Introduction

Here is the basic argument. An interest in complex co-evolutionary systems arose and continues to grow in organizations because of an increasing frequency of unexpected changes in the worlds they interact with. Traditional models that rely on independent variables and linear cause and equilibrium states cannot handle such problems. Complexity models take the facts of interdependence and nonlinearity and dynamics as foundational. So organizations look to complexity for ideas on how to handle this brave new constantly changing world.

The problem of frequent and surprising change is not new. In 1970 Alvin Toffler, in his book Future Shock (1970), had already noticed that change was increasing at an increasing rate. For that matter, so did Henry Adams with his ‘Law of Acceleration’ in 1904 (1999). Complexity theory, a later arrival that flourished in the 1980s, not only notices this but predicts it based on the increased connections, interactions and feedback loops that our accelerating processes of globalization have brought, and are still bringing about. Any organization, any node, in the global network can now expect frequent perturbations from God knows where, God knows when.

So, one characteristic of an organization adapting to this complex world lies in the recognition of a simple principle. Stability is no longer the norm, with change as the exceptional event. Now change is the norm, with stability as exceptional. Complex organizations lead, manage and practice based on that fact.

Given this principle, a large number of consequences need to be spelled out and tested in the field of organizational life, a process now ongoing. A sampling of the many implications of the principle can be found in sources such as Olson and Eoyang’s (2001) and Stacy’s (2001) applications of complexity to organizational development.

In this article, I only mean to deal with one of those consequences. The change principle needs to be linked to the way people in the organization use language to notice unexpected events and make sense out of them. Once again, the basic idea here is an old one that pre-dates complexity theory. People who actually do things in an organization will have good ideas about how to do those things better. The idea has its roots in early ‘human relations’ antidotes to scientific management (Trahair, 2005). It developed with Drucker’s early writings on the ‘knowledge worker’ (Drucker, 1998), and it surfaces in most all of the trendy management models of the last few decades.

Let’s call this simple idea the ‘local knowledge’ principle. And once again, notice that the principle is rich in consequences for leadership, management and work.

Now combine the principle of local knowledge with the principle of continual change and look at language. Perturbations, which occur with increasing frequency, are surprises – low probability events – that are first noticed in the organizational sites where they have an impact. ‘Notice’ means news, hot topics, something that we will discuss under ‘tellability’ later in this article. An organization, then, can use these local noticings to spot perturbations. Local knowledge, or, more precisely, surprises with reference to it, serve as early warning signs of changes that an organization should monitor.

But noticing, useful as it might be, isn’t enough to take advantage of all that local knowledge has to offer. Practitioners are rich resources of experience, evaluation and innovation. Practitioners don’t just notice. They talk about what they notice with each other. The talk sets the new event in context, links it to possible causes, significant actors, and related events. The talk speculates on what the change means, uses the talk to imagine a future that includes it. An organization needs to take such talk seriously, not only as an indication that a potentially important change is underway, but also as a resource to evaluate and act on it.

None of this means that local noticing and talk is the only way to build complexity into an organization that matches the complexity of an interconnected changing world. Nor does it mean that local noticing and talking are necessarily correct and that all other sources are always wrong.

But this article does argue that local noticing, and local talking about what is noticed, should be valued by an organization in a complex world. When change is assumed to be the normal situation, and when those closest to the change are viewed as sensitive reporters and creative problem-solvers, then opening a space for that talking and enabling its wider distribution makes a lot of sense.

I think this is one of the reasons for interest in story and narrative in the organization. Story and narrative point towards those moments that open with shared noticing of a change among co-workers, moments that continue with subsequent efforts to make sense of that change. If ‘complex’ organization means an organization capable of responding to frequent surprising changes, then local noticings and stories are critical resources to develop that capability.

The problem here is, ‘story’ and ‘narrative’ are used in many different ways. One popular use of the term ‘story’ comes to mind as a good example of how to adapt the term to our quest for local knowledge in a time of change. In Beating the Street (Lynch, 1994), one of his books for the new investor, the legendary Peter Lynch wrote that the secret to investing was to go after the story of a stock. Lynch encouraged us little guys to become ethnographers, to look for company details in our daily life. What were their stores like in the nearby mall? What did friends say about them? Any buzz about their products around the water-cooler at work? Lynch of course also recommended a look at quarterly earnings and debt, but it was the pattern of these little local details that gave an individual investor an advantage over the Wall St. professional. The little guy could put together a story of what was going on with a company by finding relevant material in everyday life.

And once he put all the details together, the little guy could examine the ‘arc’ of the story. Was the company moving in a positive or negative direction? Was Cinderella on her way out the door to empty the chamber pot, or off to the ball in a pumpkin turned into a coach, or rushing home before the gown turned into rags, or watching Prince Charming slide the glass slipper onto her foot? Or was it the adult version where the Prince spent the night and then never called again? Or tried to call but Cinderella changed her number. Depending on where the investor picked up the story, s/he would buy or sell accordingly.

Was Lynch really talking about a story? Well, no and yes. If you swear by the old Aristotelian virtues, a story should have a beginning, a middle and an end – a complication, a development, and a resolution. What I’ll call ‘Lynch-stories’ do not. Lynch-stories organize different kinds of information to make a best guess at the near future. Most of the information will be gathered at the same time as the story is constructed, and the future that is aimed for will be close at hand.

We can quibble over where ‘beginning’ and ‘end’ start and stop, but it’s more accurate to say Lynch-stories are actually in media res with reference to some larger, more complete story. Lynch-stories are all development, without the initial complication and the final resolution. A traditional story can also begin in media res – no question about it. But with Lynch-stories, a story doesn’t just start with in media res. In media res is all you get.

What about narrative instead of story? Is a Lynch-story a narrative? Well, yes and no. The heart of narrative is time and cause. There are post-structural arguments that neither time nor cause are necessary, that a verbal evocative swirl can do the job as well as the traditional ‘this happened and then this happened and so look how it turned out in the end’. For purposes of this article, I’d like to rule out non-narrative stories, not on any interesting intellectual grounds, but just because I’ve got enough problems to deal with as it is.

Grant me that time and cause are the heart of narrative. Is a Lynch-story a narrative? Cause is certainly key, but with a twist. ‘Cause’ in traditional narrative is about putting things together to explain something that the narrator – and maybe the audience – already knows has happened. But in a Lynch-story, cause means guessing about something that hasn’t happened yet, something that the ‘author’ doesn’t know either. There may be more than one guess, and the guesses might be wrong. The audience can’t just enjoy the way the author gets them there. There isn’t any ‘there’, there and won’t be until after the story is done.

A Lynch-story is more an open-ended exploration, where information of all types is gathered and put together to figure out what might happen. Cause in a Lynch-story is more about guessing the near future using whatever information is available. Experiences go into words, and the words reduce the size of a large space of possibilities, and within that space the story makes some paths more likely than others.

In this sense, a Lynch-story is very much about cause in a narrative sense. But the time-line is more abbreviated than usual. The ‘this happened and this happened’ are more co-occurring rather than sequential events widely separated in time. The time range of a Lynch-story is narrower than the concept of narrative leads us to expect.

Why all this obsession with Peter Lynch? After all, his advice didn’t make me rich. I draw from him here because his use of ‘story’ does offer a clear example of the problem I want to work on in this article.

The problem is this: ‘story’ and ‘narrative’ are extremely rich and ambiguous concepts with a long and contentious pedigree in many fields having to do with human expression – language, literature, film, the plastic arts, music, etc. As I look at the concepts through my background in linguistic anthropology, I find what I read and hear confusing. In the hope that some of my confusion might be helpful to others, I’d like to use this article to clarify if not resolve it.

The first clarification: ‘story’ and ‘narrative’ are actually about many different ways that language is used. The variety I will focus on here foregrounds a Lynch-type story, an in media res construction out of lived experience, without a clear sense of a beginning or end, but with a clear interest in guessing why something happened and how what happened might effect the very near future.

This of course isn’t the only way the concepts might be used, but it is one way. And this way of narrowing the concepts brings us to another well-known concept that links Lynch-stories to organizations. Karl Weick writes about what he calls ‘sensemaking’, a term I will use for the rest of the article as a label for this Winch-like use of language in organizational contexts. Weick describes it like this:

“…sensemaking is about such things as placement of items into frameworks, comprehending, redressing surprise, constructing meaning, interacting in pursuits of mutual understanding, and patterning” (1995: 6).

Not exactly a crisp definition, but clearly related to the earlier discussion of Lynch-stories. Important as his concept is, it suffers from the same problem as the use of ‘narrative’ and ‘story’ earlier in this article. It rambles all over the discourse territory.

Sensemaking isn’t the only related concept, far from it. David Boje, an editor of this special issue, talks about ‘antenarrative’, which is the “fragmented, nonlinear, incoherent, collective, unplotted and pre-narrative speculation, a bet” (Boje, 2001: 1). His book offers several approaches, drawn from literary theory, to undermine a polished story and convert it into antenarrative. Boje highlights the dynamic and improvisational nature of sensemaking and offers methods to force its occurrence. But again, he remains distant from the linguistic detail.

Jose Fonseca takes a different approach, drawing on the symbolic interaction of George Herbert Mead. He describes the conditions, the ‘enabling constraints’, that an organization must offer for sensemaking to be useful. “Innovation will be presented as the emergent continuity and transformation of patterns of human interaction, understood as ongoing, ordinary complex responsive processes of human relating in local situations in the living present” (Fonseca, 2002: 3). It is here that “innovative meanings emerge.” He spells out some of these ‘enabling constraints’, such as adequate diversity, but also commonality, trust among participants, and the like. But as in the symbolic interactionism from which he draws, there is little about language itself.

As part of his general theory of dynamic economic process, David Lane (Lane & Maxfield, 1996) introduces the notion of ‘generative relationship’ as the kind of “productive human relating” that Fonseca describes. He emphasizes how, within a relationship with enough solidarity and agreement such that communication can occur, the right balance of differences in perspective will generate innovation. Lane anticipates this article with his emphasis on continual conversation in the face of change, but again, he stops before reaching the level of conversational detail.

These different angles on sensemaking – Weick, Boje, Fonseca, Lane and others – fit the concept we’re after, how it is that people in the organization notice frequent changes and then talk them over. But they still don’t get as close to the living narrative ground as I want to go with my background and biases from linguistic anthropology. If we zoom in on one of an organization’s small groups in the middle of sensemaking, what would be the texture of what we hear? How would the talk look in a transcript?

For now let’s stick with ‘sensemaking’ as our general cover term. Sensemaking occurs when a group uses language to collectively link something that happened with a guess about why it occurred, and what changes it might bring about in the near future. Sensemaking is a social in media res construction where time and cause come together to explore possible futures.

People who do the frontline work notice a perturbation and make sense out of it. Usually we associate this sensemaking with the water-cooler, the car pool, or the long lunch. Now we want to make more of it than that. We want it to be incorporated into an organization as a regular process for monitoring and evaluating changes. Sensemaking becomes a cornerstone for the bottom-up innovation advocated by complexity-theory oriented sources cited earlier. Whether the sensemaking is useful or whether changes actually result from it – these are of course other questions that we discuss later.

Suppose someone, somewhere, built a dream organization that figured out how to support and use sensemaking among its members. Suppose we lifted the roof off and eavesdropped on the talking that went on inside of it. Suppose we listened for sensemaking, either as researchers from the school of business and management or as organizational consultants looking for some ‘positive deviance’ to amplify. Could we, with all that linguistic anthropological baggage I drag around, know in advance what a sensemaking session should sound like?

Here I can draw on decades of research on how people use language in daily life, work called discourse analysis, conversational analysis, pragmatics, and other unprintable things. This diffuse field is often anchored in the pioneering fieldwork of Bronislaw Malinowski, who worked on the Trobriand Islands early in the twentieth century. There are many other honored ancestors, such as Bhaktin and Wittgenstein as well. All of them tried to figure out how people use language to make life happen even as they live it. That’s what we want to do for organizations with the concept of sensemaking. Let’s see if we can use the ancestors to bring Weick’s concept to life.

To do so I will draw heavily on the recent work of a descendant of those ancestors, linguistic anthropologist Eleanor Ochs, whose book Living Narrative (Ochs & Capps, 2001) with the late Lisa Capps offers concepts for the grounded analysis of sensemaking as well as for the complications in the narrative concept that we seek. “Living narratives,” as Ochs and Capps call them, are what happen when people come together to “build accounts of life events.”

“[N]arrators are often bewildered, surprised, or distressed by some unexpected events and begin recounting so that they may draw conversational partners into discerning the significance of their experience. Or, narrators may start out with a seamless rendition of events only to have conversational partners poke holes in their story. In both circumstances, narratives are shaped and re-shaped turn by turn in the course of conversation” (2001: 2).

This is quite different from polished finished stories, coherent examples, or stories with a beginning, a middle and an end built on a narrative spine of chronology and cause. The narratives Ochs and Capps want to deal with, stories that arise spontaneously in ordinary conversation, aren’t so neat: “…conversational narrative will include questions, clarifications, challenges and speculations” (2001: 19).

Ochs and Capps don’t want to ignore smooth traditional narratives or pretend they don’t exist; they want to broaden the framework that includes them so that we don’t forget the messy end of the spectrum, the more frequent one, the end that contains the kind of sensemaking that Weick is after.

How do they create such a framework, one that includes both living narratives and polished performances? They outline five dimensions of narrative that show the difference between a smooth, polished story and a rough story in media res. The five dimensions are:

  1. tellership;

  2. tellability;

  3. embeddedness;

  4. linearity, and;

  5. moral stance.

Polished and smooth narratives cluster on the left hand side of the scales defined by the five dimensions. A life history that I gathered years ago, for example, was on the smooth end. It involved only one active teller and a story with guaranteed tellability – he’d told it many times and knew that audiences liked it. The narrative was not embedded in another event – it didn’t come up while doing something else; rather, the situation was organized around the telling itself. The narrative was linear, a neat sequence of cause and effect in a story that had a clear beginning, middle and end. And it had a constant moral stance – in the case I’m thinking of here, an arc of character development from down-and-out and alone to sophisticated and socially connected.

Narratives relevant to an organization can also be found at the polished and smooth end of the scale as well. For example, in her chapter in a recent overview of discourse analysis, Charlotte Linde writes of the stories organizations tell to teach employees “how things are done around here” (Linde, 2001). In her work, a polished and frequently told story promotes equilibrium rather than elaborating on change. This, like the life history narrative, is a perfectly reasonable use of the narrative concept, but it is obviously not what I’m after in this article, nor is it the kind of narrative that Ochs and Capps want to describe.

Now consider the other end of the scale, the right-hand side more typical of living narratives. Many tellers rather than just one. Several people try to tell the same story, interrupting and contradicting each other, elaborating on what one another says, etc. The story probably focuses on surprising news – tellable in that sense – but the news isn’t worked out and the conversation skids into known and even boring side-issues. The narrative is embedded in another situation, part and parcel of what’s going on even as it is being told, rather than a special event where an audience sits and listens. It will be nonlinear rather than linear, drifting and wandering around and looping back onto itself. And all those co-tellers may well dispute the moral evaluation of what is said – one thinks it’s funny, another sad, a third evil, and a fourth disgusting beyond belief.

Sensemaking lies towards this right hand side of the scale, the ‘living’ side, unplanned and occasioned rather than a bounded practiced performance. But that isn’t the only lesson to draw from Ochs and Capps’s work, not to mention from the rest of the tradition of discourse research. I want to shift now to some consequences of the match. Of what use is it to conclude that sensemaking is living narrative?

Consider the hierarchical, ‘command and control’ structures where – in the worst parody – one does not speak unless spoken to, and then only to agree with what was just said. Ochs and Capps’s narrative dimensions help see a bit more clearly what has to happen to open up the hierarchical controlled space into something different.

The smooth and polished end of Ochs and Capps’s narrative scale is more like what you’d expect in a command and control organization. One teller runs the show. A narrative is tellable because the teller says so. It is not embedded in what the listener is doing – listeners stop what they’re doing and gather around and listen. The story will be nice and linear, the clear and certain causal analysis you’d expect from the top of the hierarchy. The moral tone will be clear; no ambiguity or discussion allowed.

Contrast this with a sensemaking organization. We’d hear many people collectively making the story; we’d hear contradictions and interruptions. Several people would be tellers. We would hear tellability questioned and justified. Embeddedness would vary. Some tellings might occur in the middle of the situation that was the topic – “See, look at that, what did I tell you?” Others might happen during a break, or even in a meeting where the topic was a featured agenda item, or maybe the agenda item. Linearity – cause and effect and sequence – would be a goal, not a guideline for how to reach it. And sensemaking might end with multiple possibilities or open questions, things to be looked for and thought about between that telling and the next. Same with the moral tone – disagreement and ambiguity would abound in the beginning. In the end, rather than a clear moral tone, it might be a “does the good outweigh the bad” kind of dilemma, maybe a dilemma that would keep a few people awake that night, maybe for the rest of their lives.

Ochs and Capps take the reader through numerous examples of living narratives in their book. When they describe those examples, they draw on a discourse analysis tradition in which I also participate. This tradition offers numerous ways to describe the details of language as it naturally comes, details where the dimensions of narrative that Ochs and Capps describe come to life. In the next several sections, I’d like to introduce a few of these details to describe more precisely what sensemaking might look like. We now push down a level and look under the hood.

Consider the tellers first. Sensemaking is social discourse, talk with others, a collective enterprise where agreement and disagreement from participants shape the stories that emerge. There are multiple tellers. How do they co-ordinate what they are doing?

If we look closely, we see that some discourse mechanisms are, in fact, about the micro-politics of the group – who gets to tell what when? For example, one of the foundational concepts of conversational analysis is ‘turn-taking’. In a multi-teller conversation, no formal rules specify who speaks when. Turns have to be negotiated.

Now imagine we look at some transcripts of a sensemaking session. Look at the details. How are the turns worked out? It is normal for there to be overlaps and multiple speakers as people try to get the floor. Overlaps that occur at ‘possible completion points’ are more egalitarian than a heavy-handed takeover when a speaker is in mid-utterance. There may well be a dominant figure or two, but do they moderate or do they control? Who gets the floor when, and once gotten, do they ever give it up?

Smooth turn-taking is diagnostic of good sensemaking, since a characteristic of sensemaking is multiple tellers building a story out of their different points of view. Turn-taking reveals whether the sensemaking is a dictator looking for an audience or a group co-operating to make sense of something.

Turn-taking might show that what is supposed to be going on is in fact the opposite of what is actually going on. Consider the endless stream of articles that describe how a ‘new’ management model has failed. Typically the new model pushes in the direction of opening up space for sensemaking in a strongly hierarchical system. It advocates a move away from command and control. So, one day the new model is announced to be in place. But a close look shows that, in reality, hierarchy and command and control continue when actual discourse is underway. To use Argyris’s (1993) terms, the ‘espoused theory’ changes on the surface, but the ‘theory-in-use’ stays the same.

The contradiction between ‘espoused’ and ‘in-use’ is obvious if you look at the details. Turn-taking shows that the old hierarchy still dominates. Turn-taking control of sensemaking guarantees that it will have none of the characteristics that it is supposed to have.

Why does the group let the dictator get away with it? Everyone knows a new model is supposed to be in play. One possibility is simply that the dictator is an insensitive self-centered jerk. It’s the down side of the ‘team’ notion, one that is too infrequently discussed. That problem can be fixed, though it calls for an intervention, or possibly a personality transplant.

The more likely possibility is, the dictator controls group resources and uses them to punish those who do not stay in their subordinate role. The dictator may well control whether or not the others can be in the sensemaking session at all, or even whether their employment continues. The control might be official – a supervisor, say – or it might be informal – the brother of the CEO works in the office. The power of the dictator is obvious and direct – he or she can influence or directly control whether you continue to be part of the conversation.

Turn-taking, that narrowly defined academic concept, evaluates whether there are really multiple tellers or not. Just listen or, better, comb the details of a transcript or two. The way turns are organized will be a clear diagnostic of whether an organization is co-evolving or chain-of-command.

There are other potential diagnostics as well. Consider the adjacency pair, couplets found in every language, like “how you doing?” and “just fine, thanks.” If someone uses the first part of an adjacency pair, then someone else is required to perform the second part. If they don’t, the failure is interpreted as a breach, an expression of bad intentions. Particularly important here is the question/answer adjacency pair. If someone asks a question, it sets up a linguistic imbalance that demands to be restored with an answer. But who is asking the question? The same person, over and over again? Perhaps many are talking during a sensemaking session, so it sounds like multiple tellers. But if the same person is always asking a question, s/he is limiting and directing what it is the next person can say. Q/A adjacency pairs – like turn-taking – control the topic, and control over topic is control over how and in what terms sense will be made.

Now let’s move from tellers to tellability, the second narrative dimension that Ochs and Capps describe. A traditional narrative is made to tell. It is supposed to be tellable by definition. A living narrative, though, has to earn its tellability.

Tellability links to the question, “is it news?” Tellable means inspired by some event that departs from expectations, ranging from something astonishing and incomprehensible to something that was a mild surprise. More often than not, the news is bad, not good. To paraphrase a line from Tolstoy, happy families are all the same, but each unhappy family is unhappy in its own way.

If I tell a story that begins, “The funniest thing happened this morning. I woke up, went into the bathroom, and brushed my teeth,” I doubt that any co-tellers within range will be inspired to join in. The story is much too routine to be tellable. On the other hand, if I open with “I started to brush my teeth and the toothpaste tube was full of fire ants,” co-tellers would jump into the story with their views as to how and why such a thing could happen.

The occurrence of a low probability event is high in information – it is ‘news’ – and around this simple fact is built much of literature, the daily paper, ethnographic method, and sensemaking. Sensemaking becomes the way that its members notice and explain those events. This higher frequency of surprises is a major reason why models for organizational complexity have grown in popularity. The models – sensemaking among them – are an organizational survival strategy for an increasingly connected world.

Where do we see tellability in discourse? ‘Tellability’ links to traditional discourse notions of ‘given’ and ‘new’ information. Discourse has to reach the right balance of the two, in sequences of given and new, where the ‘given’ at time 2 might in fact have been the ‘new’ at time 1. Too much ‘new’ and the discourse is incomprehensible; too much ‘given’ and it’s boring.

In a group who have a history, shared experiences will have produced a local ‘language’ together with the connected assumptions and shared background required to understand it. I call this mixture ‘languaculture’ (Agar, 1995). Given and new are measured against this shared background.

The proper mix of given and new is a serious issue for organizational sensemaking. ‘Given’ in one corner of the organization might well be ‘new’ in another. Tellable in one corner might be boring in another. Sensemaking will vary depending on the languaculture that a group of co-tellers share. Fonseca and Lane, cited earlier, both emphasize that a group has to be similar enough to make sense to each other, but different enough to expand the space of possibilities in the process of sensemaking. The right balance must be struck.

Now let’s look at ‘tellable’ from a different angle. Say that a surprising event has produced a tellable story, where tellable means ‘news’ to all the participants. And say that a group of co-tellers sets out to collectively make sense of it. Is it also a tellable topic? Is it is something one dares say out loud? Will the organization allow sensemaking at all?

It is now a cliché, given what we’ve learned from giants such as Foucault and fields such as critical linguistics (Fairclough, 1989), to say that there are limits on what may be said. Ideological walls around the tellable buttressed by power are as much a part of an organization as they are of a society. Boje builds his book, cited earlier, around ways to make those walls visible by asking questions of narratives produced inside of them.

But what are the political consequences of using his techniques? What will be my reception in an organization if I master and apply them? That of the whistle-blower? Protected by law but disliked by all? Fired but then reinstated after a long court battle? Methods for shattering those Foucauldian barriers are available, but the personal consequences deserve some thought. Are ideological barriers so strong that sensemaking requires a teller to charge single-handed into the valley of organizational death?

Now let me bring the issue down to a more interpersonal level. Can a teller take the tellable ‘news’ and explore it with co-tellers in any way he or she chooses? If topics s/he wants to add to the sensemaking stew are more personal, more damaging, more harmful, is s/he willing to say them in front of others in the organization?

Trust is critical here (Kramer & Cook, 2004). In an organization, trust among a group of potential co-tellers increases what can be told and enriches the space within which sensemaking can emerge. But consider that a person might notice something that is not supposed to be noticed, talk about something that is not supposed to be talked about, offer explanations that no one else has uttered, and come to a conclusion that might well be that the boss should be fired. Why would anyone say anything like this at all? Why take the chance? Aren’t we talking about the whistle-blower’s fate? Only in addition to alienating colleagues and marginalizing yourself, you give them ammunition to destroy your career?

We know a few things about trust. We know that it is more likely to develop in iterated rather than one-time games, and certainly group life in an organization is iterated (Axelrod, 1997). A scheduled sensemaking session with people who have never dealt with each other before and never will again would guarantee low trust, by this account. A group that has worked together for awhile is more likely to have established trust than a team of strangers, though they may have established plenty of negative things as well, like a motive for revenge.

There is some evidence that a group bonded under shared circumstances of risk and uncertainty will yield a higher degree of trust as well. We know trust is supported when all parties see some purpose or value in maintaining the relationship. A shared incentive whose attainment benefits all would increase trust. But what kind of incentive?

Tellability turns out to be a many-splendored thing. It first of all links to the discourse concern with the proper distribution of given and new information. More than this, it links the ‘new’ with the unexpected, the surprise, the out of the ordinary, as featured in both the organizational development work based on complexity and in the linguistic work of Ochs and Capps. But then once a low probability event opens up a space for multiple tellers to do some creative sensemaking, we need to worry about the boundaries around that space. How open is it, really? We want new stories to emerge. Can they? Are they allowed to?

Two major constraints here are the ideological limits and the limits set by trust within the group doing the sensemaking. The first constraint is buttressed by power, which boils down to direct control over one’s job. The second is sanctioned by others in the group – are they likely to use what they learn to advance their own interests and/or damage the standing of others? I don’t mean to be negative here, because the notion of innovative stories emerging from collective telling has strong appeal. But experience tells me that the two constraints of organizational ideology and low trust tend to keep things inside traditional topic frameworks, and as a result keep things pretty dull and uncreative. The range of stories that can be told, the set of the ‘possible tellable’, is pretty limited, its walls are pretty thick, and on the other side of those walls you’re likely to get slapped on the wrist, or worse. Careful listening, or reading of a transcript, will reveal whether sensemaking is able to merge with the tellable, or whether the tellable is buried in old information or ruled out of bounds by ideological or interpersonal fears.

Now to embeddedness, the third dimension of narrative. At one end of the scale, a narrative is a special performance. The narrative is the event – people gather around to hear the expert tell the story. But as we move to the other end of the scale, the ‘living narrative’ end, a story might come up while a group is doing something else. The narrative emerges as part of some other event, plays a role in it, and when it ends the event goes on. Living narrative is part and parcel of the flow of life, unscheduled, often about the very thing that people are doing when they tell it.

For organizational sensemaking, a story is highly embedded by definition. People in the organization build stories to make sense of unexpected events that impact on the organization. Sensemaking is part of organizational life, not a separate special event. The most important narratives from an organizational development point of view are embedded narratives.

A concept from the tradition of anthropological linguistics helps see the embeddedness of a story. That concept is ‘indexicality’ (Duranti, 1997). Indexicality ties the surface of language to context – the people involved, the tasks they are engaged in, their shared experiences in the organization, virtually anything available in the moment when discourse is produced.

The most obvious kind of indexicality occurs with so-called ‘deixis’, the way that pronouns like ‘this’ or ‘that’ point directly to context, or the way that words like ‘bigger’ and ‘smaller’ presuppose an established frame of reference – “bigger or smaller than what?” – or the way that phrases like “you know” point to immediately relevant and shared background knowledge. But indexicality can also mean the way that language points into the realm of any ‘given’ information that a speaker assumes that others have, any background knowledge that results from shared experience in the organization, or shared experience outside it. Here we link back to the discussion of ‘given’ and ‘new’ in the previous section.

A major problem with trendy organization models brought in from the outside – TQM, BPR, Six Sigma and the like – is a lack of embedding. An outside expert comes in and presents a ‘new’ narrative of the traditional sort, a narrative at the smooth and polished end of Ochs and Capps’s dimensions. People gather around to hear it. But then those same people need to embed the narrative in that particular organization. What might that smooth and polished narrative mean here? People in the organization have to work it out later, if they are motivated to work it out at all. More often than not, the original polished narrative changes dramatically, if it survives at all. Some short term results will occur, probably from a Hawthorne-type effect. Some new rhetoric from the alien narrative enters organizational discourse. But the narrative probably does not change the organization much. Instead, the organization changes the narrative as it embeds it. It puts the outside narrative to uses that might in fact contradict the goals of the outside expert who first presented it.

Embeddedness describes a number of different connections between an organization and sensemaking. At the narrow end, it means the way specific discourse particles point directly to the setting where the speaking is underway. At the broad end, it means the way the discourse indexes a shared background for its understanding. But in all its uses, embeddedness asks, did the sensemaking arise as part of the flow of an event? Once again, an outsider can tell from the details of the language through which sense was made.

The fourth dimension of narrative is linearity, a concept that emphasizes the causal links that glue traditional narratives together. There is a famous line from Chekhov – if a shotgun is hanging over the fireplace in Act I, it had better go off by Act III. Chekhov’s message is, everything that goes into a story must play some role, from complication through development to resolution. Point to a scene in a story and ask, “What is this doing here?” and there must be an answer in terms of the story as a whole. A narrative, the traditional kind, is a finely crafted thing.

Ochs and Capps compare the smooth linear flow of traditional narrative with how living narrative slips and slides all over the place. The conflict, they write, is between ideal narrative – fixed and polished – and the open-ended and contingent narrative of everyday life. They introduce the concept of ‘sideshadowing’, something typical of living narratives. Sideshadowing occurs because the tellers want to pack in all of the experience from the situation that is the source of the story rather than to edit and re-shape experience into a polished narrative. Sideshadowing involves going ‘off on a tangent’, from a traditional narrative point of view, but of ‘covering everything that happened’, from the point of view of those who were there.

What does this mean for sensemaking? As Ochs and Capps suggest, we certainly don’t expect sensemaking to be linear. Yet, as it turns out, there are other kinds of logical links besides time and cause.

As one example let me return to my checkered past, when I worked with a colleague in artificial intelligence, Jerry Hobbs, on a life history interview I’d done in New York. One of our articles was called “The coherence of incoherent discourse” (Hobbs & Agar, 1985). It foreshadows the use I am making of Ochs and Capps’s book now, but with a slight twist. The twist is, what they would call nonlinearity still shows connections when you look closely at a transcript, utterance by utterance.

Jerry and I used the concept of ‘local coherence’ to show how this worked. The term means that two adjacent segments in discourse have some kind of connection, even if they don’t satisfy Chekhov’s shotgun-over-the-fireplace test, even if they don’t play any clear role in the model of a linear narrative. Besides time and cause, there are other local coherence relations. Some of the most important are “explanation, semantic parallelism, elaboration, exemplification, contrast, (and) background” (Hobbs & Agar, 1985: 221).

In our article, we used local coherence to show how a story worked in spite of its apparent wandering. We looked at how topics were negotiated and at what we called ‘associative slides’ when tellers wander off onto a tangent. We found discontinuous structures when tellers return to completed sections, and we also saw what I’ll call ‘topic emergence’ as tellers came up with a new high-level orientation to what the story was about and therefore changed its trajectory.

The point of our article was twofold. First of all, there was more to the potential ties among elements of sensemaking than just time and cause. Nonlinear does not mean non-coherent. The second point was that openness and contingency were well served rather than limited by local coherence relations.

Clearly, sensemaking will lie at the nonlinear end of Ochs and Capps’s scale. The old cliché ‘brainstorming’ means to legitimate just this kind of talk – no holds barred, free associate, don’t worry about the point, just let your mind go and try and come up with some new ideas. But much of Hobbs’s and my work on local coherence shows that ‘nonlinear’ doesn’t mean ‘random’; that close analysis of a transcript shows a link between one utterance and another. As psychiatrists have known for years, ‘free association’ makes a lot of sense.

Local coherence is not always explicitly marked with a discourse particle. But sometimes it is. One kind of mark is those grammatical particles usually called ‘conjunctions’, particles that show the logical connection between one sentence and another – ‘and’, ‘but’, ‘however’, and the like. Other particles signal logical flow, like ‘anyway’ to mark a return from an aside back to the narrative line the speaker means to follow. Numerous other phrases can play a logical connector role. “So, as I’ve just shown you” or “let me give another example” are simple cases in point. And other phrases or clauses can do the work of particles like ‘anyway’, such as when a speaker says something like “And your point is?”

Sensemaking will lie towards the nonlinear end of Ochs and Capps’s scale. But to achieve its goals, sensemaking will unfold with competing locally coherent sequences, a mix of linear and nonlinear, and it will offer a sequence by one teller that will receive critical and complicating responses from the others. Sensemaking isn’t just brainstorming, though at times it will resemble it. It is brainstorming constrained by and evaluated against an unexpected event that a group is trying to understand, explain, and act on the basis of.

An analysis of nonlinear local coherence will reveal the paths, taken and missed, and the way the sense-makers arrived at a final destination. It will show a nonlinear logic of sensemaking.

The final dimension of the Ochs and Capps narrative model is moral stance. In the traditional narrative, there is usually a “moral to the story,” though that phrase has come to mean most any lesson one takes away, be it moral, practical or anything else. What Ochs and Capps mean is more the traditional idea of morality, the nature of the good – the just society, right action, the values one lives by, the ‘ought’ rather than the ‘is’.

The problem is that the ‘evaluative structure’ of discourse is even harder to pin down than linearity was in the preceding section. As Charlotte Linde put it in her analysis of life stories:

“Unlike the other components of narrative structure, the linguistic forms that express evaluative structure cannot be specified simply, since evaluation can be indicated by a wide range of linguistic structures and linguistic choices” (Linde, 1993: 72).

Still, there is no question that sensemaking will reveal moral stance and disputes over which moral stance is correct.

Adjectives carry a major moral load, as in the old jokes of the form “I am dedicated, you are stubborn, he is pig-headed.” But many other ways to express moral stance are available as well. Actual cases of sensemaking can be analyzed to show their moral stance(s), and we expect sensemaking to show a variety of them. As a group of tellers work to build a sensemaking fragment that explains a surprise, different moral stances come under discussion as to what the surprise meant, why it happened, what it means for the organization, and what to do about it.

It seems immoral to allow a section on moral stance to be so brief compared to earlier discussions of the other dimensions of a living narrative, especially given recent spectacular cases of immoral organizations, both private and public. But as Linde wrote, the first four dimensions of narrative link to specific aspects of discourse. Moral tone, on the other hand, pervades those aspects and most all of the others.

Let me review where we’ve been. This article began with a problem: terms like ‘narrative’ and ‘story’ are pretty confusing for a person who grew up in linguistic anthropology, where both those terms have been used in a variety of ways for a century or so. The confusion gets in the way of the goals of this special issue, not to mention a current interest of mine, namely, to wonder what language would look like inside an organization that sees itself as participant in a complex coevolutionary system.

I tried to fit the terms to the goals with the following steps. First of all, Peter Lynch’s (1994) popular use of ‘story’ to guide new investors served as an informal and accessible example to narrow the focus, and Weick’s (1995) concept of ‘sensemaking’ brought story construction as a process into the realm of organizational research and practice. We ended up with sensemaking as a particular form of language, an in media res story that used narrative time and cause to make sense of a surprising event.

The next step took us into the recent work on ‘living narrative’ by Ochs and Capps (2001). Their five dimensions of narrative helped place our more precise notion of sensemaking in a larger context that showed what it was and how it differed from other language forms, including the more traditional idea of narrative as a polished and coherent performance.

In the final step, we looked at some of the specific devices from the tradition of discourse analysis that allow us to listen to an actual conversation or look at a transcript and evaluate whether or not it was in fact an example of sensemaking. That exercise opens a relationship between discourse analysis and organizational research and practice, particularly for those interested in evaluating discourse for its fit with ideas about the organization as a complex coevolutionary system.

The steps were taken quickly to cover a substantial amount of territory. Each of them deserves more treatment and discussion. But with any luck the confusion over ‘narrative’ and ‘story’ with which the article started is a little less severe and the steps taken will serve as a useful guide. If nothing else, the article should show that complexity can be found in the details of talk among people who bring the organization to life.

Should this approach to narrative and story be useful, a next major step will involve linguistic variations among the specific discourse devices described here. Turn-taking – the first example used under ‘tellers’ – always appears, wherever you go, since turns have to be negotiated in living narrative construction. But the way that turn-taking works varies from group to group. Turn-taking in Washington DC, where I lived for many years, is a different breed of cat from turn-taking in New Mexico, where I live now. And shifting from American English to Mexican Spanish involves changes in turn-taking as well. And these examples are just the start of a very long list.

Discourse analysis, of the sort introduced in this article, lays out many of the devices by which sensemaking is constructed. But work remains to show the different shapes those devices take in the context of different languages and different histories. Sensemaking now needs to be considered in light of other fields, such as second language learning and intercultural communication. But that is a job for another day.