As mentioned in the previous article, which was about AGI and labor, the idea of specialization can be quite dangerous. We need only consider how signals flow throughout a nervous system, or how people communicate, to see what happens when one piece of information becomes overly localized.
But first, let us consider how society becomes specialized; in what ways do we localize things? Of course, we will most likely cover a mere fraction of things which can be localized, though a few shall allow us to find the others ourselves.
Of those things which have been specialized, among them are education and occupations; certain countries in their manufacturing products for major corporations; and furthermore, certain corporations have begun to specialize in intelligence research as well. These are just a few examples of specialization, and we will discuss the possible dangers of each by treating of the subject like an information or signal theorist would. That is, when we speak of specialization or generalization, we are speaking of them as principles of organization for complex systems, not individual instances.
Culture and blindness: the dangers of specialized education
The modern university gives to its students, particularly those high in conformity, knowledge and training which are related to only their academic disciplines. The students enter first-year, equipped with little education and critical thinking skills, and become deeply entrenched in their discipline.
The foundations of their minds are laid early within their education by the initial introductory courses which are mandatory; from which follows another three or four years of thorough training within that discipline. In addition, the courses that become available past the initial years demand even more of the person, and so the time which can be allocated to extracurricular education becomes even more scarce. So, to perform well in their discipline, they need spend all their available time studying their disciplines texts; therefore, in a sense becoming culturally programmed.
Cultural programming is necessary to some extent; without having a common culture, there would be immense inefficiencies in communication, time-expenditure, and productivity. For example, if there were no classes which provided an introduction to developmental psychology, then professionals would have to explain Piaget within each article they write that regards Piaget’s work, which would lower productivity and the efficient use of time; only because his work would be drastically less popular than it currently is within psychology culture. And so, some cultural programming increases productivity since less time is being wasted on developing a standardized foundation to approach the literature from; thus, making cultural programming a necessity to some extent.
However, there are dangers to cultural programming as well, and, by extension, the principle of specialization. To give some examples, we will consider the frontal lobotomy, Freud, and psychometrics. These, I claim, all stemmed from information encapsulation.
The frontal lobotomy is a perfect example of when disciplines lack a more generalized understanding. To elaborate, each discipline usually has their own way of thinking about the world; for instance, a neuroscientist thinks more like a mechanic when it comes to understanding complex systems that have functions, even though their content is different; a psychologist thinks more like a medical doctor when it comes to the assessment of humans; and a sociologist thinks more like a macroeconomist when it comes to complex social systems. These ways of thinking are developed because of the content each discipline is interested in. Brains are complex systems that have nuanced functional components, psychology requires the analysis of people to determine a mental bill of health, and sociology involves the analysis of outputs from complex systems. So, a particular way of thinking about reality is developed when specializing in a discipline, which frequently brings about a blindness to other ways of thinking. This was one of the many reasons why the frontal lobotomy came about; that is, the impact on the whole person by alteration of the person’s constituent parts was left unconsidered by those who performed the operations. To put that in other words, the gestalt of a person’s psychology was left unconsidered; a blindness to other perspectives about humans lead to real damage. This can happen a lot in academic disciplines as well.
Another important example is Freud and his psychoanalysis. Freud started off in neurology and eventually moved to psychology; one of the reasons for this change in direction was his beliefs about neurology research. Freud believed that brain research was nowhere close to being able to develop an image of mental life, which prompted him to begin his own analysis of psychic life independent of neurological methods.
Freud specialized in armchair speculation, and he interpreted all his observations within his special theoretical framework. This lead to a great many errors in reasoning, and the reason for such still happens today; namely, his armchair analysis had never been ran up against the physical limitations of a brain, which, if it had, would have stopped some of his bad theories. For example, consider Freud’s idea about repression. According to Freud, people repress negative memories, thoughts, and events into the unconscious mind, so as to keep them away from awareness. As lovely an idea as this may be, we now know that the pathway from the frontal cortex to the amygdala is much weaker than the pathway from the amygdala to the frontal cortex. This tells us that people most likely do not suppress things into the unconscious since the impulse control mechanism is simply too weak to do so. An unfortunate example of this relates to PTSD; people with PTSD have hyper-active amygdale, and so are in constant search for a threat in the environment. And furthermore, it would seem that the impulsive “id” controls the “ego” or “super-ego”. Thus, the ideas which Freud had been emersed in lead to a bias in worldview, and that bias brought about bad ideas.
To really grasp the impact Freud had on society as well, consider that Freud, Jung, and all the other psychoanalysts. These men and women were far from dumb; indeed, Freud had contributed a significant amount to neurology in his early career, and he even drew diagrams of neural networks in his personal notebooks when thinking about the empirical basis of psychology. If it were not for psycho-analysis, these men could have lent their intellectual abilities to other areas of society, they could have developed work which was far more empirically sound and theoretically coherent. But, due to information encapsulation, they have plagued psychology and its students with massive divides, a bad public reputation, and sheer nonsense. A massive amount of inefficiency could have been avoided by something as simple as walking to the neurology department and saying, “Hey, can you tell me which of these ideas sounds neurologically impossible to you”.
And this leads us to another instance of specialization within psychology; namely, the measure of neural activity by measuring responses to questions: psychometrics. This field, psychometrics, develops theoretical constructs, like IQ, and creates an inventory of questions or statements that a person must respond to; in doing so, the person receives a score based on how they respond to the inventory of questions or statements. This method of inquiry essentially defines intelligence, personality, and other variables within psychology. Of course, there are some useful predictions which can be performed by the inventories and scores, but as people know, humans are more than the response they give on a test. A perfect example would be when someone was excluded from a psychological study of high IQ individuals, only scoring a mere 129, who then went on to win a Nobel prize. This specialized psychometric view of humans can lead people to focus on test scores rather than humans, which develops, yet again, a functional blindness to other important variables: i.e., creative capacity, base-metabolic-rate in the frontal cortex, or goal selection. These are things which are ignored by the psychometric method.
These are all instances where being emersed in a particular culture or way of thinking can result in detriment; be it bad decision making, the propagation of misinformation to thousands of people, or the over-emphasis on tests that are fundamentally incomplete. To be informationally encapsulated, and thereby specialized, can bring about massive inefficiencies and harmful net-effects.
The Brain and Complex Systems: specialization and function
As mentioned above, there are now countries which are becoming specialized at the task of manufacturing goods. The analysis we will give of this will be derived from clinical neuropsychology, as clinical neuropsychology has some of the most vivid examples of the dangers of damage being done to specialized structures.
Consider the traditional view of language production, which is now an outdated model, but will do for example alone. The traditional view of language production, known as the Geschwind-Lietcheim-Wernicke model, attributed specific functions to a network of structures; for instance, the Broca’s was involved in language production, the Wernicke’s was involved in language comprehension, and reading out loud was produced by the arcuate fasciculus. It was believed that damage to these specific structures, such as the Broca’s area, resulted in aphasias of various types: i.e., Wernicke’s or Broca’s aphasia.
Thus, damage to just a fragment of the areas which produced a signal resulted in a complete failure of the entire system. Damage to either the Broca’s or Wernicke’s area resulted in complete linguistic dysfunction. So the specialization of functions in certain anatomical areas was high risk insofar as slight damage to one area meant complete dysfunction in the entire system; whereas, a distribution of functions may result in less dysfunction upon damage.
And when it comes to the specialization of production by country, the same analysis applies. If something becomes specialized in one country, and that something turns out to be absolutely vital to the functionality of some product or job, then it is high risk to specialize; that is, if a natural disaster or political event damages the facility in that country, then the entire system which depends on that facility will also fail.
Telephones: signal distortion and specialization
Many of us have played the childhood game named telephone, but for those of us that have not, the game goes as follows: there is a linear line of 30 people, and the person first in the line starts with a message to pass along. Once the person thinks of the message, they then whisper it into the ear of the second person; from there, the function of whispering the message repeats until it reaches the last person. Arriving at the end, the last person speaks the message out loud to see whether the initial message has reached the last person or not. Majority of the time, the message heard in the middle of the line is different from the message heard at the start and end, and the message heard from the start is different from the message heard at the end. This highlights an important idea: namely, signal distortion.
Imagine each person in the line as a specialized communicator, each functioning independently. Because each of these communicators has a variation in biology and amount of air flow between one another, noise continually gets introduced into the system; therefore, the signal becomes more distorted each time it is passed along this line of individual communicators. The variation in receptors for audition, the variation in top-down expectation, and the variation in both air pressure and speech production all introduce minute distortions of information that compound as the message propagates down the line.
The point of that example is to demonstrate the idea that the idiosyncrasies of a single unit influence the output of the system, and the more individual units there are, the more signal distortion we will have. Of course, there are ways to combat signal distortion, such as the development of more standardized equipment; however, with humans, that is easier said than done. That is, one can train someone to write a piece of data in a particular format, but the syntax, word choice, and writing style would still influence the specific presentation within that format. So, specialization can increase signal distortion, which runs the danger of getting the wrong signal completely.
Conclusion: to specialize or generalize
Complex systems are all around us. A classroom is a complex system, a highway is a complex system, and an economy is a complex system. Because there are so many complex systems around us in our daily lives, so too are specialization and generalization. Consider, the education one receives is done so in a specialized discipline, the medium of exchange one operates within is usually specialized: paper money, and the way we think has become specialized as well. This should prompt us to think about the dangers of being educated in only one discipline, or trading with one another only using paper money; for instance, if one is trained in computer science only, perhaps they may want to also generalize out to linguistics as well; only so they are more likely to be hired in an AI research Job for natural language processing.
That said, we indeed cannot do away with specialization entirely; there are plenty of benefits to specialization, as Adam Smith so willingly conveys, but we need to be realistic about specialization. If we were to specialize every time fortune occasioned us with the opportunity, we would be doing so irrespective of the cons of specialization. The conclusion to take from this discussion, most certainly, is that we should critically think about the possibilities of information encapsulation, system damage, or signal distortion; that is, consider the likely hood of these events, consider whether your system is susceptible to them, and develop prevention plans. Maybe even counter-balance some aspects of specialization with more generalization in another area. Specialization is far from inherently good.