Content-type: text/html ~ Stephen's Web ~ Autonomy

Stephen Downes

Knowledge, Learning, Community

Oct 15, 2006

I think that people misunderstand what I mean by autonomy. It's like when I talk about learner-designed learning. People seem to assume I am talking about casting learners unaided into the sea to fend for themselves. As though they could never ask for advice. As though there would never be anyone willing to guide them or support them.

The same with autonomy. The presumption is that what I mean is a person who is an island, who does not depend in any way on others, who is ruggedly individualistic. Some sort of weird Ayn Rand fantasy of epistemological superhumans, a Nietzsche-inspired fantasy about people being able to completely determine, with no input from anyone or anything, what is true, what is right, what is good.

But that's not what I mean at all. Nothing close. That's why I have included openness and connectedness as additional criteria for epistemic goodness. That's why I talk about communities and networks at all. I do believe that the contributions of other people are important and essential. I am well aware how much external influences - yes, including media and advertising - can and should help determine our thoughts and beliefs. I would even draw you a picture depicting the causal relationships, how sensations effect neural states. Like this:

(Source of Diagram)

For one thing, maintaining an opposite point of view is irrational. Given what we know of human cognition, there are no belief states that are completely independent of our experiences. We are not born (contra Descartes and a whole school of misled Rationalists) with ideas burned into our brain, like some sort of mark of the Creator. What we come to believe is caused by what we experience. Our mental contents are reflections, perceptual echoes, the materials of our experiences playing back against each other, mixing and mashing and reforming.

In just the same way, contrasting autonomy with determinism is irrational. When I say that somebody's contribution to a network was 'not autonomous', I do not mean that they are under some sort of mind control, a robot at the whim of some Svengali. Yes, again, it is true that all mental states are caused by perceptions and experiences. But it does not follow (and should not be inferred) that all mental states are determined by these perceptions and experiences.

These sorts of extremes - complete independence, and complete dependence - are the result of what I might call a naive causal view of the world. This is the view (that all of us were taught as children) that the world operates like clockwork. That when you do something, there is a knowable and determinate effect. A causes B. And if there is a B, then there must be some determinate A that caused it. But the world isn't like that. Once events reach a certain level of complexity, the story about causation breaks down.

Consider, for example, a bolt of lightning. We have all (I presume) seen lightning, and know that it occurs during a thunderstorm. We are told that the cause of the lightning is the buildup of electrical charge in the thundercloud. The thundercloud, in turn, is caused by the buildup of water droplets in the air, condensation caused by the interaction of a warm and humid air mass with a cold front, this cold front in turn caused by the rotation of the Earth and the uneven heating of the Sun.

I remember once, one hot July night in Edmonton, returning home from the Power Plant, mad at the world and just wanting to get away, I saw the lightning flashing south of the city and jumped into my car to go chase it. A couple hours later I was out on the flat prairie, the lightning bolts shooting straight down, huge, towering, overwhelming bolts from the sky. I got out of my car and walked around the field, feeling the rain pelt against my face, watching the bolts streak down, one after the other, feeling so terrified by the storm I was at the same time one with it, part of it.

And I asked myself, had I been struck by lightning at that point, what would have been the cause of it? Would it have been the dismissive behaviour of those around me in the bar? Would it have been some irrational perception on my part? Would it have been my foolish walk around the field in a thunderstorm? Would it have been the buildup of an electrical charge in a cloud? Would it have been the uneven heating of the Earth by the Sun? What would have caused that bolt to have that impact at that time? And the answer is: nothing. That when we say this thing caused that thing we are placing an interpretation, based on some gross oversimplification, on the state of affairs.

There is no contradiction between saying that our thoughts and experiences are caused, and saying that we make choices. This becomes especially the case when we see that our choices in turn result in new thoughts and experiences. What we are is that entity (that amorphous assemblage of neural connections that, when thought of as a unit, can be seen as recognizing input and creating output) that recognizes certain states of affairs as states of affairs - as things, as causes, as Herman from next door.

So when I am talking about one thing being autonomous from others, I am not talking about the one thing being free from the causal influence of others, but rather, I am telling a story about how it is that the input of that one thing to the network as a whole is determined, and more accurately, how it should be seen as determined, how it should be regarded as determined, how - were we building a network of some sort - it should be enabled or permitted to be determined. When I say something is 'determined' or 'not determined' I am talking about, not some essential state of nature, where all things are one of These or one of Those, but rather, how we should consider that thing to be.

What was the cause of the lightning? If it was determined, then something made it strike at that time in that place. If it was undetermined, then the storm decided to hurl a lightning bolt at that time (neither wording really satisfies - and yet these are the words we have to work with, because our bias toward a naive causal view of the world is built into the language). What I want us to do, with respect to humans, is to take the attitude that the storm decided to hurl the lightning bolt. Not as an uncaused completely indeterminate event (because obviously it's not) but rather, seen this way, as a grounded, meaningful event (indeed, the source of meaning).

What does that mean in practice? It means that we ascribe to ourselves the possibility of choice (in fact, Gestalt alternatives, oscillating ways of seeing the world, the decision to perceive a duck rather than a rabbit), that this choice will be ascribed as the cause of our external actions, including especially our contributions to the network, in the sense that "When I say 'A' it is me that is saying 'A', and not some other person saying 'A' through me." In other words, we are saying that we see the origin of 'A' as being located inside ourselves rather than external to ourselves. It would be like saying that the cause of the lightning bolt is in the storm - it isn't some direct consequence of warm and cold air masses, and it wasn't in some sense 'drawn out' by some foolish person walking in a field tempting fate.

What this means in practice is that there ought not be an identifiable dependence (that is, an explainable correlation) between what someone else says or does, and what you say or do.

Think of it as akin to the distinction between being told to do something, and having someone suggest that you do something. These two circumstances may be perceptually indistinct. In each case, a person leans over to you and says, say, "You should vote no." And then you utter the words, "I vote no." The difference between the two states is one of interpretation, one where we decide as observers or as participants to apply one frame to it, as opposed to another. The difference between thinking to ourselves, on hearing the words, 'I have no choice' as opposed to 'I have a choice'.

In order for it to be possible for a person to rationally say that 'I have a choice' there must, in fact, be a choice. It much be possible for the person to have uttered some statement other than the one that was suggested. This implies, first, that some sort of consideration or processing of the suggestion occurs, and second, that as part of that consideration, alternative actions emerge as genuine possibilities. So that you could, as a rational person, see two possible and acceptable states of affairs, one where you said 'I vote no' and one where you said 'I vote yes' (and even one where you decline to vote at all).

What would prevent you from having that choice? First, your input might be in some way circumvented. For example, when somebody purports to express your vote for you, but substitutes their own point of view for yours. Second, your input might be coerced. For example, when the consequences of uttering 'I vote yes' are so horrible that it cannot be considered as a viable alternative. Third, you might fail to consider or process the request. For example, you way respond automatically because you have been conditioned or hypnotized in some way.

Now again, it is important to keep in mind, what these scenarios describe are ways of seeing a situation, as opposed to three ontologically distinct types of entities. This is not some sort of taxonomy that I am offering (I don't offer taxonomies). These are three vectors you can consider to be more or less the case such that, when the preponderance of the interpretation is in one direction, the choice was non-autonomous, and when the preponderance of the interpretation is in the other direction, we say the choice was autonomous.

And these vectors are very much matters of point of view. To take the most obvious case, what constitutes 'so horrible that it cannot be considered as a viable alternative'? This clearly will vary depending on the person's point of view. Some people may be prepared to tolerate anything but death or dismemberment. Others would not fear the same being done to themselves, but will fold at the thought of it happening to a loved one. Others would not consider expulsion or exclusion by a group to be tolerable. Being singled out as the lone dissenter might be unbearable for some. This circumstance - what counts as too horrible - is a matter of interpretation.

So when a person, acting as a node in a network, wishes to participate autonomously in a network, what this means is that this person would prefer that, on the whole, (a) their utterances be expressed to other members of the network accurately, (b) that there not be sanctions or punishments for making certain utterances, and (c) they be afforded the time and the capacity to consider matters in their own light before making an utterance.

So when a person, building or designing a network, wishes the participants to participate autonomously, what this means is that they would tend to (a) ensure each member's voice is communicated accurately and completely, (b) create a space or mechanism for that person such that they are shielded from sanctions or retributions, and (c) ensure they are presented with information in a timely manner and given the tools (including the education and the background knowledge necessary) to make informed decisions.

These considerations explain why I tend to disfavour small groups. See also Konrad Glogowski, 'To Ungroup a Class'. Small groups tend to fail on all three counts. First, when the decision of a group is reported, the view expressed is often the reporter's (and there are no mechanisms in place to prevent that). Second, for some people (namely, me) small groups create greater pressure to conform (especially when the group is given a task to perform or an outcome to produce). And third, the process is often constructed in such a way as to prevent consideration of the matter at hand - wither there is no time to present such considerations, or the considerations are overwhelmed by group members who have not taken the time to consider.

I haven't talked here about why autonomy is necessary in well-functioning networks. The long story is probably the subject for another day. But in a nutshell, the response is this: better decisions are made when more perspectives and more variables are taken into account. Each person in a network brings new perspectives and variables to the table. This, increasing the number of people in the network improves the functioning of the network. If, however, their participation is not autonomous, then the impact of those perspectives and variables are never brought into play. They are overridden by whatever entity is creating the non-autonomous behaviour. This weakens the network, because of the missing perspective, and worse, it disguises this weakening because the individual entity may be perceived as autonomous, even when not.

Stephen Downes Stephen Downes, Casselman, Canada

Copyright 2024
Last Updated: Jul 16, 2024 08:07 a.m.

Canadian Flag Creative Commons License.