Distributed Cognition Lecture Summary

Date: Week 11, 11th March 2009.

Today's lecture covers the topic of Distributed Cognition. Distributed Cognition builds on the tradition of Cognitive Science, which aimed to explain human cognition as computation. Distributed Cognition is a theory that looks at how cognition occurs not just in our heads, but also in the world. It applies the computational view of cognitive science to processes in the world. It differs from cognitive science in the following two respects:

Introduction

The topic of today's class, distributed cognition comes from the same line of theories and views of the world as the theories of mental models that we covered in the second week's class. Think back to the discussion we had about the idea of mental models from that lecture. Remember that the idea of mental models is that people hold in their mind mental representations of the world that they can use to reason about the world and guide their actions.

If we re-draw the illustration of mental models from a distributed cognition perspective, we might add in lots of resources around it. If you look at the picture above, ask yourself some questions such as; where is this happening, in an office? Who else is in that space, does the person have any colleagues? What other resources do they help to support them in their work? Desks, callenders, pin-boards, telephones?

The Goal of Distributed Cognition

The goal of Distributed Cognition is to identify and explain the extra tools, resources, and social relations that people draw on to carry out their work. This follows from the following observation:

Work is more than the activity of a single individual working alone and without tools.

Distinguishing Features

We can summarize three distinguishing features for distributed cognition.

Boundaries of the unit of analysis for cognition
- The unit of analysis is expanded from just being what goes on inside a person's head to include the whole system that achieves computation.
Range of mechanisms assumed to constitute cognitive processes
- It expands the range of mechanisms assumed to constitute cognitive processes from mental processes to physical and social interactions.
Extension of cognitive science models beyond the individual.

Task: Sorting a deck of cards

As a practical exercise, we tried a very simple exercise of sorting a deck of cards. First, we tried this by dividing the whole deck of cards amongst the class so that each student got 3 or 4 cards. Then students had the task to figure out how to coordinate among themselves to arrange the cards back into a sorted deck.

After this first round, we discussed some observations about how the students achieved the task. then we divided the class into two groups. One group took all the red cards and the other group took all the black cards. They took a few minutes to discuss and decide what their strategy for sorting the cards before beginning. The object was to be as quick as possible.

Functional System

The system is constrained by the set of resources appropriated into the resolution of the problem.
The system, not the individuals performs the task.

Three kinds of distributed

Researchers working from a Distributed Cognition perspective have highlighted the role of the tools that people use, their social organization and their cultural context in cognition. Distributed cognition researchers look for the ways that cognition is distributed within a computational system. They identify three ways that it can be distributed:

Between internal and external representations.
Across the members of a social group.
Through time, such that the results of earlier events can transform the nature of later events.

Cognitive Artefacts

Reduce memory load
Simplify cognitive effort
Trace changes
Transform the task into a different one.

Embodied Cognition

Minds are not passive representational engines, whose primary function is to create internal models of the external world. The relations between internal processes and external ones are far more complex, involving coordination at many different time scales

This is between internal resources

memory
attention
executive function

And external resources

objects
artifacts
at-hand materials

Socially Distributed Cognition

Culture and Cognition

The study of cognition is not separate from the study of culture.
Culture is not 'contextual knowledge', but a process that accumulates partial solutions to recurring problems.
Compare to the traditional Cognitive Science view, where culture is conceptualised as contextual knowledge on which individual cognition operates. In distributed cognition, culture shapes the cognitive processes of systems that transcend the boundaries of individuals.

Critical Questions

Distributed Cognition makes no distinction between people and artefacts. Both are treated as 'media' that hold and transform representations. Therefore, as Nardi remarks, "Messy cognitive activities conducted every day by ordinary humans, such as interpretation and imagination, are difficult to consider within such a framework." (Nardi 2002, p.273)
Is computation an appropriate metaphor for understanding all people's activities

Practical Task

Your task for this week is to analyse the activities that occur in the Library in the terms of Distributed Cognition. In groups of two or three, first identify a 'cognitive process' in the library. Observe people engaged in this process and attempt to map it out. See if you can:

Map 'information flows'
Find out how information is represented and transformed.
Find 'cognitive artefacts' or other concepts from this week's lecture and readings.

You should prepare a short (5-10 minutes) presentation of the results of your interview for next week's class.

Discussion

In the following week, after students had had time to complete the practical task, we held a discussion on the topic of Distributed Cognition.

We began by going back over some key concepts from the lecture. These were; cognitive artefacts, embodied cognition and socially distributed cognition. Students worked in groups to go back over these concepts and come up with their own explanations for each. We then took a round as a class where we wrote these explanations up on the board and discussed them together.

Following this, two groups presented the results of their practical exercise. The first group presented an analysis of students sitting studying in the shared seating area in the balcony above the library. They looked at how the students arranged themselves relative to one-another and their notes, bags and other artefacts. An interesting discussion emerged in response to this around the question of whether it is social cognition whenever two people are gathered together.

The second group had analysed the process of finding a book in the library. They used photographs to show how this activity unfolded. We observed that artefacts such as the bell at the library counter serve to make it easier to request assistance and that signs in the library are used to mark different areas.

Readings

Hollan, J., Hutchins, E., and Kirsh, D. Distributed Cognition: Toward a New Foundation for Human-Computer Interaction", ACM Transactions on Human-Computer Interaction, Vol. 7, No. 2, June 2000, pp 174-196.
Rogers, Y., "A Brief Introduction to Distributed Cognition", Discussion Paper Interact Lab, School of Cognitive and Computing Sciences, University of Sussex, 1997

References

Nardi, B., 2002. Coda and Response to Christine Halverson. Computer Supported Cooperative Work (CSCW), 11(1), p.269-275.

Web Resources

UCSD DCOG and HCI Laboratory: This lab is mentioned in the 'Hollan' reading. Home to prominent researchers in the area and has links to some publications.
Critique of Distributed Cognition: written by a student as part of a course similar to this one.
Wikipedia article: Overview article with links to related theories.