Sunday, March 29, 2009
Tuesday, March 24, 2009
Working Outline (04/15/09)
The material I have in mind divides naturally into three parts:
I. Introductions-the core structure of the argument
II. Implications-what does it mean?
III. Applications-how can it be used?
It may also prove valuable to set the table with a general philosophical framing of the need for this thesis. For now I intend to begin with the meat of the argument.
The thesis proposal has taken several layers of rewrites to achieve a stable wording (as of 06/03/09), but it finally says more or less what I want it to say. The conceptual structure it establishes will be elaborated in a series of separate posts aimed at the general reader. Hopefully, this piecemeal approach will enable more frequent posting as well as more precise targeting of comments.
Active links in the outline below will go to finished posts, and the dates indicate the most recent revisions.
Part I:
STRUCTURE OF INTELLIGENCE: The first half of Part I will be taken up with a nontechnical introduction to the structure of intelligence, thus setting the conceptual stage for the presentation of the emotional core of the thesis in the second half.
Intelligence I: Representations (04/07/09) A general definition of intelligence as pattern-recognition, retention, and use, together with a short introduction to the neurological basis of this pattern representation.
Intelligence II: Prediction (04/03/09) If I start counting "one, two, three, ...," your memory supplies the next term easily even though it hasn't yet occurred. This post introduces the central role of prediction in intelligence.
analogy An equally significant staple of human intelligence, as indicated by the ease with which "wubbun, tooboo, threebee, ...," can be decoded despite the lack of a specific memory match.
internalized concepts what confers the ability to execute complex learned behaviors (like driving) without paying attention?
computational approaches links to (and possibly an overview of) various efforts in this end of the science
INSTINCTS AND EMOTION: Once a conceptual framework for intelligence has been established, it becomes possible to lay out the heart of the thesis, which proposes that the logic of intelligence can be extended into an elegant but powerful explanatory theory of emotion. Possible subheadings include:
the biology of instincts v reflexes; types of instinctive drives
explanation and development of thesis, with examples and thought experiments
moods v emotions
connections to other theories of emotion (neuroscience, psychology, philosophy)
valence
somatic
judgement
perceptual
other
etc.
More subheadings will emerge, of course. Perhaps philosophical interludes which foreshadow several connections of eventual importance will appear between some of the Part I posts, or perhaps not. Too early to tell.
Part II will most likely deal with corollaries and logical extensions of the unfolded definitional thesis of part I, examining implied integrations, connections, and limitations. It may be organized into broad categories like science, politics, and philosophy, or may be more highly interconnected. We'll see when we get there. A placeholder post has already been set up to get things started, but it must be written with great care and is therefore on hold until a large enough block of uncluttered thinking time emerges. Its structure will be informed by the flow of the arguments in Part I and will guide the eventual shape of the outline of Part II.
Part III will look at actionable policy recommendations and general strategies for applying the ideas found in parts I and II, as it appears that there will be some of a fairly profound and systemic nature. Again more details when they become practical.
Your input is solicited!
I. Introductions-the core structure of the argument
II. Implications-what does it mean?
III. Applications-how can it be used?
It may also prove valuable to set the table with a general philosophical framing of the need for this thesis. For now I intend to begin with the meat of the argument.
The thesis proposal has taken several layers of rewrites to achieve a stable wording (as of 06/03/09), but it finally says more or less what I want it to say. The conceptual structure it establishes will be elaborated in a series of separate posts aimed at the general reader. Hopefully, this piecemeal approach will enable more frequent posting as well as more precise targeting of comments.
Active links in the outline below will go to finished posts, and the dates indicate the most recent revisions.
Part I:
STRUCTURE OF INTELLIGENCE: The first half of Part I will be taken up with a nontechnical introduction to the structure of intelligence, thus setting the conceptual stage for the presentation of the emotional core of the thesis in the second half.
Intelligence I: Representations (04/07/09) A general definition of intelligence as pattern-recognition, retention, and use, together with a short introduction to the neurological basis of this pattern representation.
Intelligence II: Prediction (04/03/09) If I start counting "one, two, three, ...," your memory supplies the next term easily even though it hasn't yet occurred. This post introduces the central role of prediction in intelligence.
analogy An equally significant staple of human intelligence, as indicated by the ease with which "wubbun, tooboo, threebee, ...," can be decoded despite the lack of a specific memory match.
internalized concepts what confers the ability to execute complex learned behaviors (like driving) without paying attention?
computational approaches links to (and possibly an overview of) various efforts in this end of the science
INSTINCTS AND EMOTION: Once a conceptual framework for intelligence has been established, it becomes possible to lay out the heart of the thesis, which proposes that the logic of intelligence can be extended into an elegant but powerful explanatory theory of emotion. Possible subheadings include:
the biology of instincts v reflexes; types of instinctive drives
explanation and development of thesis, with examples and thought experiments
moods v emotions
connections to other theories of emotion (neuroscience, psychology, philosophy)
valence
somatic
judgement
perceptual
other
etc.
More subheadings will emerge, of course. Perhaps philosophical interludes which foreshadow several connections of eventual importance will appear between some of the Part I posts, or perhaps not. Too early to tell.
Part II will most likely deal with corollaries and logical extensions of the unfolded definitional thesis of part I, examining implied integrations, connections, and limitations. It may be organized into broad categories like science, politics, and philosophy, or may be more highly interconnected. We'll see when we get there. A placeholder post has already been set up to get things started, but it must be written with great care and is therefore on hold until a large enough block of uncluttered thinking time emerges. Its structure will be informed by the flow of the arguments in Part I and will guide the eventual shape of the outline of Part II.
Part III will look at actionable policy recommendations and general strategies for applying the ideas found in parts I and II, as it appears that there will be some of a fairly profound and systemic nature. Again more details when they become practical.
Your input is solicited!
Sunday, March 22, 2009
Intelligence II: Prediction (04/03/09)
Understanding the world through abstract neurological representations (as discussed in the previous post) is a wonderfully flexible adaptive strategy which makes all kinds of learning possible. However, in order for the strategy to be viable it must somehow produce information that is accessible in real time, or at least at biologically relevant speeds. According to Christof Koch, experimental psychology suggests that the amount of time necessary to consciously resolve a visual percept is on the order of a quarter-second (the pre-conscious processing of intelligence begins to react on much shorter time frames, and lies behind such concepts as "subliminal marketing," but is also by definition non-conscious). How then can the representational platform of intelligence make it possible for humans to execute cognitively complex tasks in real time, like hunting a deer, driving a car, or performing Dvorák's magnificent cello concerto?
The answer has to do with architecture. The reason intelligence is able to effect conscious control of events on short time scales is that it confers the ability to plan ahead, thus overcoming its own inherent neurobiological speed limits, by making predictions of future events based on information already received. Very roughly speaking, consciousness can stay one or more steps ahead of the situation by using the predictive machinery of intelligence and the storehouse of memory to set goals, and then modify or extend those goals based on what is actually happening.
This is most probably achieved through some biological analogue of an architectural property known to neural network enthusiasts as autoassociation, in which the full retrieval of a stored informational pattern (in this case a neurological representation) is triggered by the input of a partial but matching subset of the full pattern, given an appropriately configured network. This allows the complete perception of a learned pattern before it has finished unfolding in time or space, and it is why you can recognize partially occluded objects or anticipate the progress of familiar sequences. The cascading of this predictive effect upwards through the layers of representations (with each layer adding its own predictions) results in more and more sophisticated predictions spanning longer and longer time frames, ultimately pushing the operation of entire subsets of successful lower-level predictive representations below the threshold of awareness.
On a fundamental level, prediction is central to intelligence.
The clearest articulation of this predictive principle I have yet come across is the "memory-prediction" framework presented by Jeff Hawkins (video link courtesy zenpundit) in his 2004 book "On Intelligence," written with Sandra Blakeslee. "On Intelligence" provides a valuable overview of a number of topics related to the general problem of intelligence, including chapters on Artificial Intelligence, Neural Networks, and the Human Brain, in order to better argue the key roles of Memory and Prediction in what intelligence does, all on its way to expressing a more specific theory of how intelligence might actually work which is based on cortical architecture. While I am not yet well equipped to critique the specifics of his proposed theory, I find the resonances between his argument for intelligence-as-prediction and my extensive teaching, practicing, and performing experience as a professional cellist to be compelling.
From chapter 5 of "On Intelligence":
The above excerpt emphasizes the way parallel predictions across sensory modalities can be combined into rich representations. Let me offer two loosely paraphrased examples to illustrate the way Hawkins believes serial combinations might work.
1) When you learned to read, you started (like all of us) by learning to recognize letters. Once your letter recognition models became sufficiently reliable, your brain was able to use their output as the basis for constructing new predictive models that could recognize entire words. This reliability is why it is no longer necessary for you to consciously process every single letter when you read, although you can, if you focus your attention appropriately. Extending the example, we can say that further modeling of phrase units, sentence structure, and the rules of grammar and composition are the subjects of yet higher layers. Your memory-predictive mastery of lower-level tasks such as letter and word recognition frees up enough processing power to employ higher-level representations, and your resulting knowledge of vocabulary and grammar actually allow you to predict what you will read before you finish each word or senten (see?)
2) The expression "muscle memory" is often used to describe deeply learned complex movements, especially those of musicians and other performing athletes. The speed and relaxed precision of their movements are due to the layers of highly accurate predictive neurological representations of body structure and function (I expect my elbow to be here when I do this), integrated with equally well developed predictive models of the execution of the task at hand, whether it's meeting a ball with a diving catch or playing a cello concerto from memory. The unconscious ease of a physical talent at work bespeaks entire subsets of accurate models pushed below the threshhold of consciousness, whether they were discovered quickly and "naturally" (a great definition of intuition!) or learned and refined more slowly and painfully through extended trial and error.
By now it should be clear that the notion of intelligence as prediction is powerfully illuminating. Before turning to the equally powerful role that analogy plays in dramatically extending the reach of the predictive representations of intelligence, I'd like to add more of my own thoughts about how it is that representations become reliable enough to build into layers in the first place. These "internalized concepts" will be the subject of the next post.
The answer has to do with architecture. The reason intelligence is able to effect conscious control of events on short time scales is that it confers the ability to plan ahead, thus overcoming its own inherent neurobiological speed limits, by making predictions of future events based on information already received. Very roughly speaking, consciousness can stay one or more steps ahead of the situation by using the predictive machinery of intelligence and the storehouse of memory to set goals, and then modify or extend those goals based on what is actually happening.
This is most probably achieved through some biological analogue of an architectural property known to neural network enthusiasts as autoassociation, in which the full retrieval of a stored informational pattern (in this case a neurological representation) is triggered by the input of a partial but matching subset of the full pattern, given an appropriately configured network. This allows the complete perception of a learned pattern before it has finished unfolding in time or space, and it is why you can recognize partially occluded objects or anticipate the progress of familiar sequences. The cascading of this predictive effect upwards through the layers of representations (with each layer adding its own predictions) results in more and more sophisticated predictions spanning longer and longer time frames, ultimately pushing the operation of entire subsets of successful lower-level predictive representations below the threshold of awareness.
On a fundamental level, prediction is central to intelligence.
The clearest articulation of this predictive principle I have yet come across is the "memory-prediction" framework presented by Jeff Hawkins (video link courtesy zenpundit) in his 2004 book "On Intelligence," written with Sandra Blakeslee. "On Intelligence" provides a valuable overview of a number of topics related to the general problem of intelligence, including chapters on Artificial Intelligence, Neural Networks, and the Human Brain, in order to better argue the key roles of Memory and Prediction in what intelligence does, all on its way to expressing a more specific theory of how intelligence might actually work which is based on cortical architecture. While I am not yet well equipped to critique the specifics of his proposed theory, I find the resonances between his argument for intelligence-as-prediction and my extensive teaching, practicing, and performing experience as a professional cellist to be compelling.
From chapter 5 of "On Intelligence":
...[Y]our brain makes low-level sensory predictions about what it expects to see, hear, and feel at every given moment, and it does so in parallel. All regions of your neocortex are simultaneously trying to predict what their next experience will be. Visual areas make predictions about edges, shapes, objects, locations, and motions. Auditory areas make predictions about tones, direction to source, and patterns of sound. Somatosensory areas make predictions about touch, texture, contour, and temperature.
"Prediction" means that the neurons involved in sensing your door become active in advance of them actually receiving sensory input. When the sensory input does arrive, it is compared with what was expected. As you approach the door, your cortex is forming a slew of predictions based on past experience. As you reach out, it predicts what you will feel on your fingers, when you will feel the door, and at what angle your joints will be when you actually touch the door. As you start to push the door open, your cortex predicts how much resistance the door will offer and how it will sound. When your predictions are all met, you'll walk through the door without consciously knowing these predictions were verified. But if your expectations about the door are violated, the error will cause you to take notice. Correct predictions result in understanding. The door is normal. Incorrect predictions result in confusion and prompt you to pay attention...We are making continuous low-level predictions in parallel across all our senses.
The above excerpt emphasizes the way parallel predictions across sensory modalities can be combined into rich representations. Let me offer two loosely paraphrased examples to illustrate the way Hawkins believes serial combinations might work.
1) When you learned to read, you started (like all of us) by learning to recognize letters. Once your letter recognition models became sufficiently reliable, your brain was able to use their output as the basis for constructing new predictive models that could recognize entire words. This reliability is why it is no longer necessary for you to consciously process every single letter when you read, although you can, if you focus your attention appropriately. Extending the example, we can say that further modeling of phrase units, sentence structure, and the rules of grammar and composition are the subjects of yet higher layers. Your memory-predictive mastery of lower-level tasks such as letter and word recognition frees up enough processing power to employ higher-level representations, and your resulting knowledge of vocabulary and grammar actually allow you to predict what you will read before you finish each word or senten (see?)
2) The expression "muscle memory" is often used to describe deeply learned complex movements, especially those of musicians and other performing athletes. The speed and relaxed precision of their movements are due to the layers of highly accurate predictive neurological representations of body structure and function (I expect my elbow to be here when I do this), integrated with equally well developed predictive models of the execution of the task at hand, whether it's meeting a ball with a diving catch or playing a cello concerto from memory. The unconscious ease of a physical talent at work bespeaks entire subsets of accurate models pushed below the threshhold of consciousness, whether they were discovered quickly and "naturally" (a great definition of intuition!) or learned and refined more slowly and painfully through extended trial and error.
By now it should be clear that the notion of intelligence as prediction is powerfully illuminating. Before turning to the equally powerful role that analogy plays in dramatically extending the reach of the predictive representations of intelligence, I'd like to add more of my own thoughts about how it is that representations become reliable enough to build into layers in the first place. These "internalized concepts" will be the subject of the next post.
Friday, March 6, 2009
What's the point? Or, Logical and Philosophical Implications of the Thesis Proposal
The thesis proposal provides a definition of emotion, but that is only half the story. This post will explain its significance.
Comingsoon eventually.
Until then, I recommend moving on to the new Working outline post.
Coming
Until then, I recommend moving on to the new Working outline post.
Intelligence I: Representations (04/07/09)
The thesis proposal made mention of the central roles of prediction and analogy in intelligence but did not explain their significance, and it introduced "representations" and "internalized concepts" without defining the terms at all. This is the first post in a short series (designated Intelligence I, II, etc. in the sidebar's "Table of Contents") which will provide a conceptual overview of these basic elements of intelligence for the general audience, thus providing a fitting prelude to the full exposition of the emotional core of the thesis.
__________________________________________
The obvious place to start a discussion of intelligence is with a definition. Broadly speaking, by intelligence I mean the general capacity to recognize new patterns in the torrent of sensory input and then to further organize those informational patterns into useful mental representations, like "food" or "dangerous animal" or "chair" or "trigonometric function." I do not mean consciousness, attention, or even awareness, although these things are all closely related topics, and I certainly do not mean particular levels of education, capability, wisdom, or experience.
So far as I can tell, this definition puts me on well-traveled theoretical ground; although the exact procedural architecture of intelligence is not yet certain, the fundamental idea that intelligence must involve pattern-recognition seems to be generally agreed upon by most cognitive scientists.
These informational patterns that we perceive, such as "chair" or "Beethoven's 9th Symphony," are recognized using the various modalities of sensory input (sight, smell, hearing, etc.), but it's worth pointing out that by the time those inputs have reached the brain they have already been converted into neuroelectrical impulses, conveying pure information. In this sense, to the brain (or at least the portions of it associated with the operation of intelligence) all patterns are equally abstract. The principal difference between neurological representations of concrete objects like chairs and abstract concepts like trigonometry lies in the directness of the methods available for testing and refining their conceptual details, not in the manner of their construction. This is especially important to remember when comparing richly complex representations of highly abstract concepts such as "self-identity," which can be spread widely through space and time, to the actual limits of the physical body.
Another point about the "information" paradigm that is worth making is that because the essence of information can be codified and defined in the purely abstract world of mathematics, it's adoption (if correct) theoretically allows us to mimic the brain's neurologically-based computational prowess in another substrate, like a computer, if we can just figure out the right processing architecture. While the extraordinary complexity of the human brain makes this is a horrendously difficult and somewhat controversial proposition, the informational approach has yielded some computational insight into what might be going on and continues to be an exciting focus of research.
The final preliminary piece to consider is that the highly organized complexity of each of the representative examples given above stems from the brain's ability to layer concepts, assembling more sophisticated representations out of older, simpler ones. This is nicely illustrated by the way in which a small number of written letters (once learned) are assembled into syllables and then words, which can be further assembled into sentences and then compositions. The higher the number of levels, the richer the capacity for meaning. At the other end of the spectrum, well below the level of consciousness, are individual neuronal inputs whose firing might correspond to the presence or absence of a vertical edge in a certain part of the field of vision. Much more will be said about this hierarchical layering in due course.
So to sum up, the representations with which we understand the world around us are abstractly neurological in nature, although they are assembled from the input of our senses, and their layered structure is the source of their sophistication. This is just enough background to introduce one of the more surprising functions of successful representations--their capacity to predict future events--which is the subject of the next post.
Further Reading:
( listed in rough ascending order of technical detail)
Jeff Hawkins/Sandra Blakeslee "On Intelligence"
Patricia Smith Churchland "Brain-Wise: Studies in Neurophilosophy"
Christof Koch "The Quest for Consciousness: A Neurobiological Approach"
These and other works can be found on my References/Further Reading post.
__________________________________________
The obvious place to start a discussion of intelligence is with a definition. Broadly speaking, by intelligence I mean the general capacity to recognize new patterns in the torrent of sensory input and then to further organize those informational patterns into useful mental representations, like "food" or "dangerous animal" or "chair" or "trigonometric function." I do not mean consciousness, attention, or even awareness, although these things are all closely related topics, and I certainly do not mean particular levels of education, capability, wisdom, or experience.
So far as I can tell, this definition puts me on well-traveled theoretical ground; although the exact procedural architecture of intelligence is not yet certain, the fundamental idea that intelligence must involve pattern-recognition seems to be generally agreed upon by most cognitive scientists.
These informational patterns that we perceive, such as "chair" or "Beethoven's 9th Symphony," are recognized using the various modalities of sensory input (sight, smell, hearing, etc.), but it's worth pointing out that by the time those inputs have reached the brain they have already been converted into neuroelectrical impulses, conveying pure information. In this sense, to the brain (or at least the portions of it associated with the operation of intelligence) all patterns are equally abstract. The principal difference between neurological representations of concrete objects like chairs and abstract concepts like trigonometry lies in the directness of the methods available for testing and refining their conceptual details, not in the manner of their construction. This is especially important to remember when comparing richly complex representations of highly abstract concepts such as "self-identity," which can be spread widely through space and time, to the actual limits of the physical body.
Another point about the "information" paradigm that is worth making is that because the essence of information can be codified and defined in the purely abstract world of mathematics, it's adoption (if correct) theoretically allows us to mimic the brain's neurologically-based computational prowess in another substrate, like a computer, if we can just figure out the right processing architecture. While the extraordinary complexity of the human brain makes this is a horrendously difficult and somewhat controversial proposition, the informational approach has yielded some computational insight into what might be going on and continues to be an exciting focus of research.
The final preliminary piece to consider is that the highly organized complexity of each of the representative examples given above stems from the brain's ability to layer concepts, assembling more sophisticated representations out of older, simpler ones. This is nicely illustrated by the way in which a small number of written letters (once learned) are assembled into syllables and then words, which can be further assembled into sentences and then compositions. The higher the number of levels, the richer the capacity for meaning. At the other end of the spectrum, well below the level of consciousness, are individual neuronal inputs whose firing might correspond to the presence or absence of a vertical edge in a certain part of the field of vision. Much more will be said about this hierarchical layering in due course.
So to sum up, the representations with which we understand the world around us are abstractly neurological in nature, although they are assembled from the input of our senses, and their layered structure is the source of their sophistication. This is just enough background to introduce one of the more surprising functions of successful representations--their capacity to predict future events--which is the subject of the next post.
Further Reading:
( listed in rough ascending order of technical detail)
Jeff Hawkins/Sandra Blakeslee "On Intelligence"
Patricia Smith Churchland "Brain-Wise: Studies in Neurophilosophy"
Christof Koch "The Quest for Consciousness: A Neurobiological Approach"
These and other works can be found on my References/Further Reading post.
Thursday, March 5, 2009
What is this Blog? (03/18/09)
Greetings, and Welcome! This blog exists to share an elegant new definition of emotion, to extend that definition into a powerful explanatory theory of human nature, and to explore the resulting structural implications for human affairs great and small. The process will benefit from the give and take of conversation, so I hope you'll comment or ask questions.
My goal is to write clearly and informally enough for the interested bystander, but also with sufficient breadth and detail to give this argument the beginnings of academic legs. Hopefully I can do both at the same time.
New posts will appear on top as they are written, but their sequence is not firmly tied to their place in the overall argument. For structural reasons a clear introduction of many of my argument's conceptual components creates chicken-and-egg type problems of presentation, so for now I am simply letting the posts write themselves in whatever order they prefer; I'll be filling the gaps as needed.
For those who prefer a more organized reading experience, I am listing the posts in what I believe to be a more helpful reading order under the "Table of Contents" heading at the top of the sidebar. As the blog expands I will also be directly connecting related posts to each other by adding appropriately placed links (for example, next up in the sequence is the thesis proposal), and I'll eventually add tag functionality for a way of searching posts by topic once things really get rolling.
One final caution! The clarity of my writing improves dramatically when I reword and revise over an extended period of time, so rather than waiting to post until everything is perfect (my purpose is conversation, after all) I intend to continue editing until I am satisfied, indicating the most recent revision of each post with a date at the end of its title. Major revisions to content will probably generate new posts, but wording changes in the service of clarity will not.
If you'd like to see a particular topic addressed sooner rather than later, drop a request into the comments and I'll see what I can do!
And now, onward to the thesis proposal!
My goal is to write clearly and informally enough for the interested bystander, but also with sufficient breadth and detail to give this argument the beginnings of academic legs. Hopefully I can do both at the same time.
New posts will appear on top as they are written, but their sequence is not firmly tied to their place in the overall argument. For structural reasons a clear introduction of many of my argument's conceptual components creates chicken-and-egg type problems of presentation, so for now I am simply letting the posts write themselves in whatever order they prefer; I'll be filling the gaps as needed.
For those who prefer a more organized reading experience, I am listing the posts in what I believe to be a more helpful reading order under the "Table of Contents" heading at the top of the sidebar. As the blog expands I will also be directly connecting related posts to each other by adding appropriately placed links (for example, next up in the sequence is the thesis proposal), and I'll eventually add tag functionality for a way of searching posts by topic once things really get rolling.
One final caution! The clarity of my writing improves dramatically when I reword and revise over an extended period of time, so rather than waiting to post until everything is perfect (my purpose is conversation, after all) I intend to continue editing until I am satisfied, indicating the most recent revision of each post with a date at the end of its title. Major revisions to content will probably generate new posts, but wording changes in the service of clarity will not.
If you'd like to see a particular topic addressed sooner rather than later, drop a request into the comments and I'll see what I can do!
And now, onward to the thesis proposal!
Subscribe to:
Posts (Atom)
