Language Instinct' Debate. continuum.
  12. Thorndike E.L.; Woodworth R. (1901) The Influence of Improvement in one Mental Function upon the Efficiency of Other Functions. Psychological Review, 8, 247-261.
  13. Weizenbaum, J. (1976) Computer Power and Human Reason: From Judgment To Calculation. San Francisco: W. H. Freeman