Academic Editing in the Humanities

by Kate Wildman Nakai

In the fields of history, thought, art history, religion, and literature, which are the main subjects of articles published in Monumenta Nipponica, the rules of academic writing are less formally structured than in the sciences, creating greater freedom but also greater possibility for uncertainty, a lack of clarity, and poor organization. Researchers have traditionally been guided as writers by their professors and by the models, advice, and comments of mentors and colleagues. Some master the craft more quickly than others. The role of a journal or book editor of a scholarly work will vary case by case but is sometimes crucial to helping not just the inexperienced but even a mature researcher to set his or her writing on a clearer and more accessible track, improve its style, and deal with numerous structural and technical issues along the way.

Refereeing and the Initial Editing Stage
For most academic publications, work is not solicited; it is submitted. Book reviews and occasionally related items will be solicited, but a journal like Monumenta Nipponica, and probably any of the academic journals in the field of Japanese studies, depends on submissions coming in from authors. After an initial filtering process, the submissions are sent out to referees for evaluation—referees being people who have some particular expertise in the subject matter of the article. The same principle holds for books published by academic presses. Refereeing is usually anonymous in that the referee is not known to the author. In the case of book publication, usually—or at least in my experience—the referee knows who the author is but the author does not know who the referee is. For journals, however, we usually use what is called double-blind refereeing, whereby neither side knows who the other is. Of course, it is sometimes possible for either to guess who the other is, but in principle we try to disguise the identities of both.

imageThe referee may point out flaws in a manuscript’s argument, raise various questions, and so on, and in many cases the decision is then made not to proceed with publication of that article. But even in cases where the recommendation is favorable, or cautiously favorable, or the referee suggests that publication is within the range of possibility, almost always there are suggestions for revision of some sort.

So this brings us to the first stage of the editing process. Here the editor usually passes on to the author the referees’ comments, or excerpts from them, and also tries to draw out of those comments some general things that the author should pay attention to. Sometimes I add suggestions myself (although these almost always concern organization rather than content), because referees usually limit themselves to pointing out problems and are less inclined to go into detail about possible solutions to those problems.

Bear in mind that it is often hard for the author to see on his or her own what the problems are. The author has been immersed in the issues and sources that he or she has been working with. There are all kinds of directions in which that material might be developed, and those possibilities are all present in the author’s mind, even as he or she has opted for one route or another. So one of the most important roles of the editor is to try to provide some specific suggestions as to how problems might be addressed and how things might be sorted out more effectively. It often is easier for an outsider (in this case the editor)—who does not have all of the options in mind that the author does and sees only the material at hand and works with just that—to chart a clearer path.

The initial task for the editor, then, is to write back to the author and lay out what the issues are and make some suggestions. In most cases these suggestions are not presented as absolutes that the author should follow exactly but as indications of problems that the author should deal with in some way. In the optimal case, when the article comes back to the editor, things have more or less fallen into place, the author having seen what needs to be done. That is the most gratifying aspect of the initial process, when in the next stage you can see how things have come together and developed from that initial critical appraisal. And from there you move into more detailed editing and copyediting. In other cases, however, even after the first round of revision the author has still not really sorted out the problems, and then the editor needs to go through another round of suggestions.

With this as background, let me give some examples of the kinds of comments that I find myself writing over and over again to different authors and that might be of use to others engaged in editorial work. The issues I’ll focus on are ones encountered not only by in-house editors but also by freelance editors, which many even quite large presses seem to be using more and more. Another kind of occasion when freelance editors might be called upon to deal with some of these issues is when an academic press has suggested that, before formally submitting the piece, the author get editing done, his or herself, in order to deal in advance with some of the problems a referee might react to negatively. So an editor might be working directly with an author who’s trying to improve his or her work prior to submission, or might be engaged by a publisher to do outsourced editing on a piece that has essentially already been accepted.

Typical Problems
The kinds of problems that come up frequently in academic editing are those that apply to any kind of writing, from term papers on up, so what I have to say here is not anything new and startling. Let me just sum up what some of the typical problems are, bearing in mind also that they are all interrelated.

  • One is that the argument or overall purpose of the article is not clear. An article without a clear focus typically has a diffuse organization and seems to ramble. There may be interesting things along the way, but you can’t see how it all fits together and you’re not sure exactly what the author is trying to do with the material being presented.
  • A second common problem is reiteration and digression. The same points or issues come up at different stages of the discussion, or the flow of the argument is interrupted by foreshadowing—mentioning that something is going to be discussed later on—or by digression from the main line of the argument onto a different track.
  • Then there are the two poles—the two evils—of over-generalization and over-specialization. The discussion may be too abstract and general, without enough specific, concrete, illustrative examples to really pin the argument down. On the other hand, the discussion may be too specialized, plunging right into the topic on the assumption that the reader already knows a lot about it, thereby losing readers who don’t have that degree of knowledge.
  • Another problem that comes up frequently is the presence of large chunks of background knowledge that just seem to be dropped into the discussion for their own sake, and which also become a kind of digression.
  • Yet another is too much unfamiliar detail. Particularly when you’re working on a subject in Japanese studies, articles often include lots of names, terms, and other minutiae that many readers may find bewildering and off-putting.
     

Common Causes of Problems
Everybody understands in principle that these are things to avoid. On the other hand, if everybody knows that you should have a clear argument, that you shouldn’t have a rambling discussion, and so on, why do these things happen? Let me try to identify what I have found to be some of the common sources of these problems.

The “reading note” problem
One trap that authors, particularly academics, often fall into is what I call the “reading note” approach to presenting the sources and material being introduced. In the course of researching the subject, naturally authors take notes as they accumulate knowledge about the various issues involved. If they’re working on a specific text—a literary or philosophical text, for instance—or focusing on a particular source, they typically follow it closely. Thus, when they sit down to write the article, there is a strong temptation simply to follow the same track that their research process has taken; in effect they reproduce that process.

The problem with this approach is that frequently it leads to the tail wagging the dog: in other words, the source being examined, which has its own structure and its own dynamics, takes over the article. A source often combines, in an intertwined fashion, elements that pertain to different parts of the argument that the article or book is trying to develop. To use those elements effectively to illustrate and support the analysis at hand, the author thus needs to disentangle and reorganize them. If the author simply follows what the source has to say, the discussion gets pulled into the framework of the source itself and trapped within what the source says. This is one major cause of the problems of reiteration and digression. Once you’re trapped within the framework of the source, your own argument becomes fragmented and is made to fit the structure of the source. Because A and B appear together in the source, you end up talking about them together in your article, too, when really, in terms of the logic of your argument, you should talk only about A first and then deal with B in another context. Letting the source dictate things is also a major cause of digression, since not everything in a particular source is necessarily crucial or fundamental to the argument that the author wants to make. The author should focus only on those aspects of the source that pertain to his or her argument. If you talk about everything in a source, you end up with an annotated commentary, not an article.

The unarticulated argument
A second typical cause of problems of organization is that the author has not really articulated a clear, overarching hypothesis at the beginning of the article or book. Authors usually stumble into this pitfall because the whole structure and how its various parts fit together are not yet clear in their minds, and this lack of clarity is reflected in their presentation. An article or book needs to sort out an overall argument that encompasses, or at least sets the stage for, the major points that will be covered in the piece and provides guideposts for the subcomponents of the hypothesis that will be developed. Otherwise it is easy for both author and reader to get lost.

Overreliance on the conclusion to pull things together
Another source of digression and diffuse, rambling organization is a tendency to set out the relevant information and wait till the conclusion to explain how it all fits together and what it means. This, too, is a result of reproducing the research process in the writing of the article. The author goes through the same stages in the article that he or she went through in the research leading up to it: there is a topic to be investigated, we find out x, y, and z about that topic, but only at the end do we find out what it all means. But in an article or book, you don’t want the reader to follow exactly the same investigative process. Rather, once you’ve reached your conclusion, you want to turn things around and use this conclusion as in fact the organizing “hypothesis” for structuring the presentation of your findings. The article may convey the “illusion” of leading the reader through the investigative process—that’s an effective presentation strategy—but it needs to organize that illusion carefully so that the argument and the evidence supporting it unfold step-by-step.

Unclear priorities
Yet another cause of poor organization is the failure to sort out what is essential detail from what is interesting but secondary. If you haven’t really thought through the structure of the article and seen how all the pieces should relate to the larger hypothesis, it can be very difficult to decide what kind of detail to choose. Among the problems typically encountered, I mentioned having either too little detail or too much. You want enough detail to illustrate your point and make it concrete and convincing, but if you’ve got too much detail and its relevance is not clear, it may confuse more than it clarifies. As the reader can hold on to only a certain amount of information, how that information is presented is crucial to enabling the reader to follow the argument. So, you really need to sort out what is essential detail from what is interesting but not fundamental to the argument as such.

Inability to let go
A related problem that is typical particularly of academic writing is the inability to let go of valuable information acquired in the research process. When you discover some wonderful point, it is very hard to let go of it; you want to keep it in the discussion somehow, even if it disrupts the argument at hand or blurs the focus. And because you are so attached to that piece of information, it can be hard even to see that it is having detrimental effects on the rest of the piece. In one recent case, for example, the author began the article with a powerful, dramatic story, one that drew the reader in immediately. The trouble was that the story was actually a counterexample to the main thrust of his argument, so it set the article off in the wrong direction. Eventually, in this case, after both the author and I had tried to find some other way to use this fascinating piece of information, he simply cut it out. It just didn’t fit in that article and would best be saved for another occasion.

Strategies for Resolving Problems
Next let me mention a few strategies that I have found useful for overcoming problems of this sort. The first one is topical organization. If I had a thousand yen for every time I have written to an author about topical organization, I might not be a millionaire but I would be quite well off. Basically, as I mentioned before, if the organization is driven by the source, and if you hold to the framework of your source, you’re likely to get trapped within it, and that makes for a very flat, “and then, and then, and then” kind of discussion. The same thing happens if you simply follow a strict chronological organization. Obviously chronology can be an important aspect of organizing a discussion, particularly a historical one, but adhering too closely to chronological sequence may lead to repetition—different aspects of the same general event may come up at different points of the chronology—or a flat presentation. If you do rely on a chronological framework, you need to divide it into discrete units, explaining why each delineated period is important or relevant. So, avoid a straight chronological or a straight, source-driven organization; instead, figure out what are the large points that you want to extract from the source material—what’s the overall story you want to tell about them—and organize your argument around that. This is topical organization, and I would call it Golden Rule Number 1.

The second strategy, which goes along with topical organization, is to formulate a clear, step-by-step architectural plan, a blueprint. This blueprint should not necessarily be visible to the reader—just as people living in a house don’t need to know what all its structural elements are as long as the house functions smoothly and they can move easily from room to room, and so on. But the person who builds the house does have to have that kind of blueprint in mind in creating the design. So, when you choose and organize your topics, you have to think about what the foundation is, what comes first, what you’re going to need to have in place for later on, what is better saved for a later stage of the discussion, and so on. You don’t want to tell the reader that this or that piece of information will be important later or that it is going to come back into view; but if you’ve put it in place, if that foundation is there, then when you do come back to it later on, it is already familiar and you can simply remind the reader by saying something like “as mentioned earlier.” Here again, though, you want to avoid weighing the reader down with premature information, because the reader is only going to be able to keep in mind so much information at any one time. If you present too much at once, the reader doesn’t take it in, so you want to be careful to distribute the information throughout the course of the article or book in such a way that the reader can absorb it step by step.

To counter the problem of too much particular detail, as a general rule don’t use proper names or special terms unless they are essential to the discussion. Five people may be involved in a particular incident or a particular literary movement, but if you’re only going to be talking about one of them later on, you don’t have to name all five. You can say that there was a group, one of whom was so-and-so, and put the other names in footnotes. I think footnotes are very useful. I am not of the school that says you should reduce footnotes, and I like footnotes at the bottom of the page where you can see them. I think of it in terms of two levels of readers. You have the ones who do know quite a bit about the subject and want those specific details, and then you have the readers who do not know so much about the subject and find too much detail troublesome and distracting. If you put the extra names and details in the footnotes, they are there for the more-specialized reader, but you are not slowing down your less-specialized reader. At the same time, you want to make sure that any reader can read the entire text without looking at the footnotes. This means you don’t want any information important to the development of the argument to be in the footnotes only; you want to make sure that all the essential elements are in the text.

So, those are my rather obvious general principles that I would try to convey to the author as needed. Assuming the author works with those guidelines and comes up with an improved version of the article, things are much simpler thereafter—though these same general principles may continue to be applied at one level or another throughout the later stages of editing as well.
 

At this point Nakai introduced a break, during which participants were invited to look over a sample illustrating the application of these principles of editing to a specific piece. A number of the questions that followed referred to this sample.
 

Sources dominate the argument
Q: Regarding the problem of authors being driven by their sources, I find that graduate students in the social sciences tend to quote excessively from their sources, often out of context or with little thought about the relevance of the quotations to their own argument. How would you deal with this problem?

KWN: It’s useful to keep in mind that in quoting from a source, whether primary or secondary, an author is free to (and should) pick up points from it in whatever order is appropriate to his or her own argument; it’s not necessary to keep to the order in which the points appear in the original source. The author should deconstruct the source’s various elements and decide which is going to fit with which part of the argument. That requires sorting out your own argument first. The research-deductive process is not the process of writing the paper. You have to turn it around: take the deduction you’ve reached ultimately, put it as your organizing hypothesis at the beginning of your paper, and then organize the material in such a way as to develop that hypothesis. Once you’ve sorted out your own argument you should go back to your sources and figure out what parts of them will illustrate your own points most effectively.

Q: Do you mean that the construction of the argument should be independent of the sources, that it should not be driven by other people’s ideas?

KWN: It’s not really an either-or issue; you need to combine both aspects. What is crucial is how to go about doing that. You have to digest the ideas of others, whether the ideas in your sources or those of previous researchers, but you also have to formulate your own argument. You may need to provide an overview of relevant previous research, but you don’t want to fall into the trap of presenting information for its own sake. If you haven’t sorted out what it is in those previous interpretations that you want to contrast your own interpretation to, you often end up just presenting A’s views and then B’s views and then C’s views—just all these large undigested chunks, each going in its own direction. So you’ve got think it through. Ask yourself: what is it about A’s or B’s or C’s interpretation of this subject that my interpretation is going to be different from? Then organize your presentation of previous research so that it is focused on issues 1, 2, and 3, and what A, B, and C have to say about those issues. You’ve got to take their ideas apart and put them together in a new way. Certainly in many fields you’ll be expected to recognize work that’s been done previously on the subject and to indicate it early in your paper. How you go about doing that is the question you have to keep in mind.

Different norms, Japanese and English
Q: As a translator of academic papers in the social sciences written by Japanese scholars, I too have noticed that Japanese authors tend to use more reiteration than would normally be expected in English text. I usually try to convince authors to omit certain reiterations, but it can be quite difficult because they feel I’m going to cut some information from their paper. Do you publish such translated articles in Monumenta Nipponica, and if so, do you have any experience with this problem and how to solve it?

KWN: We have some translations; also, quite a number of our authors are not native English speakers. We generally try to explain that repetition has a dulling effect in English that it may not have in other languages or contexts—basically pointing out that if something’s been said already you don’t have to say it again. Now, it may be that the author is not saying exactly the same thing again but rather is coming back to the same point and saying something slightly different about it. This can be very tricky to deal with, but essentially you’ve got to rearrange things so as to retain all the author’s major points while ensuring that nothing is actually being said twice. Here is one of the places where topical organization and consolidation of related matters may come into play.

Q: What if you were publishing an English translation of a well-known and influential Japanese article? Would you try to reorganize it in that case, too?

KWN: Possibly. I’d like to have a lot more translations of Japanese scholarship in Monumenta Nipponica, but it’s hard to find academic writing in Japanese that’s pitched at the right level. Monumenta assumes an audience that already has some knowledge of and interest in Japan. Ours is not a general audience; it is one that ranges from specialists in Japanese studies to people with a serious interest in that field. But Japanese academic writing is often too technical and specialized even for that audience. At other times pieces written by major academics for a more general audience end up sounding too loose and sweeping in English. It can be hard to find Japanese material that will work well in English simply translated as is. You end up either trying to give a general paper a tighter framework of background information, or else trying to make highly technical material more accessible. If a paper is too loose and general, it helps if you can get the author to follow the principle that for every generalization you have to provide a couple of specific examples to pin it down. But either way, the problem of finding the right pitch to bridge the different norms of writing in the two languages and for different audiences remains a persistent challenge.

Links to related research
Q: Often scholars don’t know what other people are doing in a related field, so editors or editorial committees may take an active role in making those connections for them. Do you actively attempt to relate one article to another in the same field, or do you leave that up to the author?

KWN: I mainly rely on referees to do that. I may offer some advice along those lines if the paper is on a subject about which I have some knowledge. It is always best to take into account other relevant research as much as possible. If you, as editor, are aware of such research, then you can give such advice, but if you are not familiar with the field, it may be difficult to do. At Monumenta Nipponica we have an advisory board of people with different areas of specialization, and they often provide that kind of input at the refereeing stage. We have referees outside the editorial board as well, but since the editorial board represents all the major areas that we publish in, individual members are well situated to give advice to a particular author about relevant research on one subject or another.

Differences of opinion
Q: Do you ever come into conflict with your authors or your referees over the content or editing of an article?

KWN: Different referees might give quite different suggestions, but the editor can try to mediate or suggest following one line rather than another. And sometimes, sure, authors disagree with suggestions either as to substance or style and ignore them. But as long as the essential problems are solved, the author doesn’t have to follow specific suggestions to the letter. There are multiple ways the material can be organized, none of which is the right way; it’s a question of devising a nicely structured, clear development of the material. At the first stage of comments, as long as the author becomes aware of what the problems are, my preference is not to suggest specific changes but to point out the problems so that the author can make the changes himself or herself (although quite often, if there are problems with organization, I’ll make general recommendations about ways the author might reorganize the material). If that doesn’t work, then I do sometimes end up making much more specific suggestions about reshaping or may move things around myself at the copyediting stage.

Q: What about differences of opinion among the referees and editorial board members? Who has priority or gets to make the final decision?

KWN: In your comments for the author you can sum up the points from the referees that seem most essential for the author to take into account, or you can suggest that any differences among the referees’ opinions mean that there are different options for the author to choose from. It requires a certain degree of delicacy—you never want to directly advise an author to disregard referee A in favor of referee B, but in your summing up of the referees’ main points you can indicate the ones that you feel really need to be addressed and leave the others up to the author’s judgment. Referees may feel that their advice has not been taken and may be unhappy about that, which could be a problem if you want to go back to that referee for something else; and you may not agree with all the advice that a particular referee gives, but as editor you have to decide how to incorporate that advice and make it available to the author.

Adding to or changing the draft
Q: Generally speaking, do articles tend to get shorter or longer in the editing process? Do you mostly cut things out or ask for additional material?

KWN: We do both; it depends on the specific case. If the writing is loose or has a lot of extra words or reiteration, then the article may get more compact with editing, but there are also cases where more information needs to be provided and the article gets longer.

Q: In the sample you’ve provided, you seem to have added a big chunk of text. Do you normally go ahead and actually write passages like that or do you normally just ask the author to do it?

KWN: It depends on the stage of the editing process. At the first stage I mostly just read through and immerse myself in the author’s argument, and try to figure out what he or she is attempting to say. I can see there are problems, but rather than thinking in my own terms of what the implicit argument is, I just ask the author to develop what needs developing. That initial stage is what I would call developmental editing. But if the problems are still there at the next stage, then I sit down and consider more specifically how I would order these materials, what the essential argument is, and how all the bits and pieces fit into that framework. Even then, however, I don’t really add ideas. I may add some information here and there, perhaps to spell out something more clearly, but in that case I would call it to the author’s attention and confirm that it’s okay. If it just seems that something is missing, I would ask the author to provide more information or a specific example. You also have to use your discretion if you think the author is particularly sensitive about having his or her wording tampered with. On the other hand, many authors don’t really mind revisions as long as the end product says what they want to say. Ideally, though, you should try to preserve the author’s voice and avoid imposing your own voice onto the text; you want to somehow bring out the potential of what the author already has. It’s not always easy.

Q: In the sample you gave us, in the edited version there is a section that has been added. Was this added by the author after the first round of editorial comments, or was it added by the editor? In the latter case, would you go back to the author to confirm that she or he accepts the change?

KWN: Every edited draft goes back to the author for review and further revision. The revision you refer to picked up on points the author had made elsewhere; it was not a case of adding any new information. The author had plunged into too much specific detail without first setting the stage for the reader, so some additional basic information was needed. The author had provided that information at various other points, so this was fundamentally a matter of setting it out at the beginning to orient the reader better—just pulling existing pieces together more topically and presenting them in a sequence that unfolds step by step.

Q: So rather than asking the author to do this, you actually wrote that passage?

KWN: Yes, since previous comments about what needed to be done hadn’t produced the desired result. But I then sent it back to the author for review and confirmation.

Style rules
Q: Are authors asked to follow a certain system for referencing and style, and if they are but still don’t follow it properly, who fixes it?

imageKWN: We have a detailed style sheet that we refer authors to and that is available on our website. The citation style we use is rather eclectic, drawing from different models, because there are various problems with citing Japanese sources in English that standard style guides like The Chicago Manual of Style don’t really deal with. As for enforcing the style sheet’s use, it would be nice to have the luxury of saying if you don’t follow our style sheet we won’t publish your article, but the reality is that we fix anything in the author’s draft that isn’t in line with our style rules. Some well-established, sought-after journals are very firm about authors following their format, but we feel that, since the author initially doesn’t know whether or not we’re going to accept the article, and may want to submit it elsewhere if we turn it down (or say that we’ll reconsider it, but only with major revision), it doesn’t seem fair to insist at the submission stage that they adhere to our specific style. It certainly makes a good impression if the author has taken the trouble to find out what our style rules are and has followed them in the initial draft. We usually ask authors to make their citation format, use of italics, and so on conform to our style only after we’ve accepted the article.

Timeframe of editing process
Q: I’m sure there’s quite a bit of variation, but roughly how long does it take from the time the author submits an article to the final edited draft?

KWN: It depends on how much revision is needed and how much time the author takes to do it, but I suppose a year to a year and half is a reasonable assumption.

Length of the introduction

Q: This introduction is three pages long. Would it be appropriate to have so long an introduction for a term paper or small research paper as opposed to a book or major article?

KWN: For introductions I don’t think there’s any absolute rule about length. What’s more important is setting out the organizing structure, the guideposts to follow for the rest of the piece.

Organization/structure
Q: How much knowledge do you expect authors to have about writing academic papers and organizing their argument?

KWN: Well, in principle, they’re supposed to know a lot about it and be good at it. But as I said, it’s always easier to see the flaws in someone else’s work than in your own.

Q: I can understand how you might expect that of an author in a writerly field like literature or journalism, but what about scientists? How much can an editor expect them to know about the writing process, how to organize a paper, and so on?

KWN: I’m not really familiar with scientific writing but my impression is that there’s more of a set framework in the hard sciences, a kind of formula to follow in terms of laying out the information, so I imagine you don’t really have the problem of building the structure from scratch, which is fundamental to and peculiar to writing in the humanities and social sciences.

Q: How do you deal with authors who tend to lean too heavily toward chronological rather than topical organization?

KWN: I would try to get the author to look at the chronology and decide what are the key turning points in the development of events it covers—how do we get from stage 1 to stage 2, what accounts for that transition, and so on—so that the chronology is divided up into distinct stages. This forms a larger picture around which the author can then build his or her own argument. The chronology is still present, but you’re focusing on the phenomena characteristic of each phase rather than all the specific developments in their real order. Some developments that occurred in the first stage might actually be more characteristic of a later stage, so you can save them until you come to discuss that later period. This is all part of holding information in reserve until the right moment in your own argument. You want your readers eventually to be aware that a certain development didn’t start right with stage 2, that there were earlier harbingers of it, but it’ll often be easier for the reader to understand the situation if the development is discussed where it really becomes significant rather than mixed in with the issues that were central to stage 1. You need to group like with like in your discussion, even if that means interrupting the chronological order. 

Q: Do you have an overall structure in your mind when you edit articles? In experimental sciences, for instance, authors generally have to follow a common, well-defined basic structure. Is there no such general pattern in the liberal arts?

KWN: No, there isn’t any fixed structure, and that’s the challenge: the structure has to grow out of the material and what the author wants to say about it. That’s why organization is such hard work; it doesn’t fall into place naturally. In scientific writing you have a kind of set framework that you’re expected to follow, and you probably think in those terms from the outset and can easily organize your materials that way. But in the humanities you’re more on your own because there’s no single approach or organizational model. Silently, without telling your readers in so many words, you have to lead them so that they can follow an argument whose essential structure is not predetermined. That’s why the first stage, the stage of getting the topical organization and architectural structure in place, is the most fundamental part of the editing process in the humanities and social sciences. You can’t start just with the language; you have to start from the structure.

Originally published in the SWET Newsletter, No. 122 (May 2009), pp. 39-56.