Two inputs coming together perturb my system to a new understanding. First up is a brief report I saw in the Duke Chronicle about research at Duke University: https://www.dukechronicle.com/article/2019/01/the-results-were-surprising-duke-university-study-proposes-new-model-of-motor-learning. Scientists studying motor learning found results that confounded their theoretical expectations, a small thing really, but oh so important, I think. Good empiricism seeks just such a finding, one that challenges assumptions, and I think the assumption challenged here is central to the orthodox view of life processes.
The Duke researchers studied motor learning as facilitated by the cerebellum (using rats but they assert that the cerebellum is fairly ‘conserved’ in evolution across species, meaning that cerebellums are similar between paleo- and modern mammals, e.g., us). They expected to find a process whereby learning comes through feedback, i.e., error detection and correction, but instead they found what I will characterize as a feedforward process, i.e., no preset standard whereby new signals are compared to old in the effort to achieve a match, but rather new signals become the old in a cascading process. Imagine two waves at the beach, one sensing the development of its shape and adjusting to achieve a desired wave form, i.e., feedback, and one building based upon current conditions of wind, bottom terrain, depth and shoreline, i.e., feedforward. (Of course the latter is what we naturally see which is why we love to watch them—they are like a musical melody giving us a sense of the immanent future: see post 3/26/16 More about musical import).
So these key cerebellar cells fired more when the movement was correct, no feedback needed. The rat knew how to move successfully without feedback, kind like when we enjoy freeform dancing, no set pattern and who cares who is watching (or what our cerebellar cells are telling the scientists) but feel the simple joy of movement without prejudice, guided only by the flow of one movement to the next according to endogenous dynamics. Know what I mean?
I can already hear some of my more intelligent readers ask why wouldn’t we expect such a process for learning? And that has to do with the cultural development in our doxa of how we conceptualize life processes, at least what I think has been the more orthodox view. If you thought to ask the question above, I think you might be a little heterodox yourself, but let’s consider this proposition.
I base my analysis here on Pierre Bourdieu’s conception of the doxa. The doxa is the whole field of discourse; it is what we are able to conceptualize for discursive discussion. Orthodoxy is, of course, the current dominant paradigm for rendering our concepts for discussion, while heterodoxy comprises some alternative ideas that challenge the orthodox view. Usually the orthodox is major and the heterodox minor (I am an old guy without much technical savvy; I tried to do a Venn diagram of the doxa etc. but cannot get it to post here), but we work at modifying the orthodox and sometimes supplanting its ideas in a paradigmatic shift. This is both a cultural phenomenon and a lovely feature of good science. In the above research the orthodox would stipulate that sensorimotor learning involves negative feedback, but they did not find that to be the case, so now that suggests a heterodoxical view—some other process, e.g., feedforward, facilitates this learning.
Okay, as to my second input I am reading a remarkable book from 1985 by Susan Oyama, The Ontogeny of Information. I am sure I will write more about it later but here is my understanding so far that is relevant today. Oyama presents a heterodox idea to supplant the orthodox ones of genes as controlling life’s flow and the long time distinction of nature-nurture. This intellectual effort is broad and deep and, what I appreciate a great deal, very well written ( & so understandable by someone outside the field like myself with some effort). Her polemic covers a lot of ground as she points out how many deride the nature-nuture distinction and so few, very few it would seem, actually come close to conceptualizing without it. At its core her argument focuses on our predilection to think genes are quite powerful and even autonomous in controlling our phylogeny and ontogeny, when they are actually the seeds initiating the chemical reactions which are multiply determined by features of their context of both external ambient and internal current states. All of life is ontogenical, as it were, to coin a word, as these chemical processes flow and cascade through time, any one moment or phase the result of its history and current states. Our ‘nature’, she says, is a product of our ‘nurture’ and all that contributes to our ontogeny, i.e., genome, somatic ecology, environmental ecology, history, developmental status, evolutionary status, etc., composes the overall process of ‘nurture’ more or less equitably. Long story short, control of biological processes is multiply determined moment by moment, a cascade of operations which we analyze on a number of levels or from several perspectives, but experience much difficulty is seeing the gift of life whole. So, Wow! In her words, “Nature and nurture are not alternative causes but product and process. Nature is not an a priori mold in which reality is cast. What exists is nature, and living nature exists by virtue of its nurture, both constant and variable, both internal and external.”
Now this is a view of life I really go with and I will finish reading this book and then read it again. It appears right now that for me this book will rank up there with Monod’s Chance and Necessity(which Oyama says ascribes too much control and power to the genome), Panksepp’s Affective Neuroscience, and Langer’s Feeling and Formand other of her works. Enough rhapsodizing now; back to the book’s input that is relevant to the motor learning research and neural processes.
The enormous intellectual power of information theory has circled a good portion of the doxa into orthodoxy, e.g., we analyze neural processes and functions accordingly. As I have long held, and I am in good company here, information machines and living organisms have some similarities and some deep differences. The latter have often been relegated to the heterodox circle of the Venn diagram above. What I learned from Langer so many years ago is that the creative vitality and autonomy of life is its own nature. We can study it, create lively artifacts, use simulation to understand it, render it through positivism and scientific analysis, convey it aesthetically and discursively, etc., but all while we live it. (Back to rhapsody: Oyama keeps her exposition grounded, I think, in just this perspective. That makes for great understanding and writing. Thank you, Dr. Oyama).
That the Duke researchers were surprised by their finding that learning was also based upon feedforward, which we understand much less because it is less amenable to mechanical operations and very difficult to simulate in life processes adequately, and not always on feedback reflects this boundary between orthodoxy and heterodoxy. How we view and understand life, as Dr. Oyama says, determines what and how we study it, what is real data, and how it can be interpreted.
To begin wrapping this post up, consider this passage from her book:
“Somerhoff particularly warns against making facile assumptions that the brain, for instance, must contain comparators and be controlled by explicit error signals and command signals, just because certain servo-mechanisms work in this way. He points out that uncritical adoption of machine concepts and reification of input-output relations may encourage fruitless searches for nonexistent brain mechanisms.” (For those unfamiliar with servo-mechanisms simple examples include a thermostat and speed control, both of which keep a relatively constant temp or speed based upon deviation from a set point; also toilet tank water level if you think about it).
The point here is that life rather continually creates its own set points, a myriad of them actually, as it multiple processes flow together forward through time. It is not only its own set point, e.g., homeostasis, but also exerts some significant control of the world it moves through. Another quote: “What is important here is the ability of such causal configurations to influence their own conditions and to do so repeatedly and consistently.”
Finally Oyama’s proposal sees life as a developmental system moving forward through time (how else you ask? Me too). Any next moment of an organism’s life is not predermined; our nature is nurtured and nurtures itself anew moment by moment. This true in the short term, ontogeny (traditionally understood) and long term, phylogeny. Again her prose: “The only way out of the problem of predetermined potential . . . is thus to see potential itself, in the sense of possibilities for future alterations in a given structure, as having a developmental history. It is multiply, progressively determined, with new varieties of causes and consequences emerging at different hierarchical levels and with time.”
Yes, feedback processes are needed for control but these, again, must be seen in context (oh, how I love this). The autonomy of life is remarkable and the essence of feedback does not contradict this. Oyama goes back to Norbert Weiner’s early work Cybenetics(first published in 1948):
“Whether one is speaking of machines, organisms, or human affairs, control without feedback tends to become derailed and therefore useless or destructive. . . . Feedback has been described as control on the basis of actual, not expected performance[my bold] (Weiner, 1967). Its very definition is the ability to control by being controlled. It is, incidentally, the influence of results on the processes that produce them that I take to be central to feedback. The implication of an explicit setpoint, the mechanical counterpart of an expectation, while useful for understanding servomechanisms and, perhaps, for certain simulations, can be misleading in treatments of biological processes.”
So kudos to Duke’s most excellent scientists finding that their orthodox expectations were not met and I am sure they now seek new answers in the heterodox circle of the doxa. And I hope that they learn of and see the beauty of Oyama’s view of life as a developmental system. I hope everyone feels the deep aesthetic as life moves forward, its sails filled by its very own winds. Oh, is there another metaphor here? One that we can apprehend in our own experience? Ponder your own life’s journey, its cascade of complexity downfield into the future, and maybe a musical melody might animate your next wanderings.
Well, enough of this rhapsodizing, with a most excellent science book to finish I will now gladly travel on.