The Death of Context

I think part of the disagreement Ricky and I have is that we’re coming at context from two different perspectives. Ricky seems to be coming at context from a purely computer science perspective where I’m taking a more social science approach. Negotiating and translating between those two domains is difficult. (My position on this is heavily influenced by Paul Dourish.)

Ricky says context has been defined to death. I say, let’s kill it. He says:

The last thing we need is more papers purporting to confer a better understanding of context unto the computer science community.

I think he’s right, in part. There are enough positivist views of context, and what it is, already. John McCarthy’s modeling of context in formal logic is possibly the apogee of the positivist approach that defines context as something that is able to be measured, captured and delineated. This view of context makes it separate from activity. In Ricky’s example of the dude standing in the road or the football field making gestures seems to be the idea that activity is what happens in a context.

Ricky’s position is right, in so far as it exists in the engineering tradition which is basically positivist in nature.

I come at context from the HCI perspective which is heavily influenced by the phenomenological approach that asserts that what we experience influences our view of the world. This is not an objective view of context but a subjective one.

The view of context in the phenomenological perspective is that context is something that emerges through interaction, not that it is something that can be represented, delineated and separated from action.

Ricky gives this example:

Or, the next time you see a person standing in the middle of an intersection wearing a blue uniform and waving their hands about in various ways, observe the way in which you and other motorists respond. It is unlikely that any motorist would make a call to the nearest hospital for the mentally ill. But if you saw a similarly dressed person making those gestures in the middle of a football field, that phone call might be made.

In the example, a person wearing a blue uniform, waving their hands is of less concern in the middle of the street than in the middle of a football field. I, for one, would be more concerned with someone practicing Tai Chi in the street!

The ordinariness of a policeman in the street and the ordinariness of the Tai Chi practitioner are contrasted with the decidedly unordinariness of the Tai Chi practitioner in the street or the policeman doing Tai Chi in full uniform in a field. They key here, that I think Rick has touched on in the policeman, is that of context is not only an issue of location or other tangible measures but action within context and the meaning that arises from specific actions in particular contexts. A shorter way of saying that is to call the link between meaning and action practice.

This is not to say that context is not important, because it is for “understanding, contextualising and disambiguating forms of activity and information” (Dourish). By looking at practice it is possible to understand context(s) and use them to disambiguate actions. Context arises from activity and activities continually generate new contexts.

Ricky is also right when he says, when talking about what context is for: “all you have to do is observe the way in which a typical conversation proceeds between two people and see how they fill in the blanks in the absence of explicit information.”

However, I find this, again, to be related to my point that context is mutually constituted and emerges from the activity at hand. Going even further with the conversation example, context is “sustained and managed” (Dourish) in the course of a conversation, it is not a thing that a conversation takes place within but is as relevant to the conversation as the content, and as potentially dynamic. This is exactly what the quote from the IFTF article says:

First, “context” isn’t something that programmers or designers can either thoroughly describe or comprehensively predict; it’s something that emerges in the moment.

Second, context can change very rapidly. Making sense of context cannot merely be a matter of computers getting faster; it requires helping people, who are already better-equipped than any computer to figure out context.

Ricky concludes his post by saying:

What is needed are solutions to the problems of modeling, incomplete information, filtering, scaling and the derivation of context information from existing context information. Basically we need to figure out how to remove the current limiting factors so that it can be applied more generally and with predictable results that make sense to the user.

which is again the positivist view. I say, what is needed are solutions to the problems of supporting people in their emergent practices at work and at play. We need to appreciate how people create (new) meanings for the technology they use and figure out how to support people in appropriating technology into their practice. Context, in this view, is not the thing to focus on but is part of the emerging mutable practice.