designing UX conflict prevention solutions in the emerging “conversational commerce”


In early 2015, Chris Messina, Developer Experience Lead at Uber (and, as Business Insider noted: “the man credited with inventing the now ubiquitous hashtag”), posted a short article on Medium in an effort to highlight what he saw as a growing trend in the communications and messaging space — the idea of a “conversational commerce.”

The original article cited several examples: “From Path’s pivot to emphasize Talk, which enables users to message local businesses, to Facebook’s $22B purchase of WhatsApp, to Fetch’s texting shopping assistants or Fancy Hand’s on-demand assistants, to the recent real-time translation innovations in Google Translate, to Ethan’s army of textable Ethans, to Rise, Lark, and Better in the healthcare space.” The article also cited the rise of voice assistants, highlighting offerings by tech giants Apple (Siri), Google (Now), and Microsoft (Cortana).

A year later, Messina released a follow-up article, branding 2016 as the year that conversational commerce became the dominant trend of consumer computing apps. Messina highlights Uber’s integration into Facebook Messenger and data from Business Insider showing that messaging apps have eclipsed social networks in monthly actives. Further, he notes a big move by WhatsApp to engage in the trend:

“Starting this year, we will test tools that allow you to use WhatsApp to communicate with businesses and organizations that you want to hear from. That could mean communicating with your bank about whether a recent transaction was fraudulent, or with an airline about a delayed flight. We all get these messages elsewhere today — through text messages and phone calls — so we want to test new tools to make this easier to do on WhatsApp, while still giving you an experience without third-party ads and spam.”

Messina’s second article contains several more great resources that you can (and should!) check out, but for the purpose of this paper, I’m going to pause the rambling here and say that I agree with him. I’m also going to steal a definition of sorts from him:

“I want to clarify that conversational commerce (as I see it) largely pertains to utilizing chat, messaging, or other natural language interfaces (i.e. voice) to interact with people, brands, or services and bots that heretofore have had no real place in the bidirectional, asynchronous messaging context. The net result is that you and I will be talking to brands and companies over Facebook Messenger, WhatsApp, Telegram, Slack, and elsewhere before year’s end, and will find it normal.”

Assuming we buy into this idea (I do), Messina has effectively identified a new frontier for designers. New mediums mean new visual, structural, ethical, and historical considerations (among many others). Essentially, this new medium involves taking an old form (conversation) and constraining it to a commercial context. When a new constraint is placed on an understood medium, we have the opportunity to mold the syntax of that medium in new ways. This is a chance to get into the rough of it, to develop and test new theory work — and designers should be thrilled about it.

Messina has done his part to aggregate any developers and designers attempting to innovate solutions for a conversational commerce. He highlights Jonathan Libov’s ideas on text, Slack’s standardization of Slash Commands, and Peach’s Magic Words as efforts to “provide the most utility to users with the least effort and the least complexity.” I think these two attributes, effort and complexity, will continue to be of the utmost importance — particularly as the trend of conversational commerce permeates into excessively convoluted fields like medicine and law — industries ostensibly designed to help the members of society they serve (I recognize this can be interpreted as naive thinking, but here’s hoping).

Keeping all of this in mind, conversation is very familiar to us on a base level. I do not think it’s a gigantic leap to assume that many of the conversations that exist in customer service settings are generally centered on a conflict between customer and company (e.g. I don’t call Comcast representatives to ask how their day is going). This paper, then, is an attempt at ideating how, within our new medium, institutional interfaces might facilitate a mutual understanding between the two engaged parties that lowers effort and complexity on the part of the customer, while allowing the institution to protect its right to operate efficiently.

What follows is a bit of (semi-dense) language theory and a design summary of how an institutional interface in this medium might not only enable, but educate — and that there’s a graceful way to do so. Specific suggestions are centered on an annotation mechanism, a solution which seems to have some industry weight noting the modest funding of Genius (and its extension, Bookmarklet) and VC investors’ (see: Marc Andreessen) continued interest in the notion itself.


In any institution optimizing itself for a certain social purpose, language acts as a natural self-defense mechanism against ideological change of the institution’s current purpose by outside actors. It does so by redirecting conflicts that exist between institutional-laymen (also referred to as outside actors) and the institution in such a way that a discrepancy between lay and institution ideologies makes laymen more likely to blame the particular member of the institution they are interacting with rather than the rules that govern and constitute the institution. This latent function of language creates inefficiencies of communication at the agent level that promote redirection of the layman’s criticisms from institutional practice to individual-within-institution. This function occurs when the institution reinforces in practice the lexical choice and semantics of individual agents within the institution. These institution-generated gaps in pragmatic competence function by shifting the context of the conflict and often serve to reinforce rigid institutions that become closed off to change that would benefit their audience. This paper defines a specific type of conflict and how this type of conflict might serve as an example of this phenomena. Further, this paper proposes a program for ethical conflict resolution that functionally assists institutions in communicating with laymen such that laymen are empowered to suggest and elicit policy change on the institution, rather than have their conflicts redirected.

institutions as social networks

In general, social networks are self-organizing, emergent, and complex, such that a globally coherent pattern appears from the local interaction of the elements that make up the system. These patterns become more apparent as network size increases. Practical limitations of computing power, ethics and participant recruitment and payment currently limit the scope of large-scale social network analysis. The nuances of a local system may be lost in a large network analysis, hence the quality of information may be more important than its scale for understanding network properties.¹ Thus, social networks are analyzed at the scale relevant to the researcher’s theoretical question. Although levels of analysis are not necessarily mutually exclusive, there are three general levels into which networks may fall: micro-level, meso-level, and macro-level.²

At the micro-level, social network research typically begins with an individual, snowballing as social relationships are traced, or may begin with a small group of individuals in a particular social context.³ Meso-level networks are low density and may exhibit causal processes distinct from interpersonal micro-level networks. In general, meso-level theories begin with a population size that falls between the micro- and macro-levels. However, meso-level may also refer to analyses that are specifically designed to reveal connections between micro- and macro-levels.⁴ Finally, macro-level analyses generally trace the outcomes of interactions, such as economic or other resource transfer interactions over a large population.⁵

For the purposes of this paper, when referring to a “social network (and by extension, an ‘institution’),” I am referring to a meso-level network, more specifically: a formal organization that exists as a social group with collective goals. Network research on organizations has traditionally focused on either intra-organizational or inter-organizational ties in terms of formal or informal relationships.⁶ Intra-organizational networks themselves often contain multiple levels of analysis, especially in larger organizations with multiple branches, franchises or semi-autonomous departments. In these cases, research is often conducted at a workgroup level and organization level, focusing on the interplay between the two structures. This paper will focus on how meso-level social networks interact with actors that exist outside of said networks and how these social networks use language, consciously or unconsciously, to protect their collective goals and strategies from nonaffiliated actor-initiated change.⁷

Cooley denotes two types of social groups, primary and secondary. He categorizes primary groups as small social groups whose members share personal and lasting relationships. People joined in primary relationships spend a great deal of time together, engage in a wide range of activities, and feel that they know one another well. In short, they show real concern for one another. In contrast to primary groups, secondary groups are large groups involving formal and institutional relationships. Secondary relationships involve weak emotional ties and little personal knowledge of one another.⁸ This paper will focus on meso-level social networks involving secondary social groups.

A great deal of past work on internally-sparked dispersal and transformation of social networks exists. Results dictate that weakening of the common purpose once a group is well established can be attributed to: adding new members; unsettled conflicts of identities (i.e., territorial problems in individuals); weakening of a settled dominance-order; and weakening or failure of the leader to tend to the group.⁹ There are numerous reasons for stable groups to “malfunction” or to disperse, but essentially this is because of loss of compliance with one or more elements of the definition of group provided by Sherif.¹⁰ The two most common causes of a malfunctioning group are the addition of too many individuals, and the failure of the leader to enforce a common purpose, though malfunctions may occur due to a failure of any of the other elements (i.e., confusions status or of norms).¹¹

At this point it’s important to introduce and explain the term “optimizing” used thus far. This term is important in defining the specific type of conflict this paper is interested in (which is done at length in a latter section). Taking the standard definition of “to optimize”: “to make the best or most effective use of (a situation, opportunity, or resource)¹²”, we can say that to optimize a social network is to elevate its effectiveness in accomplishing the common purpose around which the network has formed to achieve. Because social networks are organic structures and as such are constantly in flux, we can say this process of optimization is constantly ongoing and exists as the anti-function of network malfunction and dispersal. Therefore, we can come to the conclusion that any malfunction (and the internal reasons for that malfunction) within a social network is a result of a failed attempt by the network at further optimization. Each specific act in this process produces subtle transformations in the core purpose of the network, and from this notion we can reach a final important conclusion about social networks: The existence of a network is not predicated by a specific purpose, it only need possess a purpose. What that purpose is only matters inasmuch as actors in the network continue to rally around it.

From this vantage, this paper will consider how conflicts between a social network and an individual agent who exists outside this network (but nevertheless is compelled to interact with it) might be leveraged to optimize the network by shifting its social purpose toward a direction that benefits both network and outside actor.

on conflict

So what does conflict accomplish? Research that centers on discourse has helped prompt scholars to view conflict as a crucial part of the social construction of reality. Studies of disputes among children and adolescents suggest that conflict provides a central facet of the socialization process. Eder notes that resolution is not even a relevant goal in the case of conflicts between boys.¹³ Schiffrin suggests that arguing can serve as an important means of conveying sociability among adults.¹⁴ “Resolution” may indeed be a relatively rare outcome; among the family disputes examined by Vuchinich, only 26 percent ended in submissions or compromises, whereas 66 percent resulted in standoffs.¹⁵ Social theorists similarly have come to view social and cultural forms as historically contingent products that emerge through conflict, and the process of differentiating an Other from a Self is often deemed to be more crucial to the creation of both identities and communities than the presence of a shared cognitive substratum. Many scholars would thus agree with Simmel’s observation that conflict provides a central force for the constitution of social relations.¹⁶

Myers and Brenneis have pointed out that simply gathering people together to talk can signal the creation or re-creation of community and political order.¹⁷ Attempting to construct shared interpretations of past, present, and future events — or contesting attempts to impose unaimity — can shape social memory and social organization as well as impose limits of perception. Comaroff and Roberts have similarly argued that “It is in the context of confrontation — when persons negotiate their social universe and enter discourse about it — that the character of that system is revealed.”¹⁸ Each of these authors argues that discourse goes beyond questions of individual interest; it rather provides a fundamental basis for the creation of social community and the social construction of reality. Bourdieu seems to build on this, suggesting the existence of relatively fixed and homogeneous relationships between language and social relations.¹⁹

That said, uncritical adoption of these insights leads to a tendency of identifying all cultural forms as instances of power relations and to see power as being everywhere the same; historical and cultural specificities often fall victim to a universalistic and functionalist deus ex machina. Further, the idea that having power = gaining more benefit ignores a great many instances in which an individual at a power discrepancy outside of a Network A stands to gain benefit in a Network B by entering into conflict with Network A.

W.I. Thomas emphasized the importance of definitions and meanings in social behavior and its consequences. He suggested that humans respond to their definition of a situation rather than to the objective situation itself. Hence Thomas noted that situations that we define as real become real in their consequences.²⁰ All conflict essentially consists of multiple beings each trying to impose their specific subjective interpretation of an experience on the objective experience itself and on whatever the consequences of the experience are (this can be thought of as the objective experience’s historical relevance). A non-resolution of conflict leaves two parties with existing and separate subjective interpretations. When many similar conflicts surrounding a certain issue fail to be resolved, the basis of a hegemonic social hierarchy begins to form. A resolution of conflict implies that one party has conceded their subjective interpretation to another or that an amalgamation of the presented subjective interpretations has been agreed upon. The latter outcome is more advantageous to society in this case, as it’s most often a closer representation of the objective experience.

In this way, objective reality can be thought of as a limit, and each conflict can be thought of as an opportunity for disputing parties to reach a resolution that best approaches that limit, or not — but a reality without conflict is a reality in which one subjective interpretation of existence has superseded not just all other subjective interpretations, but objective reality itself, at all levels of society. This notion is a dangerous one because it steals from us our ability to protect our own self-interests, or to even have them in the first place.

From this, we can deduce that the true power of conflict is that it allows the economy of subjective experience to persist. At the borders of all social networks, from class to institution, exist boundaries of subjectivity. Engaging in conflict with these boundaries is beneficial to both the individual outside the network and those individuals that exist within it, despite being potentially detrimental to the immediate contextual purpose of the network. This is of little consequence, as optimization of a network implies the notion of a constantly shifting purpose, which, if not in revolutionary terms, at the very least occurs incrementally.

Conflict, then, is perhaps best viewed on an instance-to-instance basis as an attempt by an individual to optimize some system that they are interacting with within a specific contextual scenario. Attempting to give each type of conflict context means attempting to give each conflict consistency at a ‘local’ level that doesn’t necessarily exist at a ‘global’ level. This contextual view of conflict protects us from making unfounded universal claims about power and its relation to conflict while still allowing us to recognize and analyze patterns that occur within these contexts.

What follows is a description of a specific type of conflict given as example and a look at how language within this specific type of conflict is currently prohibiting actors from outside of a network to engage with said network in a productive way.

a specific type of conflict

As previously stated, we can think of social networks with a social purpose as institutions, that is, institutions transcend the individuals that comprise them and the intentions of these individuals by mediating the rules that govern them to produce valued, recurring patterns of behavior. The specific term “institution” commonly applies to a custom or behavior pattern important to a society, and to particular formal organizations of the government and public services.

We can distinguish between different models of institution by assuming two caveats provided by Luhmann: 1) the institution as a whole is operatively closed on the basis of communication and 2) that the function institutions emerging within society conform to, and embody, the principle of operative closure and, therefore, will exhibit comparable structures despite factual differences between them.²¹ Comparisons between models of institutions derive force when we recognize that the compared realms differ in all other respects; we can then highlight what is comparable and charge it with special significance. The main differentiator this paper will be using is social purpose, as this purpose delineates the rules that mediate and therefore constitute the institutions with which we’re interested.

Rules, then, become exceedingly important. It is a commonplace within philosophy that institutions can be understood in terms of constitutive rules. Proposals for understanding particular institutional phenomena or institutions generally in terms of such rules have been made by, among others, Rawls (1955), Midgley (1959), Searle (1969, 1995), and Goldman (1970).²² Searle’s account of constitutive rules is perhaps the most developed one, and it remains the main point of reference in the contemporary literature on the subject. Over the years, however, it has attracted severe criticism.²³ It has been argued, inter alia, that constitutive rules do not capture the essence of institutional phenomena, that all rules are constitutive, and that the difference between constitutive rules and regulative rules, to which Searle contrasts them, is merely linguistic rather than ontological. Nevertheless, the notion continues to be used in fields as diverse as the philosophy of language²⁴, the philosophy of law²⁵, and most recently artificial intelligence.²⁶ Although partial answers to some of these criticisms have been proposed, there is as yet no comprehensive theory of constitutive rules that answers all or even the most pressing ones. In light of the continuing significance of the notion we need a better understanding of what constitutive rules are.

Searle uses the term “constitutive rule” for rules that define institutional activities. He draws a contrast between constitutive and regulative rules in terms that are reminiscent of Rawls. Regulative rules pertain to activities that are logically independent of the rules, whereas activities that fall under constitutive rules logically depend on them.²⁷

In line with this, Searle writes: “[R]egulative rules regulate antecedently or independently existing forms of behaviour […]. But constitutive rules do not merely regulate, they create or define new forms of behaviour. The rules of football or chess, for example […] create the very possibility of playing such games.”²⁸

According to Searle, constitutive rules define and create (the possibility of) institutional forms of behaviour. The underlying conception of institutions is that they are (systems of) constitutive rules.²⁹

In criticism of Searle, Joseph Ransdell expands this idea by distinguishing between two aspects of institutional terms: connotation and import.³⁰ Roughly speaking, connotation concerns the descriptive conditions that have to be met in order for an institutional term to apply, while its import pertains to what can and cannot be done once those conditions are met. Using baseball as his example, Ransdell maintains that ‘the application of the game-term ‘Bat’ to a given object connotes that the object has a certain size, shape, construction, etc., but the import of the term concerns what can and cannot be done with the object in question within the game, as, e.g., it can be used to Swing at the Baseball by a Batter but cannot be used by a Baseman to trip a Runner, etc.’.³¹

More dialogue on constitutive rules can be found with Hindriks³², but for the purposes of this paper, we won’t go into further detail. Each argument can essentially be reduced to the same thing: when a new, not naturally occurring action is performed, language must be formulated that describes this action in such a way that it can be replicated. Doing so forms a ‘rule’. When many of these rules work in tandem to functionally drive a larger process towards a larger goal, an institution is formed. Thus, rules constitute an institution because they are the only reason the social purpose of the institution can continue to exist.

In this way, the social purpose and the self preservation of the institution are intimately linked and of utmost importance. As defined by its rules, an institution exists to perform a social purpose. If the rules that make this process possible are not preserved, the institution ceases to exist, and therefore can no longer perform. Ian Lustick suggests that the social sciences, particularly those with the institution as a central concept, can benefit by applying the concept of natural selection to the study of how institutions change over time.³³

By viewing institutions as existing within a fitness landscape, Lustick argues that the gradual improvements typical of many institutions can be seen as analogous to hill-climbing within one of these fitness landscapes. This can eventually lead to institutions becoming stuck on local maxima, such that for the institution to improve any further, it would first need to decrease its overall fitness score (e.g., adopt policies that may cause short-term harm to the institution’s members). The tendency to get stuck on local maxima can explain why certain types of institutions may continue to have policies that are harmful to its members or to the institution itself, even when members and leadership are all aware of the faults of these policies.

Lustick himself notes that identifying the inability of institutions to adapt as a symptom of being stuck on a local maxima within a fitness landscape does nothing to solve the problem. At the very least, however, it might add credibility to the idea that truly beneficial change might require short-term harm to institutions and their members.³⁴ Bradley Thayer points out that the concept of a fitness landscape and local maxima only makes sense if one institution can be said to be “better” than another, and this in turn only makes sense insofar as there exists some objective measure of an institution’s quality.³⁵ Assuming that performing specific social purposes are the only reason specific institutions exists, it seems logical to measure the objective quality of said institutions on how efficiently and effectively they perform their social purposes. As we’ve established that self preservation is inherently looped into this performance, we can say that the objective quality of an institution is how efficiently and effectively it performs its social purpose long term or historically. In Lustick’s argument, institutions must occasionally self-harm in the short-term to overcome local maxima in a fitness landscape, and while the sentiment is interesting, I believe the word harm gives off the wrong impression. Rather, institutions must occasionally allow actors from within their audience to correct institutional rule-based behavior through conflict by allowing the subjective experiences of their audience to influence the internal interpretations of these rules. These conflicts can lead to institutional policy changes that might seem harmful in the short-term, but eventually offer the institution a net-positive and allow the institution to overcome the local maxima it faces.

Our outside actor, existing as audience member and hopeful receiver of the institution’s social purpose, only comes into contact with the institution for two reasons: 1) to interact with whatever social purpose the institution is successfully performing and benefit from it, or 2) to interact with a member of the institution in order to resolve a conflict over the ways in which the outside actor is interpreting the performance of the social purpose by the institution, so that they may benefit more efficiently from said purpose in the future. The actor on behalf of the institution is constantly coming into contact with outside actors, but still only has two operations once contact has been made: 1) to steward whatever social purpose the institution is successfully providing to a member of that institution’s audience, or 2) to interact with a member of the institution’s audience in order to resolve a conflict over the ways in which the outside actor is interpreting the performance of the social purpose by the institution, so that he may either clarify the purpose of his practice or change his practice to further increase the efficiency of the institution’s ability to perform the social purpose. Because the institution needs an audience to perform its social purpose and to survive, and because the audience needs the social purpose provided by the specific institution, we can think of their relationship as one represented by (in ideal conditions) mutualistic disjunctive symbiosis.

Conflict within this type of relationship necessarily revolves around the interpretations of rules, as varying interpretations of rules are the singular reason for the existence of these conflicts. So, we must ask: what can be done with rules? There are internal and external equivalencies that exist within these conflicts:

an outside actor can:

  • accept an institutional rule
  • challenge an institutional rule
  • ignore an institutional rule
  • cease interaction with the institution

an institutional actor can:

  • enforce an institutional rule
  • edit an institutional rule
  • ignore an institutional rule
  • remove an institutional rule

The fatal flaw of institutional rules is that they are only as enforceably objective and interpretive as the institution within which they are contained. If we buy that institutions with a social purpose rely on their audience to subsist and language to internally express their rules, then they can never be truly autonomous, as they answer to the institution they are contained within: language. From this, we can say that in times of conflict where the goal of each (internal agent vs. external agent) is productive dialogue, an institute must present an individual from within the organization to interact with the individual outside of it to debate the interpretation of the rule or rules with which the conflict surrounds. The alternative would be a scenario in which an outside individual tries to impress their subjective interpretations of rules onto the rules themselves, without interacting with anyone within the organization that is free to interpret and translate those rules. This benefits neither party, as it would cause the institution to fail at fulfilling its social purpose and cause the audience member to doubt the ability of the institution to do so.

The type of conflict we are interested for the purposes of this paper, then, are conflicts of language. Wittgenstein wrote that all philosophy is just a byproduct of misunderstanding language. One relatively simple example given in the beginning of Philosophical Investigations is that of a sort-of typical philosophical question — “What is a number?”, or more specifically — “What is the meaning of five?”. Wittgenstein presents a “use of language” where the number five cannot possibly have any mysterious “meaning” as many philosophers want to think, and the question itself doesn’t even arise: what five is, is simply what we do with it.³⁶ This concept is true for anything and is inherent to any human interaction involving communication. All conflicts with institutions are problems of language, as language dictates the full capacity of our rules to dictate ideas.

These specific types are conflicts between institution audience members who carry their own subjective interpretations of institutional rules, and actors from within the institution that represent the subjective interpretation of rules provided for them by the institution itself. A successfully navigated conflict sees the institutional actor interpreting and using the institutional rule in a way that both ensures the continued survival of the institution and services the needs of the outside actor.

language without context

The lexical differences between institution representatives and laypersons that we’re discussing in this section are non-contextual — meaning, outside actors are tripped up by vocabulary that is regulative in nature. As a quick demonstration, let’s look at Searle’s example of a regulative term: Say an outside actor, Viewer X, sits down to watch a Football game. While watching, a player within the game is charged with taunting. Viewer X does not know what the connotation or import of taunting is within the institution of football. But, he has a vague idea about its connotation as a general action because the term taunting does not depend on the rules of Football to exist. We can say the word is not constituted by the rules governing Football, but by the rules governing a larger institution, that of general human interaction.

This issue becomes compounded when the outside actor is unable to leverage context at any institutional level (more on this idea later) to come up with even a vague idea of a regulative term’s connotation within the institution they’re currently conflicting with. Let’s expand this idea with another example: Patient X sits down with his doctor in an institution, Hospital A, and is told that he is to have thoracic surgery. Now, if Patient X is well-versed in the constitutive language of the larger institutions of medicine and biology, he is able to form some idea of what thoracic surgery might entail. But, if he — like many patients — is not familiar with either field, the words can be thought of as totally foreign to him and effectively can be treated as unproductive and useless language pending explanation. In this way, specialized lexicons become dangerous gambles that trade on the social purpose of the institution itself.

Further, research done by Pennebaker suggests that function words also play an important non-contextual part in defining roles across the timeline of a conflict. “The,” “this,” “a,” “and,” “an,” “there,” “that,” are all examples of what are called “function words.” They can be thought of as the glue that bind together, so-called “content words” — which convey information by denoting key people, places, things, and situations — into meaningful statements. The hypothesis he poses that we’re interested in is this: that by analyzing the language that two people exchange in a given relationship, you can tell who holds the power and each person’s relative social status. This hypothesis centers on the relative use of the word “I”. The person with the higher status will use the pronoun less often. Further, speakers displaying self-­assurance have a lower frequency of I-words.³⁷ In the specific type of conflict outlined above, an outside actor has less power in the immediate interaction, as they are interacting with the institution to gain something from it. They are initially at the whims of the institution they are interacting with. The institutional actor, on the other hand, is also able to display more self-assurance, as their subjective interpretation of the rule or rules around which the conflict is based has been provided by the institution itself and therefore comes from a place of institutional authority. Problems arise when language style matching (LSM) comes into play, a relatively new tool that stands to shed light on synchrony in real-life relationships in an unobtrusive measure of nonconscious verbal coordination.³⁸

LSM is a dyad-level measure of the degree to which two people in a conversation subtly match each other’s speaking or writing style. Although it naturally occurs in most everyday conversations, LSM is undetectable by both speakers and trained observers.³⁹ Furthermore, like eye gaze coordination, LSM is thought to map directly onto the interpersonal coordination of psychological states.⁴⁰ In seeking a bond, people readily accommodate to one another’s manner of speaking through “language style matching,” getting their function words, or their use of pronouns, in sync. So, within this specific type of conflict, the outside actor, who is acting of their own volition, not the institution’s, and is necessarily at an initial power disadvantage, is much more likely to have a high frequency of I-words. If LSM occurs in such a way that the institutional actor begins matching language with the outside actor, the institutional actor’s frequency of I-words increases, removing him from behind the veil of the institutional authority and making the scenario in which the outside layman blames the particular institutional actor for their problem, rather than the institutional rules themselves, all the more likely.

language with context

A context has physical and communicative dimensions such as: time, space, names, signs and symbols. Manipulation of any of these dimensions results in a changed environment of interpretation. For the purposes of this paper, we’ll assume all physical dimensions (including time) within a given conflict are normalized in such a way that the only dimension truly affecting the outcome of the conflict is communicative, specifically at the concept level, and how the language being used by the institution can often confuse or alienate outside actors by masking the institutional context necessary to fully understand the concepts being presented.

As stated previously, any system — and by extension any institution — is beholden to the limits of the institution that it is contained within. As established, language operates as the governing institution of human social systems. Inherent to language is its interpretability — and this flexibility in both use and understanding is one of its greatest strengths. It makes language adaptable for a variety of functions, allows humanity to retain history and to create narrative, and enables the employment of a variety of devices that make emotion expressible through art. A natural byproduct of this interpretability also acts as one of language’s limitations during times of conflict: it must be interpreted.

To clarify, let’s reintroduce the idea of connotation and import. The connotation C of term X functions as a set of attributes or conditions that can be labeled as X. What can be done with term X once its C has been clarified exists as another set, import I.

Contextual language issues in conflicts arise for different reasons than their non-contextual counterparts. Let’s return to Searle’s football example to better explain contextual language issues: He offers touchdown as a constitutive term in contrast to taunting as his example of a regulative term.

A general, non-contextual connotation for taunting can be summarized as such: “to provoke or challenge (someone) with insulting remarks.”⁴¹

Consulting the NFL rule book gives us a good summary of the connotation of taunting within the game of Football:


(c) The use of baiting or taunting acts or words that engender ill will between teams.

(d) Individual players involved in prolonged or excessive celebrations. Players are prohibited from engaging in any celebrations while on the ground. A celebration shall be deemed excessive or prolonged if a player continues to celebrate after a warning from an official.

(e) Two-or-more players engage in prolonged, excessive, premeditated, or choreographed celebrations.

(f) Possession or use of foreign or extraneous object(s) that are not part of the uniform during the game on the field or the sideline, or using the ball as a prop.

Notice that (c) is extremely similar to our general connotation, and that (d), (e), and (f) are mostly derivative of (c). This is a natural occurrence with regulative terms, because they exist and have connotation outside of the immediate institution. On the other hand, the import of taunting in a general sense is open to fairly broad interpretation. Further, the import of taunting within the game of Football (a specific penalty is called, a team receives a punishment in yards lost, etc.) is fairly unknowable if the agent in question is unfamiliar with the game.

Contrasting this example with our constitutive term, touchdown, demonstrates the main difference between constitutive and regulative vocabulary.

A general, non-contextual connotation for touchdown does not exist, as the term has no connotation outside of the game itself. Within the game, the connotation for touchdown (again pulled from the NFL rulebook) reads as follows:


A touchdown is scored when:

(a) the ball is on, above, or behind the plane of the opponents’ goal line and is in possession of a runner who has advanced from the field of play; or

(b) a ball in possession of an airborne runner is on, above, or behind the plane of the goal line, and some part of the ball passed over or inside the pylon; or

(c) a ball in player possession touches the pylon, provided that, after contact by an opponent, no part of the player’s body, except his hands or feet, struck the ground before the ball touched the pylon; or

(d) any player who is legally inbounds catches or recovers a loose ball that is on, above, or behind the opponent’s goal line; or

(e) the Referee awards a touchdown to a team that has been denied one by a palpably unfair act.

The import of a touchdown within the game is noted as such:

Points are scored as follows:

(a) Touchdown: 6 points;⁴⁴

or: the import of a touchdown within the game of Football is that the scoring team is awarded 6 points.

Because of this interplay, one can know the import of a constitutive term without knowing its connotation. A new viewer to Football might understand that touchdown means a team is awarded 6 points, while not understanding why it is this is true. This is the opposite of a regulative term, where one can potentially understand the term’s connotation without being any more clued into what its import might be.

This phenomena is a direct result of a core idea of this paper: that systems are limited in scope by the system they are contained within. This creates a nested hierarchy of institutions that functions like a set of matryoshka dolls. We can visualize our touchdown and taunting examples in this way:


social interaction -> taunting





social interaction


football -> touchdown

As a second demonstration, we can call on our hospital example. Patient X is told he’s going to have thoracic surgery, which can be thought to be constitutive on a different institutional level than something like code 44 might:



medicine -> thoracic surgery






hospital -> code 44

These are simplified diagrams, but they evoke the necessary considerations for creating workable solutions. This kind of awareness allows us to acknowledge that the institutional level on which speech is operating should be treated as one of, if not the most important coordinates within the communicative dimension to be mindful of when designing for language use in conflict.

our core problems

As outlined above, it seems particularly apparent in the case of conflict discourse that conveying meaning, including the communication of referential content and speaker intention, may not often be the central goal of interlocutors. Bloch’s work on political oratory suggests that the referential function may be highly subordinated to indexical and iconic elements when discourse is saturated with authority.⁴⁵

A ready example of this exists in the specialized lexicons associated with high prestige occupations, such as law and medicine, which are generally justified as means of creating precise meanings — but they also make non-specialists painfully aware of the constraints they face when participating in the production and reception of successful discourse. O’Barr and Conley suggest that the legal system’s monopoly over the discourse practices through which meaning is generated produces confusion and alienation even in small-claims courts, where discourse is supposed to be maximally accessible to laypersons. Cicourel⁴⁶ and Mishler⁴⁷ have argued that serious misunderstandings often arise when professionals use specialized terminology and highly asymmetrical discourse routines while interviewing patients — and thus jeopardize the diagnostic process. Clearly, narrators often do seek to specify meanings and convey them to their interlocutors. When one set of interpretations comes to be accepted as the authoritative account of a conflict and as the basis for shaping outcomes, however, this outcome must be seen less as a natural product of communicative processes than as a particular social construction; attention can thus be directed toward grasping what types of material and symbolic capital (Bourdieu 1991) were mustered in achieving this result and the ways in which they were deployed.⁴⁸

Basso, Brenneis, and Herzfeld argue that narrative discourse presented in settings of conflict don’t necessarily provide revealing insights into the social order. Rather, they argue that lay litigants fail to perceive the systemic and pervasive nature of the control that legal professionals exercise over their participation in the court system.⁴⁹ O’Barr and Conley suggest that even as litigants learn to present their narratives in ways that reflect the ideologies and discursive forms that are predominant in the legal system, they may continue to be motivated by contradictory ideologies — such as the idea that courts exist to right social wrongs. When this discrepancy between lay and legal ideologies leads to discontent on the part of litigants, they are more likely to blame a particular judge or other court official than to hold legal institutions and procedures themselves responsible.⁵⁰

We’ve also identified three major properties of language that often work to create ideological discrepancies. They can be summarized as such:

  1. regulative terms and the rules they are used to generate offer no practical way for an outside actor to determine their import.
  2. constitutive terms and the rules they are used to generate offer no practical way for an outside actor to determine their connotation.
  3. a shift in personal pronoun use during a conflict can cause a shift in power, and as a result, a shift in blame.

Our goal then, is to design solutions for institutions that promote alignment, or at least a cohesive interpretation, of the rules with which the conflict is concerned. These guidelines should not be purported to solve conflicts — rather, they should offer each party a way to better contextualize the other’s interpretation of the specific sets of institutional rules that the conflict surrounds. If an institution is truly dedicated to its social purpose, the understanding that these guidelines offer to both parties should keep the focus of the conversation on the institutional rule at hand rather than allowing the conversation to drift towards personal disagreements based on perceived negligence, dishonesty, or other factors that might exist on the personal level that do not exist at the institution level on the part of the individual actors.

Attempts at design solutions for each of these problems follow below.

design proposal 1: mitigating rule-based confusion

Our first design proposal is a process which attempts to mitigate potential damage from our first two identified problems. To reiterate:

  1. regulative terms and the rules they are used to generate offer no practical way for an outside actor to determine their import.
  2. constitutive terms and the rules they are used to generate offer no practical way for an outside actor to determine their connotation.

Essentially, these problems can be reduced to a single statement: outside actors are capable of misunderstanding institutional terms, and institutional actors are are capable of assuming outside actors understand everything they are saying. To reduce even further: words are interpretable, and this can confuse people.

Before we fully launch into our strategies, let’s first explicitly outline what happens in any conversation (and therefore, any cooperative conflict). In the diagram that follows, our outside actor’s actions will be in regular typeface. Our institutional actor’s actions will be in bold.

This back-and-forth process proceeds until a conversation ends, regardless of outcome. In our case, the outcome goal of a cooperative conflict is to reach a positive outcome for the outside actor without sacrificing the core social purpose of the institution. To state again, the proposed process is not designed to ensure this result. At each intersection of interpretations and expectations between the two parties, the opportunity for misinterpretation occurs due to potential misunderstandings of the pre-supposed connotations and imports either actor associates with the terms they’re using. Rather, then, the proposed process is designed to ensure a mutual understanding of communication that might facilitate a positive result by helping to remove misinterpretation of either party’s intention within the language they are using.

Our process for proposal 1 is going to include four steps: identify, provide, annotate, adapt.


An old school participatory design technique called ‘card sorting’ has in the past been used to explore how participants group items into categories and relate concepts to one another.⁵¹ This technique has been used in digital interface design before. Participants are given cards with printed concepts, terms, or features on them, and are asked to sort them in various ways. One of the most common reasons to do a card sort is to identify terminology that is likely to be misunderstood, either because the terminology is vague or because multiple meanings are associated with it. The technique has also been used as a method to generate options for structuring information, as it can identify different schemas for organizing navigation, menus, and taxonomies. Finally, the technique can be used to evaluate categories by helping to identify items that are difficult to categorize or perhaps aren’t as important as others. The method can validate that the concepts and categories presented by a service or product actually reflect the mental model of the audience the service or product is interacting with.

Our identify strategy is going to operate as an innovation on the card sorting method. Essentially, we’re going to propose that any time a chat conversation occurs between an outside actor and an institutional actor, a mechanism exists that allows the outside actor to ‘tag’ a word within the conversation as ambiguous.

If the outside actor takes the action of tagging a word within the conversation as ambiguous, this information is stored by the institution. A prototypical display of how this mechanism might look follows:



Our second proposed strategy assumes that an institution has followed through with our first, that is to say, that they have aggregated data on words that outside actors have tagged as ambiguous. This data allows the institution to visualize which words are regularly tagged by outside actors and which are not

With this data, the institution can then launch an addition to the preliminary tagging mechanism that annotates popularly ambiguous words. This proposed secondary mechanism would prompt users who are in the action of tagging a word they find ambiguous to read and learn the institutional definition of the word. This definition should be displayed in a simple layout that highlights both the connotation and import of the term. By doing so, we cover our regulative and constitutive bases in one action while effectively removing the ambiguous nature of the term and adding it to the outside user’s vocabulary.

As the institution works through tagged words in a top-down fashion, it can roll out definitions for the most often tagged ambiguous words, work its way down to semi-often tagged words, and so on. This process is a two-fold positive for the institution: it requires them to consider the language being used to negotiate with customers, while also educating their users in a way that is low-weight on the institution’s end once it’s set up.

A prototypical display of how this mechanism might look follows:



Our third suggested strategy in the process is an elevated way of aggregating customer data. Any time an outside actor engages with a concept to either tag it as ambiguous or to learn about the institutional understanding of it, they should also be given the opportunity to annotate it with their personal understanding of it. This strategy has several key upsides:

  • The institution can aggregate data to the word on which of their communication policies are failing.
  • The institution can train their users to think of concepts in terms of their connotation and import by developing annotation prompts that drive users to define (in their impression) both the cause and the impact of the ambiguous concept.
  • By prompting for ‘cause’ and ‘impact’, the institution can aggregate a better picture of which terms built into their communications policies are being treated as regulative and which as constitutive by studying which side of the terms the outside actor is getting confused about. This allows the editing process to have real, immediate impact and helps institutional actors communicate more effectively the existing policies of the institution in terms their audience is already familiar with.
  • Finally, this policy allows the institution to see which terms are failing the outside actors in non-ambiguous ways. If outside actor and institutional actor definitions are in alignment, but the social purpose of the institution is still unsuccessful, we can say that the issue is not with the language, but with the institutional policy. Therefore, this strategy acts as a failsafe that ensures cooperative conflict resolution that goes beyond policies of language by offering outside actors an opportunity to elicit organizational change, or to at least alert the institution that their organization is failing in some way.

A prototypical display of how this mechanism might look follows:



Any valuable design process includes a review stage. This allows the process to be cyclical and therefore continuously improvable. We’re calling this final strategy adapt, as the ultimate goal of this design process is improving the success of the institution to deliver on its social purpose by learning how to optimize its policies via direct communication with its customers.

As with our initial step, we’re going to propose a slight innovation on an old school design method: the cognitive walkthrough.⁵² Essentially, the cognitive walkthrough is a usability inspection method that evaluates a system’s relative ease-of-use in situations where preparatory instruction, coaching, or training of the system is unlikely to occur. In these situations — when a person must actively engage with an interface to know what to do next, rather than relying on preexisting knowledge of the system — each step of the interaction with the system can be assessed as a step that either moves the individual closer to or further from his goal. The cognitive walkthrough provides a systematic way to identify these points during an interaction sequence, and then evaluate whether each step is more likely to fail or succeed in helping people make the next correct decision in the interaction.

Generally, the method centers around asking users the same four learning theory-based questions for each step in the action sequence:

  • Will users want to produce whatever effect the action has?
  • Will users see the control (button, menu, label, etc) for the action?
  • Once users find the control, will they recognize that it will produce the effect that they want?
  • After the action is taken, will users understand the feedback they get, so they can confidently continue on to the next action?

Our proposal includes testing for these questions, as well as our innovations on this general process, which are two-fold. First, we believe that cooperative conflict resolution requires both parties to understand the process in similarly productive ways. Traditionally, the cognitive walkthrough is used to train users to interact with a non-negotiable system. As part of our goal includes creating more flexible systems of service, the adapt stage of this process includes running institutional actors, as well as users, through the cognitive walkthrough scenario. This process allows the institution to learn something about the experience of its user and to adapt accordingly.

Second, we propose the adapt strategy develops a program that institutes several long-term follow-ups to the cognitive walkthroughs offered to early adopters and users of this specific design process. If our goal is to empower institutions and their users to collectively create more usable policies, we need to confirm that the process described is accomplishing that goal. These long-term surveys allow the institution to be reflectively adaptable with the proposed design process and to engage the outside actor in a way that validates their history with the institution.

design proposal 2: removing the “i” from team

The final core problem we’ve identified requires a bit more of a rigid approach than the fluid undertakings of the aforementioned steps detailed in design proposal 1. To reiterate:

c) a shift in personal pronoun use during a conflict can cause a shift in power, and as a result, a shift in blame.

As a reminder: Pennebaker’s research demonstrates that, when seeking to bond, like one that might occur as a result of attempting cooperative conflict resolution, people readily accommodate to one another’s manner of speaking through language style matching — e.g. syncing their function words, particularly the use of the personal pronoun “I”. When this occurs in such a way that the institutional actor matches personal pronoun use with the outside actor, a power shift occurs that removes institutional authority from anything else he might say or do when interacting with the outside actor.

There seems to be a rising trend in customer service that dictates that institutional actors within an organization should appear to customers as individuals who are momentarily part of an institution. Anecdotally, consider the last time you called in to get help from a representative when your cable stopped working. The representative probably started the conversation the same way they always do, with some version of: “Hello, I am _______. I’d love to help you with your problem today. Let’s get started.”

In fact, if we engage 3 companies ranked in the top 10 worst at customer service in 2015 in a live chat with a representative (when available), here are the first sentences we are greeted with:


  • Corazon: Hello Drew_ [sic], Thank you for contacting Comcast Live Chat Support. My name is Corazon. Please give me one moment to review your information.
  • Corazon: Hi! Glad to have your [sic] on chat! I look forward to helping you today.


  • Mary Rose R.: Thanks for chatting with! I’m Mary Rose R. and I’ll be assisting you today.

Bank of America:⁵⁵

  • Ayeashia: My name is Ayeashia. I assist with personal credit cards as well as auto loan applications. May I please have your name?

In each interaction, before we, as outside actors, have the chance to say anything, the institutions that we are interacting with have given up their institutional authority by way of their own internal actors. Now when we enter into conflict with these named individuals, we do not seek to immediately blame the policies they are enacting, rather, we blame Corazon, Mary Rose, or Ayeashia for being incompetent at their jobs. We hang up and hope for a “better” representative next time, or we cease interacting with the organization.

An argument that can be made is that this is beneficial for big businesses. Why fix policy problems if their customers have to pay them anyways? Personalized representatives act as a defense mechanism in this setting, by deflecting any complaints the outside actor has with institutional policies to the competence of the immediate institutional actor. This protects business practices that might otherwise be changed to offer better service at an incurred cost. And while this seems understandable, there are interesting implications when this approach moves beyond the realm of big business.

Take, for example, the institutions of law and medicine, both of which were ostensibly designed to serve a population and have since shifted to behave more like businesses. Again, this is understandable — where there’s money to be made, money will be made. And it is, often and in great amounts. Or, take the approach of weak AI systems like Siri, Cortana, or even simple GPS guides — tools designed to offer specific services as part of a larger product. In John Frank Weaver’s book Robots are People Too… he denotes several instances in which accident participants have blamed weak AI — most notably GPS — for the accident:

  • In 2009, a driver followed the directions provided by his GPS, until he was trapped on a narrow cliffside path. Police arrived to tow him back to the main road. The man went on to blame the GPS system in a British court of law.
  • In 2011, a judge found that a GPS device was partially to blame for a fatal car crash, stating that a driver placed too much reliance on the GPS’s map of the road.
  • In 2012, a driver drove through a “No Entry” sign, pulling into the path of another vehicle. The driver claimed he was following directions from his GPS, despite the fact that similar signs existed for 180 yards leading up to the intersection.⁵⁶

Each of these weak AI systems has been given a name and speaks in the first person. When we hold the home button on our iPhones to prompt an interaction with Siri, the first text we meet is: “What can I help you with. Some things you can ask me — .” And what happens when Siri is unable to help or does so poorly? We blame Siri, not the larger rules that govern her.

So, the solution to this particular phenomenon seems fairly simple. As the market becomes more conversational, any institution that actually wants to improve the service it is offering should forego the use of the “I” pronoun when extending an institutional actor to interact with an outside actor. Let the “human” component occur within the communication itself (via the process suggested in design proposal 1, not within the larger context of the situation.

in conclusion

In 1964, Marshall McLuhan published Understanding Media and with it coined the phrase, “the medium is the message.” Underpinning the meaning of this phrase is an adaptation of the Gestalt psychology idea of a figure and a ground. He leveraged this idea to relate a form of communication (the medium or figure) to its context (or ground). Essentially, he argued that the figure necessarily operates through its ground.⁵⁷

In order to grasp fully the effect of a new technology, he argued, one must examine figure and ground as one — as the two cannot not be substantiated independently. Media, then, must be analyzed in their historical context, particularly in the context of what preceded them. The conclusion McLuhan reaches is this: the message which the medium conveys can only be understood if the medium and the environment in which the medium is used — and which, simultaneously, it effectively creates — are analysed together.

In “An idea about texting,” Jonathan Libov pulls a quote from Always bet on text: “Text is the most powerful, useful, effective communication technology ever, period.” He also writes: “I think many of us are quite convinced that in the not too distant future, messaging is going to mean a lot more than it does today.”

If these assumptions are true, and the paradigm shifts mentioned by Messina in his article series (1, 2) come to fruition, then the real goal of this paper is to initiate design that acknowledges the limited bandwidth for context that text-based messaging currently provides to the parties involved, while managing to expand this bandwidth in ways that do not make future interactions more convoluted.

This paper offers an attempt at dealing with a very small sliver of these interactions, but the process seems like a replicable way to think critically about the solutions we as designers begin to offer in the expanding conversational commerce.


1 Kadushin, C. (2012). Understanding socil networks: Theories, concepts, and findings. Oxford: Oxford University Press.

2 ibid

3 e Nooy, Wouter (2012). “Graph Theoretical Approaches to Social Network Analysis.” in Computational Complexity: Theory, Techniques, and Applications (Robert A. Meyers, ed.). Springer. pp. 2864–2877.

5 ibid

6 Riketta, M.; Nienber, S. (2007). “Multiple identities and work motivation: The role of perceived compatibility between nested organizational units”. British Journal of Management 18: S61–77.

7 Thomas, Richard K. (2003). “Society and Health: Sociology for Health Professionals.” Kluwer Academic Publishers. pp. 104–105

8 Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–47). Monterey, CA: Brooks/Cole

9 Sherif, Muzafer. (2015). Group Conflict and Co-operation: Their Social Psychology.

10 ibid

12 Google dictionary: op·ti·mize ˈäptəˌmīz/, verb make the best or most effective use of (a situation, opportunity, or resource).

13 Eder, Donna. (1990). “Children’s Peer Cultures” in Annual Review of Sociology, Vol. 16: 197–220 (Volume publication date August 1990).

14 Briggs, Charles L. (1996). Disorderly Discourse: Narrative, Conflict, & Inequality. Oxford University Press.

15 ibid

16 ibid

16 ibid

17 ibid

18 ibid

19 ibid

20 Mooney, Knox, and Schacht. (2007). “The Three Main Sociological Perspectives” in Understanding Social Problems, 5th edition.

21 Mattheis, Clemens. (2012). “The System Theory of Niklas Luhmann and the

Constitutionalization of the World Society” in Goettingen Journal of International Law 4.

22 Hindriks, Franks. (forthcoming). “Constitutive Rules, Language, and Ontology” in Erkenntnis.

23 ibid

24 ibid

25 ibid

26 ibid

27 ibid

28 ibid

29 ibid

30 ibid

31 ibid

32 ibid

33 Lustick, Ian (2011). “Institutional Rigidity and Evolutionary Theory: Trapped on a Local Maximum”. Cliodynamics 2

34 ibid

35 ibid

36 Wittgenstein, Ludwig. (1958). Philosophical Investigations. Basil Blackwell Ltd.

37 Pennebaker, James W. (2011). The Secret Life of Pronouns: What Our Words Say About Us. Bloomsbury Press.

38 Gonzales, Hancock, & Pennebaker. (2010). “Language Style Matching Predicts Relationship Initiation and Stability” in Psychological Science OnlineFirst.

39 Ireland & Pennebaker (in press).

40 Richardson, Dale, & Kirkham. (2007). “The Art of Conversation Is Coordination: Common Ground and the Coupling of Eye Movements During Dialogue” in Psychological Science.

41 Google dictionary: /tônt/, verb, gerund, or present participle: taunting, provoke or challenge (someone) with insulting remarks.

42 “Rule 12, Player Conduct” in NFL Rule Book. (

43 “Rule 11, Scoring” in NFL Rule Book. (

44 ibid

45 Briggs, Charles L. (1996). Disorderly Discourse: Narrative, Conflict, & Inequality. Oxford University Press.

46 ibid

47 ibid

48 ibid

49 ibid

50 ibid

51 Martin & Hanington. (2012). Universal Methods of Design: 100 Ways to Research Complex Problems, Develop Innovative Ideas, and Design Effective Solutions. Rockport Publishers.

52 ibid

53 Initiated by author, January 2016.

54 ibid

55 ibid

56 Weaver, John Frank. (2013). Robots Are People Too: How Siri, Google Car, and Artificial Intelligence Will Force Us to Change Our Laws. Praeger.

57 Old Messengers, New Media: The Legacy of Innis and McLuhan, a virtual museum exhibition at Library and Archives Canada