Why Poetry Cannot Emerge From GOFAI and why The Extended Mind Approach is a Better AlternativeGOFAI was once thought of as a pathway that could eventually lead to the emergence of a machine intelligence that supervenes from circuitry and in the same way that human intelligence was thought emerge from the connectedness of neurons. Intelligence breaks down to reasoning, which is the manipulation of representations (PHIL342, Lec 7b). GOFAI does is based upon symbolic logic. GOFAI was once theorized as the way that machines could have minds. But while GOFAI systems have been able to mimic human behavior, and sometimes even eerily resemble it, these systems are essentially just glorified calculators. Yet theorists in the 70s began backing away from thinking that GOFAI could lead to actual AI. Speaking of the then cutting edge natural-language understanding program, SHRDLU, Ubert L. Dreyfus writes that though this “seemed” a step toward intelligence, that this line of thought rested on incorrect premises (Dreyfus, 147). One of the highest outcomes of complexity of human intelligence in can be found an artistic endeavor such as “poetry” which relies on metaphors, something unique to human intelligence. Poetry requires intelligence to be written. Since the demise of the promise of GOFAI, other architectures of the mind have filled the gap. The extended mind approach is one of the most compelling models since the demise of GOFAI and provides a better model for understanding how poetry emerges from the human poet. Though much research remains to be done, this vein of thought could perhaps be the route or a component of what would be necessary for a machine mind to one day write a poem rivaling something created by a human mind. To create poetry, a poet uses language in non-obvious, often contradictory ways. It requires much from the reader in order to arrive at understandings. A metaphor is a “brief, compressed comparison that talks about one thing as if it were another’s,” (HelpMe.com, n.p.). A good example of how this works can be found in Shakespeare’s Sonnet 73, which contains three metaphors. In the first metaphor, Shakespeare compares himself to a tree. He writes, “That time of year thou mayst in me behold when yellow leaves, or none, or few, do hang upon these boughs which shake against the cold” (Shakespeare, 2-3). A literal interpretation of this would mean that the speaker (Shakespeare) had leaves and branches. But everyone but the densest readers knows that this is not what Shakespeare means. He is comparing the fall with human aging and mortality. Just as a tree losing its leaves in the fall, humans towards the end of their life begin to lose certain abilities they had during their youth. A human mind makes this connection seamlessly. Yet there is not a machine in the world that could understand this. So this understanding provides a good benchmark between what seems like intelligence and what is actually intelligence. R. A. Brooks writes, “Abstraction is the essence of intelligence and the hard part of the problems being solved. Under the current scheme the abstraction is done by the researchers leaving little for AI programs to do but search” (Brooks, 139). Abstraction is necessary for understanding poetry. In the example of the sonnet, the abstraction necessary to understand the metaphor is done instantly, easily and seamlessly in the human mind. Yet this is something that a machine simply cannot do. Scientists still have no idea about how the human mind is able to make the cognitive connections in order to understand metaphor. One model that has been thought of as an alternative to GOFAI is Connectionism. It includes thoughts, different approaches, a typical psychology and science and the minds philosophy. Connectionism is divided into many forms but neural network model is a common form. It is low-level model with a quite general framework that is provided to a human. It does not target any specific thing (Smolensky, pp.10). Connectionism rests on the principle that with enough interconnections, intelligence is emergent from these complicated systems. It rests on the principle that patterns, complicated enough will lead to human intelligent. Yet this reduces human beings to just a mind, not the fullness of the brain and the body and how somehow from the interplay between the two, the mind and human intelligence emerges. Because of this, it has many perceived limitations. So while AI may one possible, and we may one day even be able to build machines that could understand (or even write) poetry, it seems that connectionism may be one component of intelligence, since certainly the connection of neurons is part of the equation, there is likely much more to intelligence. Likewise, dynamic systems have also been proposed as an alternative to GOFAI in creating actual intelligence. Yet this has the same drawbacks as connectionism, as it is mathematically based. Dynamic systems. Yet there is a part of this related to the preferred model in this, that is, that multiple approaches and perspective might be required for intelligence: “Development can only be understood as the multiple, mutual, and continuous interaction of all the levels of the developing system, from the molecular the cultural” (Thelen and Smith, 258). Another alternate is dynamical system as it provides a value with the reference to the time but basically it is a mathematical term so it majorly involves calculations. It explains spring pendulum or clock pendulum or may be a water flow rate in a pipe (Tong, pp.10-15). In addition, a dynamical system gives a description on how points in geometrical space rely on time. It is like, for instance, the wavering of a clock pendulum. Moreover, in a dynamic system, small changes in system state create change in numbers. There is a fixed rule that explains the future states based on the current state. The difference between GOFAI and Extended Mind Approach (Robotics) is that GOFAI programs attempt to adapt the ability of a human brain to represent the globe with symbols and reason logically in manipulation of the symbols. On the other hand, Robotics were designed to handle specific dynamic function because of the tails for balance, joints in particular places and springy limbs.in addition, robotics do not respond to computed commands. (Tong, pp.10-15) This is a merit for positive though, since the human mind does not respond to computed commands either. This provides the best model to explain poetry because this model is not one specific thing. Human intelligence is complex, and needs many models to explain and understand it. The Extended Mind Approach is not limited to one thing, and therefore it is the best model that can be used to explain poetry. It certainly provides a better model than GOFAI, which relied on logic, since poetry often defies logic in order to express meaning. Human behavior can be analyzed from a variety of viewpoints. Neuroscience looks at brain activity, or the underlying causes of behavior. Psychology looks at behavior in order to understand the brain. In the biological assumption, the brain always understands. It seems to work like a switchboard or computer. But a computer was designed by engineering. The model compares to the brain with co-related work in neurophysiology, which discovered that neurons produced a burst of electricity. The burst is taken as the brains system unit correspondingly to the information in the computer. In addition to that, psychologically assuming, the computer model fails to match fully the brains operations. The current research, development and theory suggest a computer system able to grasp poetry cannot be based upon a reductionist theory if emergence is to be a property of it. An example is money. A quarter can be reduced down to molecules, or the metal that composing it, but in the process, the concept of what a quarter is, money, is lost. The extended mind approach offers a middle ground where multiple systems and approaches can be mixed together. It reduces a human effort and can mix various models. It does not involve a mindset of thinking beyond the limitations and no strategies or plans are there to control or manage all the scattered thoughts. Understanding poetry involves understanding concepts, which transcend their parts. By allowing things outside of the system to be part of the system, it helps integrate the mind-brain-body question. As models develop and move away from GOFAI, it may turn out that the body is more important to the mind. GOFAI used flawed assumptions about how the brain operated and essentially left out the body. But the extended mind approach leaves the door open to include the body and the organic chemistry. It is likely that understanding poetry does not reduce into something as simple as a complexity of connections, but instead is combination of patterns, connections, chemicals and the language that brain uses to think and communicate with itself.
References:
Brooks, R.A. “Intelligence without representation” Artificial Intelligence Volume 47, Issue 1-3, 1991, 139
Dreyfus, Herbert L. "From Micro-Worlds to Knowledge Representation: AI at an Impasse." 1 Jan. 1979. Web. <http://www.cogsci.ucsd.edu/~coulson/203/dreyfus.pdf>."Metaphors in Sonnet 73." 123HelpMe.com. 02 Dec 2014
<http://www.123HelpMe.com/view.asp?id=9407>.
"Shakespeare Sonnet 73 - That Time of Year Thou Mayst in Me Behold." Shakespeare Sonnet 73 - That Time of Year Thou Mayst in Me Behold. Web. 2 Dec. 2014. <http://www.shakespeare-online.com/sonnets/73.html>.
Thelen, Esther, and Linda Smith. "Dynamic Systems Theories." 1 Jan. 2005. Web. <http://www.iub.edu/~cogdev/labwork/handbook.pdf>.
Smolensky, Paul."On the proper treatment of connectionism." Behavioral and brain sciences 11.01 (1988): 1-23.
Tong, Howell. Non-linear time series: a dynamical system approach. Oxford University Press, 1990.
