Conversation Theory Literature Review
What is conversation? How can we design for effective conversation?, Hugh Dubberly and Paul Pangaro
In What is conversation? How can we design for effective conversation?, Dubberly and Pangaro explain what conversation is and what it is not, the process of conversation, what conversation “offers”, and the limits of conversation.
Early in the article, it is said that “we talk all the time, but we’re usually not aware of when conversation works, when it doesn’t, and how to improve it. Few of us have robust models of conversation.” This is an idea I have played with for awhile, not only in regards to conversation, but also in regards to our general actions. We have very little understanding of why we do things. For instance, if I ask an individual why they did something 5 minutes ago, they will most likely struggle to provide a complete justification for why they took that action. Just how to provide a user with the tools to better understand why they are feeling a certain way or why a conversation has completely changed course (those situations where you wonder how did we get here and try to replay the conversation in your hand, but are not entirely sure what triggered the unexpected trajectory) is something I will most certainly need to address this year, especially in the context of a romantic relationship, where it is that much that easier for logic to be pushed aside (when you question why we are in this argument).
When thinking about intervening within the context of a romantic relationship, I will never be able to directly touch every conversation a couple has. What I can do is create tools that allow an individual within a relationship or both individuals within a relationship to have a conversation with a system that can help facilitate improved conversations both in real time and in the future. Dubberly and Pangaro note that “only in conversation can we learn new concepts, share and evolve knowledge, and confirm agreement”, when looking at that statement, I specifically focus on the words “new” and “evolve”, I hope this is where some artificially intelligent system can come in handy. This is not scientifically proven in any way, but I have suspicion this is the precise reason why some relationships are plateaued. Couples aren’t having new conversations, they aren’t acquiring the needed information for their relationship to evolve, and instead in this continuous cycle of struggle. Just like a relationship counselor, I hope that the system I end up designing creates an environment for that couple to “learn new concepts, share and evolve knowledge, and confirm agreement.”
Dubberly and Pangaro then go into what exactly constitutes a conversation, by listing the following steps:
- Opening a Channel
When one individual sends another individual “an initial message”. This doesn’t necessarily signal a conversation has started, but that there is a possibility for a conversation. In order for a “conversation to follow, the message must establish common ground; it must be comprehensible to participant B.” - Commitment to Engage
When the individual receiving the initial message, “commit[s] to engag[e]” with the individual who sent the initial message. It is also important to note that “for conversation to persist, the commitment must be symmetrical, and either side may break off for any reason, at any time. Put another way, each participant must see value in continuing the conversation, which offsets the personal cost of being engaged: what we call the “bio-cost,” or the energy, time, attention, and stress required.” - Construct Meaning
“Conversation enables us to construct (or reconstruct) meaning, including meaning that is new to the” individuals within the conversation. In order for that that conversation to “construct meaning”, “messages are composed with topics or distinctions that are already shared, on the basis of prior conversation or shared contexts, such as common language and social norms.” Dubberly and Pangaro introduce the term “prescriptive dynamics”, as this “kind of ‘glue’ that explains just how these topics interact to make up the new concept”, this is exactly the point where I imagine a machine can assist a couple when attempting to have a productive conversation, if that machine is able to provide that “glue” in a way the two participants are unable to. - Evolution
When both individuals in a conversation “are different after the interaction.” It is important to not that Dubberly and Pangaro note “an ‘effective conversation’ as an interaction in which the changes brought about by conversation have lasting value to the participants.” - Converging on Agreement
In some situations, one individual may attempt to confirm their understanding of the other individual’s understanding of what was discussed. In order to do that, the individual trying to confirm their understanding, must communicate their understanding of their “model of the concept”. When that understanding is conveyed, the other participant attempts to understand what was just said and further messages may be sent until both participants “judge that the[ir] concepts match sufficiently” or “they reached an ‘agreement over an understanding’”. - Action or Transaction
If successful in converging on an agreement, “sometimes one or more of the participants agrees to perform an action as a result of, and beyond, the conversation that has taken place.”
Dubberly and Pangaro also note that conversation allows for learning, coordination, and collaboration. I want to pay particular attention to what they say about conversations ability to facilitate collaboration, specifically that “coordination of action assumes relatively clear goals, but many times social interaction involves the negotiation of goals.” Just how to accomplish within the context of a socio-technical system I am not sure. What exactly are people are willing to establish within such a system? What infrastructure needs to be in place for people to people to establish “clear goals.”
Dubberly and Pangaro also discuss what is needed for conversation and mention “when assembling a design team we ought to ask, What expertise and what collaborative style(s) do we need? What variety is required to succeed?” While not a direct comparison, it begs the question how do two people in a committed relationship have the required variety within their conversations? They also note that “we generate new ideas by combining old topics in new ways,” can we expect couples to be able to do this without outside influence?
Dubberly and Pangaro end the article with a set of questions for designers. A couple ones that are particularly relevant to what I plan to do are:
- “What channel is being opened to begin the conversation? Is the interruption reasonable in how and when it intrudes? What is the bio-cost of the intrusion relative to its benefit? Are there better ways to interrupt?”
- “Once accepted, does the ongoing exchange convey the potential benefits in continuing the engagement? Is there learning or delight? Is curiosity or interest stimulated? At what bio-cost? How can it be improved?”
- “Is meaning easily understood; that is, do the messages speak to the participants’ context, needs, interests, values, and in their language? How difficult is it for users to “put together”? How can messages be made more efficient or clear or entertaining, as appropriate?” Especially important to think about these questions in the context of machine learning and natural language processing.
- “How can users convey intention and meaning to the software? Are those means sufficiently expressive or easy or delightful? Where do they fall short?”
- “How do physical age and technology exposure change predilections for media, modes of collaboration, and personal values?”
Cybernetics and Conversation, Paul Pangaro
In Cybernetics and Conversation, Pangaro hits on lots of topics he and Dubberly discussed in What is conversation? How can we design for effective conversation? But he does hit on a point, that is particularly interesting to think about in regards to conversation. He notes that “if instead of observing a relationship of objects in our environment, we take a position of observing ourselves in conversation with others, a similar phenomenon occurs: the participants in the conversation are defined by the looping-around. Our features, feelings, opinions, boundaries, differences are computed by the interaction. Thus we find ourselves being constructed (defined, identified, distinguished) by that conversation. From this point-of-view, our selves emerge as a consequence of conversation. Expressed more fully, conversation and identity arise together.” If I was able to create an environment where an individual could observe how their relationship was being constructed (what was causing their relationship to be defined a certain way, individuals especially those within a relationship could only benefit.
Architecture of Conversation Theory, Paul Pangaro
In the Architecture of Conversation Theory, uses “Pask’s Conversation Theory (CT), especially as presented in “An Approach to Machine Intelligence” (reprinted in Soft Architecture Machines, edited by Negroponte, MIT Press, 1975)” to derive a modeling technique that is a formalization “for describing the architecture of interactions or conversations.” Pangaro’s examples are pulled from an “example of modeling an organization or company”, but can be applied to a more typical conversation or conversation between two partners.
One particular highlight to pull is the existence of a “comparator” or “the specific mechanism whereby the feedback information is used by comparing the actual result to the desired result, or original goal.” I’m trying to think about what exactly a “comparator” is in regards to a couple’s communication. Potentially it could be how their parents communicated or a couple they know that they really admire, just how to bring that to light will be key in being able to create an artifact that the individuals interacting see as beneficial to them and their relationship.
Modeling Engagement — Designing Conversation and Managing Complexity, Paul Pangaro
In Modeling Engagement, Paul Pangaro provides a variety: one for successful conversations, one for model for client engagement, and another for designing engagement conversation. I wanted to pull specifically one thing from this deck, the point on “selecting participants is part of a larger planning function that enables the right external information to enter the next conversation.” On the same page it is noted that “just as participants feed the next conversation by adding needed variety, information feeds it by adding variety in the form of qualitative and quantitative statements and context that are considered necessary to achieve the goals of the next conversation.” This makes me think of two things:
- A system can provide the needed variety in terms of information, as long as it is feed sufficient information to provide that variety.
- Is there potential for such a system to have multiple personalities? For instance, could two personalities be watching over a conversation and provide different opinions? What if a new personality was introduced randomly to an interaction, would the participants in that conversation welcome that new personality?
Cybernetics: Book of Models, Paul Pangaro
One model that I found particularly interesting is W. Ross Ashby’s model of Conversation.

The explicit differentiation between the upper “immaterial aspects” and lower “physical world” show that “actions take place in the physical world, while goals do not. Goals, the province of cybernetics, are the ‘immaterial aspects’ of interaction [W. Ross Ashby].” It is also noted that “the dotted lines indicate that recursions via conversations are as if we are interacting directly at the level of goals, while in practice we are not.” This brought up a couple of questions for me. Could an intelligent system make it possible for humanity to interact “directly at the level of goals?” Could that intelligent system make those goals more visible so that people may not directly-directly interact at the level of goals, but work in a much more direct manner of goals.

