A Linguistics for Ethnography

Why Not Second Languaculture Learning and Translation?

Michael Agar

Professor Emeritus, University of Maryland, College Park


Language and ethnography have always gone hand in hand. In this article two kinds of linguistics are explored that seem to have a close relationship to ethnography, namely, second languaculture learning and translation theory. The article shows how the former resembles the ethnographic research process while the latter is similar to the usual ethnographic product. The irony is that neither of these two kinds of linguistics have played much of a role in ethnographic research in the past. By applying them to the authorís past work on ethnographic logic and exemplifying them with samples of applied ethnographic work, languaculture learning and translation help re-think the notion of "method" beyond the usual social science conversations while highlighting features of intercultural communication as it happens on the ground.

Keywords: Ethnography, Linguistics, Second Language Learning, Translation Theory, Applied Research

In late 2005 the Graduate School of Education at the University of California, Santa Barbara, invited me to give a series of lectures. The second lecture was charged with answering a specific question, namely, given an orientation to ethnography, what kind of linguistics would be most useful?

Answers to the ethnography-language question are available, since several links between the two are well-established--ethnographic semantics, sociolinguistics, conversational analysis, pragmatics, and the list goes on. The relationship has a long history that I assume readers of this journal already know, at least in broad outline. But as I thought about the question that colleagues posed, two other kinds of linguistics came to mind. I had never considered them as candidates for a core linguistic focus of ethnography. Until now.

Based on that lecture, I want to explore one language focus that corresponds to ethnography as a process, and another that corresponds to ethnography as a product. For ethnography as a process, we need a model of second language learning, or as I will justify in a moment, second languaculture learning. For the product, we need to think of ourselves as producing a translation. Second languaculture learning and translation are related, of course, and it will turn out they are related in the same way that ethnography as a process is related to ethnography as a product, already a good sign in terms of a language/ethnography fit.

The informal tone of the lecture will be preserved in the writing of this article. Ethnography remains the focus, since it was the lecture topic and since it is at any rate a genre of intercultural communication. The relation between the two will be commented on in the conclusion.

An Intra-Cultural Example by Way of Introduction

Let me begin with an example of how language played both a process and a product role in a recent ethnographic project. Stephen Guerin of The Redfish Group and I were asked by researchers with the California Superior Court system to help them figure out what "performance" meant. They didnít mean linguistic performance, but rather performance in the sense of measurement of use of resources. Since the central office of the courts had just been assigned the job of allocating budgets to the separate counties, their question was driven by more than academic interest. They had to give out money and then defend their decision, and they had no idea of how to do that. Performance measures were to be the answer.

First we went to work with the ethnographic process, which involved second languaculture learning. Note that I say second languaculture rather than second language. I borrowed and modified that concept from Paul Friedrichís notion of linguaculture and wrote a book called Language Shock around it. The reason for the awkward concept is that learning a second language is more than vocabulary and syntax and sentence drills and dialogues about buying a ticket for the train.

Earlier forms of linguistics already included powerful and elaborate systems of sociocultural background, each in their own particular way. But I wanted the "culture" in languaculture to remind us of all of them at once. So, Iíll replace the usual abbreviation for a second language, L2, with LC2 for second languaculture. And what I want to argue is, the process part of ethnography and LC2 learning are the same thing.

So was the court in California in fact an LC2, a second languaculture?

If you do not think a court is a second languaculture to us ordinary people, you havenít been in one lately. Lucky for you, too, as most people donít walk in to celebrate a happy occasion. When we started work in an actual court in the San Francisco Bay Area, numerous "rich points" appeared, right away. Rich point is a term Iíve used for years to label those problems in comprehension that come up and turn into a research focus. Rich points mark an LC1/LC2 boundary, in this case a boundary between the languaculture I brought in with me and the one in play as court work got done.

The most striking rich point: Everyone talked and wrote about "pro pers," often with "crisis" appended at the end. Actually that rich point showed up well before our arrival when we scanned the court web pages and related links.

A pro per (called "pro se" in some states) is a person representing him or herself without an attorney, and there were a lot more of them than there used to be. We worked in the area of family law, so this meant that all those pro pers brought in divorces, arguments over child custody and support, paternity cases, domestic violence restraining orders--cheery light-hearted situations like that.

The proportion of pro pers in the courts had increased dramatically over the previous decade or so. No one knew what to do about it, even though reports had been written and conferences had been held and recommendations had been made. It was the classic case of a state bureaucracy in a situation of knowledgeable paralysis, the situation that explains why our civil service produces such a large number of stressed-out nihilists seeking early retirement.

Why fewer lawyers? Have you asked one for their hourly rate lately? Besides, in the court where we worked, a high percentage of pro pers were working-class to poor. Poverty increases the life stresses that make family law issues more likely to surface and decreases the chance that the people involved in those issues will have money to pay an attorney. A classic economic Catch-22: The less you can afford to pay a lawyer the more likely it is that youíre going to need one.

The term "pro pers" wasnít the only rich point we encountered, just the first and most frequent and most passionately discussed. The languaculture of the courts handed us a much longer list as time went on. We had to learn how the so-called "Domain" database worked and what was in it, what an "order to show cause" entailed, what clauses like "churn the calendar" meant, why a domestic violence restraining order was called "a poor manís divorce," and on and on it went.

But mostly we learned that a pro pers who walked into Family Law for the first time wouldnít know this strange courthouse LC2, what it meant, how it worked, or what one was supposed to do with it. We certainly didnít. A few pro pers--we called them the pro pro pers--had been working the system for so many years that they had figured most of it out. They had done their own ethnography out of necessity, so to speak, like an immigrant to a new country. But even they, or at least the few we spoke with, needed help and slipped up now and again.

Time was much too short to dive deeper into court LC2. But even with the time limits, our brief ethnographic foray into court life revealed an obvious LC1/LC2 boundary that raised stress and frustration levels for both court staff and clients and wasted everyoneís time and energy in exactly the same way over and over again. Pro pers werenít lawyers, but the court was designed for and run by lawyers. How obvious could a translation problem be?

So now to the second link between language and ethnography. Why not aim towards a useful ethnographic product and build a translation to fix the problem? Sad to say, there was another LC1/LC2 rich point that taught us that a translation wasnít going to be easy. That rich point had to do with "legal information" versus "legal advice." Court staff were allowed to convey legal information to clients; but legal advice was reserved for private attorneys, which of course a pro pers didnít have. Over and over again we heard a staff member say to a pro pers, "I canít give legal advice," and we heard the savvy pro pers open a question with, "I know you canít give legal advice, butÖ" It wasnít at all clear where the boundaries between information and advice actually were. Much of what needed to go into a translation fell into this semantic purgatory.

Because of the information/advice problem, it was impossible to build a public translation between LC1 and LC2 that could smooth court/client encounters, whether at the counter in the main office or in the courtroom. We summarized the problem as translating from personal narrative to legal schema and back again. The personal narrative was the tragic story of relationships gone wrong. The legal schema was a framework with very few, very specific slots on a form that needed to be filled in. The narrative and schema were worlds apart and needed to be translated, one into the other. No one had figured out how to do that, except for the pro pro pers whoíd figured it out for him/herself through experience.

In the end we showed how a translation could be constructed if the legal advice/legal information boundary could go the way of the old Berlin Wall. But it had staying power, that boundary did. It was inherited from the days when everyone had an attorney, and it was supported by a system owned and operated by law school graduates as well. Preserving the exclusive role of attorneys as the intermediary between court and client of course makes no sense in a time when a vast majority of clients walk into the court without one. In spite of that, the obvious solution was impossible. Such are the annoying politically-supported contradictions you run up against when you do projects in the real world.

But LC2 learning did model the process, and translation did name what could have been a useful and interesting product.

The California court example provides a good introduction to this essay. It is narrowly focused and familiar to English speaking readers, especially if they have been through a divorce. It is useful because it shows how LC1/LC2 contact produces rich points that then trigger language learning. It shows how translation between LC1 and LC2 is the point of it all, with certain meaning problems foregrounded and explained by different contexts. LC2 learning and LC1/LC2 translating arenít just helpful as an adjunct to doing ethnography. They are the major share of it.

Ethnographic Process as Second Language Learning

Letís take a closer look now at ethnography as LC2 learning, and then in later sections weíll look more closely at the product, translation.

An ethnographer and his/her audience probably share an LC1. Thatís usually the case, but certainly not always. An ethnographer might actually be a native of LC2, as in what they call "indigenous ethnography," where the former "native" subjects of American and European ethnography do their own work at home and point out how the LC1 foreigners saw things differently. Or an ethnographer might be a native of yet another LC, LC3, and then set out to learn both LC2 and a LC1 and build a translation bridge between them. Many variations on the translation theme exist, especially nowadays.

For now keep the LC mix simple and just think in terms of LC1 and LC2, with LC1 the languaculture of both the ethnographer and his/her audience. Notice that LC1 and LC2 could be as different as my version of American English and a dialect of Western Balochi spoken by opium farmers Iím working with in Afghanistan. Or LC1 and LC2 could be as similar as the example I just described, my version of American English and the language spoken at work by clerks and judges in the Family Law Department at a California court. Any two LCís will be similar to some extent and different to some extent. The degree of similarity and difference will vary dramatically, depending on which two LCís are brought into contact.

The differences may be great; or the differences may be small. Whatever the case may be, you recognize the LC1/LC2 border when contact produces rich points. Again, rich point is shorthand for the surprises that come up in ethnography, the moments when expectations are violated and an ethnographer wonders what that moment that he/she didnít understand was really all about. Rich points are the foundation of the ethnographic enterprise. They are also foundational for LC2 learning. In Van Lierís framework, rich points are a disturbance, a perturbation, in the LC1 ecology. Some new semiotic species appears in the environment to which the LC2 learner must adapt .

Say a rich point occurs and it is taken seriously. Taken seriously means that the logic so characteristic of ethnography is called into play. "Abduction"--a core part of that logic--means taking the unexpected seriously and trying to figure out what might explain it . It means taking the surprising rich point and thinking creatively and pursuing new information to explain why it happened in ways that werenít at first understood. The logic drives both LC2 learning and ethnography. Once again, ethnography is languaculture learning. (More details on the logic are available in the published version of the first lecture at http://www.qualitative-research.net/fqs-texte/4-06/06-4-36-e.htm).

But not all rich points are created equal in terms of their impact. A rich point might be trivial. An example I used in Language Shock was a simple translation problem for a boring object. An "ashtray" in English is an "Aschenbecher" in German is a "cendrier" in French is a "cenicero" in Spanish. There might be some different associations in there to play with to make it a little more interesting. Not many, though. The words are mostly about "ash." This is a simple rich point well within the bounds of traditional language learning. You donít know what ashtray means? Look it up in the dictionary.

Compare such trivial rich points with those that are more like an epiphany. My dictionary defines an "epiphany" as a "moment of sudden revelation or insight" and links it with religious experience. I mean it in the secular sense, though, like religious conversion, it will have an impact at several different levels and across many different domains.

How does an epiphany-like rich point act? It "percolates," from the Latin for "strain through." Maybe "diffuse" would be a better term. Percolate might sound too slow. But an epiphany feels powerful because it will have profound consequences, though it will take time to work through what they are and what they mean, bit by bit. A hurricane passes through in a few hours; its consequences work themselves out for months. An epiphany will percolate. It will strain through the nooks and crannies of the original LC1 and work its magic to change it into something that works for LC2 as well.

Percolation is what gives an LC2 learner, or an ethnographer, the sense that a new phase of learning has begun. Any reader who has seriously tangled with an LC2 knows the feeling. At first you learn a lot, fast, but then you hit a wall. Then one day, somehow, youíre through that wall, but then youíre worse at LC2 for awhile than you were before, because you need to reorganize what you thought you knew to fit what youíre learning now. And what youíre learning now is taking you closer to communicative competence in the LC2 world. How do you know? It is obvious from how the LC2 world reacts to you and how you feel, as I like to romantically put it, as the flow from your brain and your heart to your mouth and ears smoothes out. Some things that previously felt like a lot of effort now flow without conscious thought. But then you hit a wall again, and the process repeats, and then youíre even closer to communicative competence than you were before. And the process repeats again, and again, andÖ

Doesnít this sound like ethnography? Of course it does, because ethnography and LC2 learning, one more time, are the same thing. In fact, some British colleagues developed a program and wrote a book around just that concept, Language Learners as Ethnographers . Iím proud that their book, and the program it described, was based in part on my work.

A couple of ethnographic moments from my own checkered past will show epiphany-like rich points in action, rich points that percolate and take me over a wall and into a deeper understanding of the LC2 Iím trying to learn.

Remember the California court example I described earlier? When I talked with a few pro pers and court staff, right at the beginning, several mentioned the letter. "What letter," I wondered, "why was a letter such a big deal to both staff and clients?"

Recall that pro pers represent themselves without an attorney. Most donít know much about court procedure. Probably most have the idea that they need to tell their story. And the way many tell their story is to write a letter to the court when they file for divorce, or they mail it to the judge after they get a court date. The letter, as far as the staff and the judge are concerned, is irrelevant. What they want are the proper legal forms, properly filled out.

I sat in a courtroom and a young man stood up as his hearing began. He held each arm high in the air like a cheering fan, a hand-written page in each hand, and said, "Your Honor, I wrote this letter that explains everything." Later that day I spoke with an experienced pro pers who had been learning the system for years and told him the story. He smiled. "When I started, mine was seven pages long."

Such observations and stories, right at the beginning of the project and continuing throughout, turned into a rich point that percolated. It signaled the difference between the aggrieved person who wanted to tell his or her story and the court official who wanted a form filled out in just the right way. I called it the narrative/legal schema conflict and it and its many implications organized much of our final report. In fact, I started the final report by writing that family law was my baby done left me versus order to show cause.

What about foreign LC2s, the usual thing we think of when we think of second languaculture learning? I had lived and worked many years in Austria before it hit me, this epiphany. The concept of "Schmäh" was extraordinarily rich and the usual translation into English as "joke" was way off. I decided to take the concept seriously and it turned into an epiphany that not only led me deeper into Austria but also inspired me to start my book Language Shock. Schmäh was a term I had treated as a minor rich point years before. But when I took it up again, it turned into a rich point that percolated, not only across the LC1/LC2 boundary but also deeply into my life.

Schmäh was used in a variety of ways. The heart of it all was a fundamental premise of Viennese LC2--Nothing is what it seems, what it actually is, is much worse, but thereís no point in getting upset. You might as well joke about it. Once I figured out this simple premise I broke through a LC2 wall and saw people and events differently, and I used my Austrian German in ways that better fit the flow of conversations. The problem wasnít in the content of the premise--it was pretty simple. The problem was in appreciating how deeply and widely it percolated and changed so many details of my LC1 as I went into LC2.

LC2 learning parallels ethnography in ways that ethnographers havenít talked about nearly enough. In both cases, learning isnít about an isolated skill that one tackles, masters, and then applies when one is done. Instead, it is repeating, cyclical, with each round taking one deeper into new material that wasnít even known to exist before the rich point that initiated the cycle. LC2 learning and ethnography resemble Hegelís dialectic, and in that sense both are Hegelian, or historical, sciences rather than Newtonian ones.

LC2 Learning as a Process

Second language researchers know that LC2 learning isnít a linear, all-or-nothing proposition. They know that a question like "Do you speak X" makes no sense, because there isnít a simple yes or no answer unless youíre completely fluent or completely ignorant. Most learners of languaculture X move around somewhere in between those two extremes. They learn, talk for awhile, then learn something that takes them to a higher level. And this happens more than once. In fact, it probably keeps happening as long as they live. There will always be something new to work out at the LC1/LC2 boundary, no matter how fluent they think they are. If nothing else, the boundary will have changed.

When you know some LC2 but arenít fluent, when youíre in that in-between space--this situation is called "interlanguage" . A learner is on the way but still somewhere between LC1 and LC2. But the discussion of rich points that percolate across the LC1/LC2 boundary shows that there isnít just one interlanguage. There will be several. A learner jumps from no LC2 to a little, moves along for awhile adding a bit here and a bit there, and then jumps to a new level after an epiphany-like rich point percolates. Those jumps happen over and over again, until the LC2 student calls a halt or drops dead. Several jumps, several quantum leaps, through several different interlanguages. Each leap results in a new interlanguage, one closer to LC2 than the previous one. Learners build LC2 as they go.

Restructuring LC2 from one level to the next involves more than traditional grammar and lexicon, as the concept of languaculture already makes clear. In fact, tinkering with specific rules is probably the easiest part of the change, though a few of those rules might be rich indeed. More difficult to handle, for the learner, are the percolating rich points that show how moving closer to LC2 will change who you are. LC1 is bound up with identity and competence in everyday practices. Moving away from LC1 weakens who you are and threatens the famous self-esteem. Who wants to look like an incompetent babbling idiot? A shift from one interlanguage to another to yet another takes you ever more distant from LC1 home base into new and previously unfamiliar LC2 identities and practices with which you have less, or perhaps no experience.

A little tinkering isnít a problem. But rich points that carry a learner from one interlanguage to another are. I think of them as an interlanguage wall, with apologies to Pink Floyd and marathon runners everywhere. Sooner or later a rich point does come up that offers an epiphany. Professional ethnographers seek these moments. Many second languaculture learners panic and retreat when they encounter them, but if they aim for communicative competence in LC2, they have to deal with them sooner or later.

When I give ethnography workshops I use two pictures in the powerpoint presentation. The first shows a cartoon figure slumped over a desk. Fatigue, depression, despair--the very image of the newcomer to LC2. The second shows a cartoon figure standing on the bow of a sailboat. Exhiliration, at one with nature, smooth sailing--how you feel when you make the leap into an interlanguage closer to LC2 than you were before.

The pictures work for both LC2 learning and ethnography, because--once again--LC2 learning and ethnography are the same thing.

The Ethnographic Product as Translation

Letís shift now from LC2 learning to translation. Recall that the question that my colleagues posed was, given your way of thinking about ethnography, what kind of linguistics would be most useful? I divided the question into two. For ethnography as a process, LC2 learning is the place to go. Now I would like to shift to the second half of the question. It has to do with ethnography as a product. The product is a bridge between LC1 and LC2 that an ethnographer builds and puts into public view. It is a translation between LC1 and LC2.

An ethnography is a translation writ large.

By translation writ large I mean that languaculture is in play, not just language alone in the sense of syntax and vocabulary. If I were writing my field methods book The Professional Stranger now, I would call it The Professional Immigrant. Ethnography is what immigrants do to figure out how to use a language in their new world, and using a language in a new world is more than conjugating a verb and looking up a word in a dictionary. An ethnography is a bridge between LC1 and LC2 worlds, just like a translation is.

Translation seems to have been a non-field for most of its life, a very long life when you consider that Sumer, the first civilization that left a written record, was already multilanguacultural. Someone had to have been translating. This long silence reflects an outsiderís view of a translator as someone in the background who transforms one language into another. Keeping translators out of sight preserves the delusion among monolinguals that translation is no problem at all, a simple mechanical task. A monolingual can believe that two texts, original and translated, mean exactly the same thing. A translator is just a sort of secretary who speaks in tongues. "Here, Iím going to lunch, put this into Mandarin for me before I get back, will you?"

Anyone halfway experienced at tacking back and forth between LC1 and LC2 knows better than that.

Translation is one of the most difficult jobs in the world. I hate it, especially if it is simultaneous interpretation. It is an extremely stressful job and quickly wears me out. And it feels unnatural. Living in LC1, fine, it is where I grew up. Working my way into LC2 and staying there, fine. I always enjoy the different view from LC2. LC2 makes life more interesting. But flipping back and forth between the two, it feels somewhere between flaky and promiscuous. When I reviewed a German language book for an American journal, I described it as a feeling that I had strained a rich stew through cheesecloth before serving it up.

George Steinerís book After Babel brought the formerly invisible translator out of the closet. Steiner wrote in detail of the richness and subtlety of the work, including an argument that figuring things out in your own LC1 was already a form of translation. And he showed how a translation could twist, distort, and impose outside interests to make it look like something the original material didnít intend at all.

He takes seriously the old translatorís proverb, "traduttore, traditore." I still canít find where this chestnut came from, but it means "to translate is to betray." It is Italian, so who knows, maybe Machiavelli? A Pope? Berlusconi?

Encounters since I moved to Northern New Mexico are my most recent examples of the Italian proverb. On arrival I thought I could use Spanish and it would be a welcome gesture. But until 150 years ago, the area was part of Spain and then Mexico. After General Kearney rode in and took over for the U.S. in 1846, the descendants of the original Spanish were promised their land and rights, which of course turned into the opposite story, as conquests usually do. As a result, the LC1/LC2 boundary between English and Spanish in these parts carries a heavy historical and political load. Early on I had the feeling that local Spanish speakers didnít like it if I used the language. Recent immigrants from Mexico were and are delighted when I speak Spanish. But with local descendants of the original Spanish/Mexican families, Iíve learned to be cautious. For many native New Mexicans, my use of their language echoes the Anglo invasion of their world. I could go on and on with more context and evidence that has accumulated since the original lecture in late 2005. The "traduttore, traditore" theme is holding up very well indeed.

Ethnographers may not know the Italian proverb, but they have the same problem.. Any ethnographer worth his or her salt has obsessed over how some of what he or she translates from LC2 was none of an LC1 outsiderís business. And he or she also knows that what was conveyed at the end of the work--the product--was just a fragment of what was experienced along the way. Perhaps the only thing worse than a betrayal--traditore--is a superficial betrayal. Give the inside scoop on LC2 to LC1 but leave a lot of LC2 out. A cross between the National Enquirer and Readerís Digest. It is the worst nightmare. The fact that it is the same nightmare for ethnographer and translator, though, is a sign weíre on the right track as far as a lingustics/ethnography match goes.

Getting the Translation Right

Since Steinerís book a substantial literature on translation has blossomed and then grown like wildfire . It is an area in which I was never trained as a lingustic anthropologist. In fact, the question my colleagues asked makes me wonder why linguistic anthropology hasnít been littered with references to translation studies for decades. We do have a better record in second languaculture learning, though not by much. I can only guess that both fields--translation and LC2 learning--are too applied for the academic purists.

A core concept in the translation field is "equivalence," and the debates center on just how one tells when one has achieved it. One major conflict is the well-known difference between word for word and sense for sense translation. In other words, does a translator slavishly follow the original text, or does he/she go with what the text conveys to him/her and reformulate it into a translation, even if the language doesnít match up item by item? What equivalence should a translator be faithful to?

For example, how do I translate the Spanish "meter la pata?" Word for word it means "to put the paw (into something)." First of all, it makes sense to hear paw as something human, not animal, as in the English "get your paws off of me." Except "paw" in that English clause means "hands" and in the Spanish expression it associates more strongly with "foot. At least I think it does. And "foot" carries the sense of getting caught in a trap, which Iím guessing is part of the association the phrase means to call up.

Spanish often uses an article rather than a personal pronoun, so "put the paw" means "put oneís paw." But then it doesnít really mean "put your paw (into something)." It is more like--in my experience--get involved in a mess, or start up a hassle, or commit a faux pas. The last time I heard it was at a dinner in San Pedro Sula a few years ago, in Honduras. A guy arrived late with his foot bound up in a cast, walking with a cane. "What happened," I asked in Spanish. Another guy at the table said, "metió la pata," "he put his paw (in it)." Everyone laughed long and loud.

The preceding paragraph took a lot of words just to translate one line at a dinner. Where along the way would I get equivalence in English? My take on it, given that it was a bunch of guys talking raucously, would be to say the joke meant to play with the double meaning--literal and metaphorical. How could I build a double meaning in English that ties "he messed up" with "he sprained/broke his ankle/foot" to get that translation right? Maybe I could have said: "He put his foot in his mouth while chewing gum." This is not word for word, but it is a better equivalence.

So how does an ethnographer, or LC2 learner, know when a translation is equivalent? What counts as LC2 equivalence in an LC1 world? At this point the conversation among social researchers usually shifts to issues of methodology. If you collect and analyze data using the right methods, then the translation is equivalent. Only now that weíre thinking of translation, you see the problem right away. An elegant display of social science methodology means the translation has lost most of its relationship to both LC1 and LC2. It probably wonít make sense to anybody except another social scientist. Thatís not equivalence,.

Equivalence evaluates the relation between two different languacultures. The relation is evaluated by comparing the original and the translated version. In the jargon of the field, one speaks of a source of and a target for the translation. It does not make sense to say "a translation of a Magritte mystery" without knowing for whom the translation was done, and it does not make sense to say "a translation for the Americans" without knowing what it is that was translated.

What does all this mean for ethnography? The ethnographer is a translator. He/she experiences rich points and comes to understand them. But then what? An ethnographer needs to foreground problems in meaning--the rich points that represent significant problems in understanding between source and target--and then explain those problems in terms of differences between LC1 and LC2, in other words in terms of context.

Here is where the concepts of meaning and context come alive. Meaning names a problem that is foregrounded, a rich point, a difference that emerges at the boundary between the LC2 source and the LC1 target. The meaning problem will depend on which source and which target are linked.

The ethnographer/translator will then provide material to explain the rich point, why it is there and how to understand it. This material that explains the rich point is context from LC2, the material that was missing from the LC1 world such that the rich point was a problem in the first place.

Meaning is the problem; context is the solution. A translation is a collection of LC2 meaning problems for LC1, the rich points, together with LC2 contextual material that shows how the rich points can, indeed, make sense, from an LC1 point of view.

But note once again--a translation makes no sense unless we specify a source and a target. The meanings that will be explained with new contexts for the target audience depend on which LC1 target and which LC2 source are bought into contact. Change the source/target pair and you change the translation, just like with ethnography.

LC2 and Translation Together Again

Equivalence is a major issue in translation theory. How do you tell if an ethnography/translation is any good or not?

For many years Iíve been a fanatic about ethnographic methods. I was raised in a tradition--ethnoscience and cognitive anthropology--that argued that the lack of ethnographic method was a scandal. It was, too. But strange things have happened to me here and there that I only appreciate in retrospect, things that are less about method and more about consequences of a translation. These strange things have to do with members of a group I worked with saying that what I wrote made sense to them.

Return to the court example with which this article began. We presented our results to a variety of small groups of court staff. Reactions varied in terms of where a group thought the analysis should go in the future, but all agreed that we had made sense of some part of their world. Stories from non-English LC2s work in similar ways, though Iím emphasizing English-to-English examples here for the readerís convenience. I wondered why I hadnít talked about such reactions in the past. I never took them seriously as part of method.

Translators donít convince LC2 people of their competence by describing all the sentence drills they went through, all the conversations they were involved in, all the mistakes they made, so that finally in the end they could speak. They convince them by showing that they can speak. Ethnographic equivalence, at least part of it, is about how well a paraphrase works in the LC source world rather than a recital of the specific steps taken to learn how to produce it.

Here is a link between the process and the product, between LC2 learning and translation. An ethnographer does this kind of paraphrasing all the time. He/she does it during that part of fieldwork that seldom gets reported, during the simple act of being there and talking and listening in all sorts of situations that just happen to come up.

These paraphrases, these reformulations that an ethnographer/translator checks with LC sources, what do they have to do with getting from that LC to the target LC, from LC2 to LC1?

Recall that we made the simplifying assumption that the LC1 target of a translation is also the native LC1 of the translator/ethnographer. Under this simplifying assumption, an ethnographer is working out the translation even as he/she tries out paraphrases among source LC2 people.

He/she is trying out paraphrases with the LC2 source, but paraphrases of a certain type. Our translator/ethnographer is creating an interlanguage paraphrase, one that reformulates the rich point into a shape that LC2 people probably never said or heard before. For them as native speakers, the meaning in question was not a rich point, so no context needed to be included to explain it. Translator/ethnographers have to include that missing context on the way to LC1, their native languaculture, in order to show why the rich point makes sense. They pull the rich point and its context in the direction of LC1, but still say it in LC2 to see if it makes sense in the source world.

Translators take a rich point that they know will be a problem for the LC1 target and cast it into a different shape, a between shape, an interlanguage version, still expressed in LC2. They check with their LC2 sources to see if it makes sense. They have made context explicit to explain the rich point, but the paraphrase should still make sense to LC2 as well. In fact, translators will often get an aha reaction, because they have told the source something they knew, but they didnít know that they knew it, certainly not in the way the translator formulated it. It is creative work. You say something in LC2 that everyone in LC2 recognizes, but they have never heard it put that way before. Youíve othered the rich point and its LC2 context, but in a form that LC2 natives find recognizable and interesting, maybe even enlightening, possibly amusing, now and again annoying or even dangerous. It makes sense of LC2 to LC2 in a new and interesting way. But notice how this half of the equivalence problem is less about method and more about use and feedback and revision.

I organize method displays after Iím already pretty sure Iím right. The displays are more a rhetorical device and a focused test for a social science audience. And the test could prove me wrong if I follow the science rules properly. But the main reason I know Iím right about the meanings and contexts that will organize a translation? It isnít because of social science methods. It is because I can formulate that translation in a way that is different from anything the LC2 source would ordinarily say, in a way that will map onto LC1, and in a way that shows LC2 a different perspective on who they are that--sometimes--elicits an aha reaction from them.

The proof of the translation pudding, the demonstration of equivalence, lies in how well it works, not in details of the procedure by which it was acquired. But how do I check the second part of the translation? How do I know it works with the LC1 target? How do I know Iíve gotten equivalence on that side of the LC1/LC2 boundary as well?

Here is the sad part of the story. Iíve learned that getting the translation right for LC2 is usually easier than getting it right for LC1, the target of my efforts. Why should it be easier to learn LC2 than to translate into LC1? After all, weíre assuming LC1 is the native LC of the translator. Going into LC1 should be the easy part.

It is worth another article to explore this question, because it has less to do with communication and more to do with power. Here I can summarize and say that much of my work sets up the following problem: The LC1 group who pays me already has a bad translation. I come in and correct it. But the bad translation had a history of serving LC1 interests and, more often than not, my new corrected version doesnít serve those interests nearly as well. That history of course is the reason the bad translation was held in place for so long in spite of abundant evidence that it was wrong.

Hereís how it usually goes. (Though please keep in mind that my checkered applied research past is not the academic ethnographic norm.) The LC1 target has power over the LC2 source. By power I mean they control resource flow to LC2, where resource can mean any number of things, including all those different kinds of capital that social theorists write about--material, symbolic, social, cultural etc etc.--singly or in combination. The LC1 target asks for the ethnography. Theyíre motivated because resource control isnít working like it is supposed to and they want to correct the situation so that it does. They hire an ethnographer to go forth into LC2 country and find out how to do that.

The ethnographer, more comfortable than your average professional among those lower on the power scale, will do a fine job of learning LC2 and figuring out how to translate it into LC1. The problem is, the translation will go well beyond fixing the previous LC1 translation in some simple way that satisfies the LC1 goal, namely, to continue to control LC2 resources. Instead, the translation will show discrepancies between LC1 intended use of resources and LC2 consequences. All manner of complications will be produced by linking resource flow with numerous other contexts of which LC1 was unaware. In fact, the translation will raise deep issues of control and use of those resources, and, just for good measure, the translation might well raise questions about why resource control over LC2 should be the province of the LC1 sponsors of the translation in the first place. Worse still, from LC1ís point of view, the translation will work in the other direction, as it should, and make LC2 aware of some things about their situation in an "aha" kind of way. They knew it, but they didnít know they knew it until they talked with this LC2 learner/ethnographer who made the context of what they said explicit with his or her paraphrases.

My experiences suggest an odd measure of successful LC1 equivalence. Successful equivalence of translation for the LC1 target means that LC1 understands it perfectly well, dislikes it intensely, and fires the ethnographer. Equivalence, in the sense of a good translation that bridges source and target, is rejected by the target, more often than not, when equivalence, in the sense of resource control, is skewed between target and source. Instead of "traditore, traduttore," to translate is to betray, we get "kill the messenger."

There are of course more balanced translation jobs that donít involve this kind of skewed relationship. For example, the translation of a Swedish mystery by Mankell into English is something that author, publisher and English-speaking reader can all support, the rare triple win situation for source, target and translator. And there are academic translations, where no resources are really at stake beyond the event itself and the academic career of the translator. Not that they arenít often interesting and useful. Imagine my surprise in Austria when I learned that I could access Soviet linguistics because the East Germans had translated them into my second languaculture. Not that there isnít a political story in there as well, to put it mildly.

But still, the notion of target audience success shares some characteristics with source success. A translator doesnít show the detailed steps by which something came to be known, the usual method mentality. Translators show that, however they learned, whatever they learned, it works. It has consequences that the translator was after, visible in the reactions of the LC2 source, the LC1 target, and the subsequent actions of both.

Translators shift the shape of LC2 into an inter-languacultural form and try that out in LC2 with paraphrases. Then they reshape it into LC1 for their target audience. The question is, did an LC2 rich point make sense to LC1 in ways that preserved what it meant for LC2 in the first place? Paraphrase in LC2 was the first step. Reformulation into LC1 was the second. It all happened through the mediation of the LC2-learning/LC1-native translator. Translators know when they get it right because of LC2 and LC1 reactions. The main evaluation comes out of the consequences of what the translator translated, from both sides of the process, the source and the target.

I want to argue that a translation must satisfy what I call a "triad test." (This is actually an old-fashioned name for a psychometric technique, but I mean something different). The term "triad" refers to the three points of view involved in a translation--the source, the target, and the translator. A translation has to work intersubjectively among all three at once. It has to connect distinct frameworks among the ethnographer/translator, the people studied, and the consumers of the translation under construction.

Most traditional ethnographies fail this triad test, because they are written for other professional ethnographers without much attention to source, target, or other consumers outside the professional peer group. Judging by the stories Iíve heard over the years, the people about whom an ethnography is written often--not always, but often--find it boring and irrelevant to their lives. And not only the people about whom it is written. In one famous line in the landmark book Writing Culture, Mary Louise Pratt asked of ethnography, how can such interesting people doing such interesting things write such boring books?

Translators, on the other hand, have to pass the triad test or they will be fired, since they have to understand and clarify what LC2 speakers mean in ways that convey what those speakers are trying to communicate, and they have to make sense to LC1 consumers who are not themselves capable of understanding LC2. Comprehension by both source and target are the criteria.

A translation must connect, on the one hand with available frameworks in LC2, and on the other hand with available frameworks in LC1. And the actual work of making the connections cannot be done unless the translatorís frameworks connect with those of LC2 and LC1 as well.

Vygotsky, about whom I learned much thanks to education colleagues who were the audience for this talk, writes of the "zone of proximal development," or ZPD. He applied the concept to learning and education. A crude summary might be, one can only learn within limits defined by the state of knowledge when learning commences. What one learns must connect with available knowledge but also develop it into new knowledge. Too far inside the zone and it is boring. Too far outside and it is incomprehensible.

Translating is the same, except translators stretch between LC1 and LC2, so they must know the ZPD of both. And they can only operate in terms of their own ZPD with reference to both LC1 and LC2. The translator, the person doing this stretching, must find places where the rich point ends for LC2 and begins for LC1, using his/her own ZPD. Then the translator builds a context that makes sense to LC2 given LC2ís ZPD. Then our translator uses that context to show how the rich point makes sense to LC1 by adding just the right additional knowledge, given LC1ís ZPD.

This is beginning to sound like a booklet of incomprehensible instructions for assembling a home entertainment system. Now you see why I hate to translate? The triad test just hints at what a massive amount of work it is. And the ZPDs move around even as you unearth and then connect them. One part of a translation stretches ZPDs and therefore makes other translations possible. Just like ethnography, I have to start with the basics and then build on what the basics made possible, then build on that in turn, and so it continues until Iíve finished the translation.

When we present ethnography to audiences who expect traditional social science, we need method. But I now appreciate that issues of method are important for the form of the conclusions, for certain limited audiences, but not for the truth of their content, or not important at all for most of the non-academic audiences I now aspire to reach. Truth lies in the ability to make sense of LC2, both to LC2 people themselves and to LC1, and truth is evident in comprehension, understanding, and the consequences of what is understood. How the food tastes matters more than the recipe by which I prepared it.


Translation, then, is the area of linguistics that makes sense for the product of ethnography, just as second languaculture learning makes sense for the process of doing it. The fact that these connections havenít been taken more seriously--at least as far as I know--amazes me in retrospect.

I guess it shouldnít. I didnít take them as seriously as I should have until now, forty years after my first fieldwork in a South Indian village. Consider the lowly image of the translator or the second language teacher, the person subservient to the two LCs and, invisible in the background. It is a job, a chore, in the eyes of monolinguals. The world of social science inherited from the Enlightenment an obsession with methods to obtain universal objective knowledge. Why would an aspiring science look to lowly job descriptions for clarity and inspiration?

Things are different now. The profession of translation and the concern with second languacultures in our global society is growing by leaps and bounds. As I worked on the lectures on which this article is based, I realized that translation is what I do, what Iíve always done, and second languaculture learning is how I acquire the ability to translate. In fact, my career, scattered across many different areas of study, has always been motivated, at least in part, by the discovery of bad translations, followed by enough LC2 learning to straighten it out. And the motivation usually has something to do with a lack of equivalence in the social and economic sense as much as with the rich points and contexts to explain them.

Though this article focuses on ethnography and languaculture, it is also about intercultural communication. How could it not be? Anthropologist Edward T. Hall is considered one of the founders. Years ago I wrote an article, amazed at the parallels when kismet cast me into an intercultural communication role for a Mexican and American joint venture. All three fields deal with languacultures in contact, however similar or different those languacultures might be. This is not the time or place to do so, but the similarities among these three--and many other fields that concentrate on "cultural differences"--call out for some synthesis and generalization. Perhaps the job has been done and Iím just not aware of it. It would be a job well worth doing.

Thinking of ethnography as second languaculture learning and translation makes perfect sense, to me now at any rate. Both involve the learning of a different way of seeing together with the way to express it--a languaculture--together with a demonstration of that different way of seeing to an audience who didnít previously have it--a translation. LC2 learning and translation theory are the two kinds of linguistics that fit best with ethnography. It is a conclusion Iíve never reached before, but at the moment it sounds like the right answer to the question that colleagues who invited me to lecture asked me to address.


AGAR, M. (1994) The Intercultural Frame. Journal of Intercultural Relations, 18, 221-237.

AGAR, M. (1995) Language Shock: Understanding the Culture of Conversation, New York, Wm. Morrow.

AGAR, M. (1996) The Professional Stranger: An Informal Introduction to Ethnography, New York, Academic Press.

AGAR, M. (2006) An Ethnography by Any Other Name... Forum Qualitative Sozialforschung/Forum: Qualitative Social Research. http://www.qualitative-research.net/fqs-texte/4-06/06-4-36-e.htm

BAKER, M. (Ed.) (2001) Routledge Encyclopedia of Translation Studies, Routledge.

CLIFFORD, J. & MARCUS, G. E. (Eds.) (1986) Writing Culture: The Poetics and Politics of Ethnography, Berkeley, University of California Press.

FRIEDRICH, P. (1989) Language, Ideology, and Political Economy. American Anthropologist, 91, 295-312.

KASPER, G. & BLUM-KULKA, S. (Eds.) (1993) Interlanguage Pragmatics, Oxford, Oxford University Press.

PEIRCE, C. (1906) Collected Papers of Charles Sanders Peirce, Cambridge MA, Harvard University Press.

ROBERTS, C., BYRAM, M., BARRO, A., JORDAN, S. & STREET, B. (2001) Language Learners as Ethnographers, Buffalo NY, Multilingual Matters.

STEINER, G. (1975) After Babel: Aspects of Language and Translation, Oxford, Oxford University Press.

VAN LIER, L. (2004) The Ecology and Semiotics of Language Learning: A Sociocultural Perspective, Dordrecht, Kluwer Academic.

VYGOTSKY, L. S. (1978) Mind and Society: The Development of Higher Mental Processes, Cambridge MA, Harvard University Press.

About the Author

Michael Agar received his Ph.D. in linguistic anthropology from the University of California, Berkeley. He has worked in both university and applied research settings, most recently at the University of Maryland where he is now emeritus. Currently he works independently as Ethknoworks on projects involving a blend of ethnography and complexity theory in social service organizations. He is the author of The Professional Stranger, Language Shock, and the recently published Dope Double Agent: The Naked Emperor on Drugs. This article is based on a lecture delivered at the Gevirtz Graduate School of Education at the University of California, Santa Barbara. The lecture series was made possible by Judith Green and Janet Chrispeels with the indispensable and generous help of Audra Skukauskaite.

Author's address

Professor Emeritus, University of Maryland, College Park
Consultant, Ethknoworks, Santa Fe NM

Journal of Intercultural Communication, ISSN 1404-1634, issue 16, April 2008.
Editor: Prof. Jens Allwood
URL: http://www.immi.se/intercultural/.