David Ganzhorn
ganzhorn@cs.rochester.edu
CSC240
Paper 7
Book Review: Descartes^Ò Error

Descartes^Ò Error, by Antonio R. Damasio, takes a unique approach to creating a system of understanding of emotions and feelings. Rather than avoiding the often unclear topic, Damasio starts from scratch and slowly builds up a possible theoretical understanding of what emotions are and why they are important. Although he often cites cases and studies to back up his speculation, Damasio^Òs goal is to show that there is a logical (if not entirely correct) explanation of how emotions are critical for reason and decision. Although Damasio goes on to tentatively theorize about consciousness, the meat of the book is all about emotions role in rational thought.

The story all begins with Phineas Gage, the famous case where a man had a iron rod blasted through his head, and survived with no immediately apparent permanent damage. However, in months to come, it became clear that he no longer could make sound decisions; he was horribly rude, could not keep a job, and was unable to take care of his life. In recent times there have been several more cases like Gage, and it was once widely thought that their problems with decision making were due to psychological issues that were the fault of the individuals. However, it is now generally accepted that severe damage to the pre-frontal cortex will result in severely impaired decision making. What is so mysterious about these cases is that the patients seem to have perfectly intact cognitive abilities; when tested they have high IQ^Òs, a good sense of morality, and can generate numerous possible consequences of a given decision. But, these patients simply can not decide on which decision to make. Damasio^Òs whole case for the role of emotion in decision making stems from how all of these people have lost almost all sense of emotion. When shown films of disasters, death, and mayhem, the patients remain unaffected. If they nearly get in a car accident, they will later talk about it without a hint of anxiety. Yet, these patients are still capable of showing short bursts of emotion, such as a burst of anger, or startled fear. What is missing is the ability for these people to take into account the emotional guidance that normally steers one towards decisions that are good and bad. Somatic markers, markers of emotional significance, would normally allow these people to evaluate their possible courses of action as good or bad, but because of the damage to their pre-frontal cortexes, these patients no longer have access to somatic markers.

To explain somatic markers, emotion needs to be explained first. Damasio considers emotion to be the body's response to certain stimuli. Now, it is critical to note the distinction between emotion, which is a body state, and feeling emotion, which is a mental process that deals with the emotion. The body comes with several innate emotions: fear, happiness, sadness, anger, and disgust. However, life experiences give rise to various new emotions: boredom, elation, curiosity, and so forth. These emotions arise from the incredibly complex interactions of the body and brain; the brain release various chemicals, hormones, and signals into the body, which in turn sends chemicals, hormones, and signals back to the brain, in an endless cycle. The complex body states that arise are the emotional states of an individual. These emotional states turn out to be valuable sources of information to the brain because the body can be used as an extra information processor; the brain sends information about the current situation of the body, and the way that information affects the body determines the information the brain receives back. A good example of this is the brain signaling to the body that some primal fear instinct has been activated, perhaps by the perception of strange sound, and the body responding to this by stiffening up and increasing metabolic rate. Although the brain initially releases the hormones and chemicals into the body, the brain has no way of predicting the reaction the body will have; without a body, the brain would be deprived of a unique source of information. Damasio argues that this source of information is extremely valuable in deciding whether certain stimuli are good or bad for the perceiving individual. Over the years, this constant feedback from the emotional state of the body allows the brain to create an efficient system for reasoning and deciding: the brain generates numerous possible courses of action, and each course of action is evaluated by emotion (a poor idea might result in a bad gut feeling, a good idea may result in a sense of elation). Because it may not be possible to change body state fast enough to consider all of these possibilties, Damasio postulates that after enough emotional feedback, the brain can simulate the bodies response to known stimuli, in order to more rapidly process information. The ability to assign value to different possibilities allows the brain to both quickly eliminate poor ideas, so that no time is wasted considering them, as well as to eventually identify the best idea. The emotional value assigned to each possibility is what Damasio calls a somatic marker.

Damasio admits that this theory of emotional decision making is not by any means an absolute certainty, but he does provide enough evidence in support of the constituent assumptions that his idea ends up seeming quite plausible. If Damasio is correct in his theory about emotions role in decision making, then there are significant consequences. First, Emotion could no longer be considered to be a lowly aspect of humanity, inferior to logic and morality. Second, computational models of the brain would most likely need to take emotions into account.

Although some might see computationally modeling emotions as being impossible, I see it as possibly simplifying the overall problem of AI. One stumbling block in models of the brain has been trying to decide upon courses of action without having to calculate every conceivable ramification. Backgammon is a good example of this; all of the possible combinations of pieces on a board is far too large for any modern computer to logically with. However, a very good program has been created that takes an almost emotional approach to playing - it evaluates certain features of the board by how good they are, where good is defined as how indicative they are of winning. By making moves that are good, the program can very successfully play backgammon.

Extend this example to real life, and it can clearly be seen that trying to make a computational model of the brain, that logically computes every possible course of action, is likely impossible. I believe that the first highly successful models of the brain will make extensive use of emotion like systems to help reduce the complexity of computation that must be done to arrive at very good, if not perfect, decisions. After all, humans are not perfect, so a computational model of humans must be imperfect to be accurate.

I highly recommend Descartes^Ò Error to anyone who is interested in computational models of the brain, and to anyone who is interested in reason and emotion in the brain. My only reservation in reccomending this book is its cumbersome style, but the very provactive content is worth the effort to read Descartes' Error.

Books Cited

Damasio, R. Antonio. Descartes^Ò Error: Emotion, Reason, and the Human Brain. Avon Books, New York, New York. 1994