“Conversations for Action” – Perspectives on the Design of Cooperative Work


In 1987, Terry Winograd, Professor Emeritus, Computer ScienceStanford University wrote a compelling paper ‘A Language/Action Perspective on the Design of Cooperative Work’ which was published in Human-Computer Interaction 3:11 (1987-1988), 3-30.

In the article, Winograd describes what he and his fellow collaborator, Dr. Fernando Flores, called ‘Conversations for Action’ which he asserts form the central fabric of all cooperative work.  “Language is the primary dimension of human cooperative activity.”   Winograd concluded that a language-action perspective would play a major role in developing the field of ‘computer-supported cooperative work’.

He was right, but it has taken more than 25 years and several software attempts to realize his vision.  4Spires is the most recent company to design work management software based on this perspective.

Below is a synopsis of the Winograd article:

People act through language.  The language-action perspective focuses on the form; the meaning and use of language to get things done.  A ‘conversation for action’ follows a certain structure – one party (A) makes a request to another (B).  Each party interprets a future course of action that will satisfy the request.  B can accept (and thereby commit to an outcome), decline, or propose a counter-offer with alternative conditions.  Each of B’s ‘moves‘ then lead to different ‘moves‘ by A, and the conversation can be seen as a dance that eventually leads to a mutual understanding that the requested action has been done or that the conversation is complete without it having been done.

The perspective deals with the structure and coordinated sequence of acts by A and B rather than the actual doing of whatever is needed.  Conversations for action are the central coordinating structure for human organizations.

“We work together by making commitments so that we can successfully anticipate the actions of others and coordinate them with our own. The emphasis here is on language as an activity, not as the transmission of information or as the expression of thought.”

Winograd (and we) are concerned with designing computer systems that support these conversations for action.  Email is still the dominant electronic communication tool, though email does not provide sufficient structure to properly support taking cooperative action.  Email, for example, offers only one, generalized way to begin a conversation – ‘compose email’, and it does not offer any distinction between information sharing and making a request.

The system Winograd and Flores conceived allows for a user to initiate a ‘request’ form which prompts the user to specify a performer, others who will receive copies, a related domain of interest, and a description of the desired outcome and due date.

The recipient, on the other hand, is prompted with the various options for responding (e.g. Agree, Decline, or Counter-Offer) that are generated by a conversational state interpreter.  At each stage of the conversation the user is presented with a display of only those actions that could sensibly be taken next by the current speaker (i.e. A or B).  The program deals with the structure of the conversation, not the content.

“The system has no magic to coerce people to come through with what they promise, but it provides a straightforward structure in which they can review the status of their commitments, alter those commitments they are no longer in condition to fulfill, make new commitments to take care of breakdowns and opportunities appearing in their conversations, and generally be clear (with themselves and others) about the state of their work.”

Unlike email, the basic unit of work is a conversation, not a message.  Rather than just linking email messages by the use of Re: in headers, each message belongs to a conversation.  This key distinction enables a much more powerful retrieval and monitoring of work in progress.  To begin with, answers to basic questions like who has the ball, and what do I have to do next become readily apparent.  Messages can be retrieved based on status, or stage (e.g. open or closed), or role (e.g. performer or observer), or domain (e.g. goal or account), etc.

The system replaces typing parts of the contents of an open email message with more direct and structured interactions which are more efficient.  It is a generic tool in the sense that it is intended for a particular kind of communication (i.e. taking cooperative action) without regard for the topic or functional domain.  It is not built for arbitrary sequences of messages, but for the requests, promises and completions that are at the heart of coordinated work.

Systems designed to support conversations for action are not intended to replace face-to-face verbal interactions, or to lessen the importance of interpersonal relations.  Language acts, in general, can be less effective in the absence of personal relationships.  Much of business involves meetings and the social acts of persuasion, negotiation, and, at times, argument.  Trust is developed and built-up over time and is a key factor to productivity along with the mood and motivation of individuals.  Systems, however, can add substantial value by recording and tracking these agreements and work tasks.

Winograd’s article lays out a compelling case for a new generation of tools designed specifically to support conversations for action.  Email is as inadequate to this purpose today as it was 25 years ago when he wrote the article.  Like Winograd, we here at 4Spires believe the basic unit of cooperative work is a conversation that turns into a commitment to act.  Our solutions draw their differentiation from this work management perspective.

Leave a Reply