推荐给各位科研工作者一本好书：writing science。
这本书不同于其他教写作的书，不是从写作的角度教你具体的写作方法，而是强调一个理念：科研写作实际是讲故事。
阅读了这本书的前九章，感觉受益颇多，难怪亚马逊上有人评论：看过他的书之后，我想把我之前写过的论文全部重新写一遍。
强烈推荐各位科研工作者和写论文感到痛苦的小伙伴阅读这本书，并完成每章的练习。

• 作者推荐阅读的书籍（粗体是必读书）：

  • style: toward Clarity and Grace
  • Made to stick
  • On writing well
  • The Chicago Guide to Communicating Science
  • Bird by bird
  • Ten Lessons in Clarity and Grace
  • Writing Tools
  • On Writing

• 附上前九章的笔记：

Writing science

Preface

• Although I believe I have become a good writer, I got there through hard work and hard lessons.
• My writing has improved because I worked on becoming a writer. That doesn't mean just writing a lot. You can do something for many years without becoming competent.
• I have learned to write throught a number of avenues: guidance from my mentors; the trial and error of reviews and rejections; thinking about communication stratege; working with students on their papers; reviewing and editing humdreds of manuscripts; reading and rereading books on writing; and importantly, participating in my wife's experiences as a developing writer, listening to the lessons from her classes, and watching how real writers train and develop. I have tried to meld all these lessons into science writing incorporating writers' perspectives into the traditions and formlas of science.

Principles versus rules

• Most of the time, following the rules will improve your writing , but good writers break them when it serves their purposes.
• If you violate principles, your writing will suffer
• When following a rule conflicts with following a principle, flout the rule freely and joyously.

Acknowledge

• I've worked on manuscripts with a number of graduate students and postdocs. They helped me develop my own writing tools and my analytical understanding of those tools so I could teach them to others.
• Finally, I would like to note two books that have greatly influenced my thinking on writing and communications: Joseph Williams's Style: Towrd Clarity and Grace, and Chip and Dan Heath's Made to stick.

1 Writing Science

• As a scientist, you are a professional writer.
• You don't succeed as a scientist by getting papers published. You succeed as a scientist by getting them cited.
• When I needed examples of good wrting, I could usually go to the leaders in various fields--most write exceptionally well.
• This is the widespread notion that "to write clearly, you must first think clearly." This sharp little maxim may appear logical, but it is really rubbish.
• "clear thinking can emerge from clear writing."
• In many cases, writing is the process through which scientists come to understand the real form and impliations of their work.
• As you focus on writing clearly, you force yourself to think more clearly.
• After we get the data, we "write up" the paper. This is an unfortunate approach. Because writing is a critical tool, you shold study it and develop it as thoroughly as your other tools.
• Even the most successful writers struggle with writing.
• But our work gets read and cited because we made our points well enough that readers could follow what we were saying. Our proposals are funded because we were able to make our ideas clear, compelling, and convincing to reviewers. Our success, then, comes from our ability to communicate our ideas as much as from their inherent quality. As the author, therefore, your jobs is to make the reader's job easy.
• It is the author's job to make the reader's job easy.
• I take a different approach--treating being a writer as something a scientist is. You should sharpen your tools or expanding your toolbox, "writing toolbox".
• If you want your writing to be effective, become a writer.
• Thus. a large part of the book is about story and story structure--how you lay out issues, arguments, and conclusions in a coherent way.
• Then I move on to finer scales, from overall story structure through paragraphs and sentences to how we choose individual words.

1.1 Writing versus rewriting

• This is really a book about rewriting, not writing.
• First drafts, though, don't matter; no one else sees them. Trying to get a first draft perfect can be paralyzing, a phenomenon well recognized by the best writers on writing.
• A warning: if you think about these principles as you draft, you may never draft anything. Most experienced writers get something down on paper or up on the screen as fast as they can, just to have something to revise.
• Rewriting is the essence of writing.
• Shitty First Drafts. All good writers write them. That is how they end up with good second drafts and terrific third drafts.
• Not one of them writes elegant first drafts. All right, one of them does, but we do not like her very much.
• Writing can be a painful process of rewriting, rewriting, and more rewriting until your work gets good enough to send off.
• How do you get to Carnegie Hall? Practice, practice, practice! Howe do you get an award letter from the National Science Foundation or ther National Institutes of Health? Polish, polish, polish! If you are going to be a successfull writer, learn to embrace the pain and enjoy the process.

2 Science writing as storytelling

• Elizabeth Kolbert once said that the problem she has with scientists is that we don't tell stories.
• That lack of recognition raises several issues that scientists should consider. The first is the formalism of how we wirte papers and proposals. The second issue is how to become better storytellers and better communicators.
• To tell a good story in science, you must assess your data and evaluate the possible explanations--which are most consistent with existing knowledge and theory? The story grows organically from the data and is objective, dispassionate, and fully professional. Where you run into problems is when the authors know the story they want to tell before they collect the data and then try to jam those data into that framework.
• Lamott highlights the importance of listening to your charactors to draw the story out of them, rather than imposing it on them.

2.1. Finding the story

• In the discovery of the structure of DNA and the molecular basis for heredity, it wasn't Rosalind Franklin's Photo 51, the critical X-ray diffraction image of DNA (figure 2.1a) that gained fame but sketch of the molecular structure of DNA that Francis Watson and James Crick built from it (figure 2.1b)
• 加图片2.1
• 加图片2.2
• Because they used the data to tell a story about nature and how it works, developing an intellectual model of DNA structure and what that implies for heredity.
• The role of scientists is to collect data and transform them into understanding. Their role as authors is to present that understanding.
• In the case of DNA, Photo 51 was data--an image of X-ray scattering. Franklin used that data to produce actual, critical information on the atomic structure of crystallized DNA. Watson and Crick used that information to produce knowledge--the double helix structure. The last step isunderstanding--taking that knowledge about the molecule's structure to explain how it allows cell replication and heredity.
• The story grows from the data, but the data are not the story.
• Develop your story from the bottom up, then tell it from the top down. Start with the data, think about them, listen for the story they are trying to tell, and find that story. Don't listen just to your characters' loud proclamations, though; listen also to their quiet, uncertain mutterings.
• Overinterpret your data wildly, and consider what they might mean at those farthest fringes. Explore the possibilities and develop the story expansively. Then, take Occam's razor and slash away to find the simple core.
• Why go through this "elaborate and slash" process? Isn't elaborating a waste of time if you're going to come back to a simple story in the end? Why not start there? Well, if you start with the first simple story that comes to mind, you are probably imposing plot onto your characters and falling into the trap Lamott describes. Only by exploring the boundaries and limits of your data can you find the important story.
• For his doctorate, he worked on how hill slope steepness controlled soil depth in the Pacific Northwest. Most of the data fit a nice tight relationship (figure 2.3), which made a perfectly good story. But there were outliers where soils were much deeper than they "should” be. He could have ignored them and focused on the main story. He didn't. He looked at the deep soils and what created them; he found that along a hill slope, the bedrock is uneven and in places forms hollow "wedges" (figure 2.3). Over time, those wedges fill up with debris and soil. Once filled , they are's obvious on the land scape, but woe to the person who buys a house below one--in a heavy rainstorm, they can flush out, creating lethal mud flows. Evaluating the processes that fill and flush these wedges became a focus of Dietrich's early research career. Because he listened to his characters carefully, recognized that the most important story wasn't in the average but in the outliers, and then explored those outliers, he came up with more novel, exciting, and important science.
• 加图片2.3
• Learning to explore the fringes of your data, however, can be difficult and frustrating.
• Mary saw more of the system and how it fit together. She was exploring the issues to deepen our thinking, to ensure we found the story that tied together the sometimes apparently contradictory data, and to identify issues that might trip us up later.
• So listen to your characters carefully--take the time to hear what they have to say and figure out what they mean. Fight the pressure to publish prematurely. One good paper can launch a career; Many mediocre ones build a rather different one. Think well, write well ,and then think some more while you write. Let the story grow from the data and then structure the paper to tell that story.
• When we recognize that writing a paper is writing a story, it raises the obvious point that we can become better sotry tellers, better writers, and better scientists by studying what makes a good story, how other writers do it, and how to apply those ideas to science.
• There are three aspects to effective storytelling. The first is content; the second is structure; the third is language.

3 Making a story sticky

• A sticky idea is an idea that is more likely to make a difference.
• Others may stay with you for your entire life and be passed on to your children. Some are so powerful that they have lasted intact from the dawn of civilization.
• In their book, Made to Stick, Chip and Dan Heath frame this question as "What makes an idea 'sticky?"
• Heath and Heath ideatify six factors that make an idea sticky and organize them in a simple mnemonic: SUCCES.

• S: Simple

• U: Unexpected
• C: Concrete
• C: Credible
• E: Emotional
• S: Stories

• They are fundamental to good storytelling and thus to good science writing.

3.1. Simple

• Ideas that stick tend to be simple.
• Hence not simple, but simplistic.
• Most science is driven by simple ideas. Frequently, the simpler an idea is at its core, the larger its swath of influence.
• I have to make things simplistic enough that I can understand them
• A simple idea, therefore, is one that finds the core of the problem.
• There are different ways to find and express a simple message. For some it would be an equation; for otehrs, a verbal description.
• I have always felt that I don't understand something until I can draw a cartoon to explain it. A simple diagram or model--the clearer the picture, the better.
• 加图片3.1
• We went back and banged our heads for several weeks trying to find the truly simple story in the data.
• That paper has been cited over 100 times, largely because the reviewers held our feet to the fire to do a better job of finding the simple story in the complex data.
• Of course, it's better when reviewers hang tough than when they are "nice" and let you publish less-than-perfect work. The pain of an embarrassing review lasts a few days, the pain of an embarrassing paper lasts a lifetime.

3.1.1. Simple Language: Schemas

• Part of being simple is expressing your thoughts in language that builds off ideas that your readers already know.
• Why are schemas so important to create messages that feel simple? They are how people learn; we start with existing schemas and then attach new information to develop new, more sophisticated ones.
• For example, if you were describing how alligator meat tastes, you might say:

It's a light-colored, finely textured meat, with very little fat. It cuts easily and is moist if not overcooked. The flavor is mild.

• Or you could say:

It tastes like chick, but a little meatier.

• Similarly, in molecular biology we start with the simple transcription/translation model of DNA->RNA->protein, and the idea of one gene/one product. Only after establishing those schemas do we start introducing ideas such as post-translational modification of proteins and overlapping reading frames (a single stretch of DNA may actually be part of two separate genes). Each step takes a simple schema and modifies it, making it increasiingly elaborate and nuanced.
• To communicate effectively in science, we need to know what schemas our audience holds so we can build from them.
• We think and write based on the schemas we and our closest colleagues hold, limiting the reach of our writing to a narrow community.
• When you use ideas and terms, stop and think about whether they relate to schemas held by the target audience. If not, don't be afraid to redefine your ideas in simpler terms and more broadly held schemas.

3.2. Unexpected

• Really good papers go beyond incremental to novel--they say something new and unexpected.
• Few data sets don't provide the opportunity to develop new insights. Conversely, few data sets are so imbued with novelty that you can't use them to tell a boring and uninsightful story. Your job is to find what is novel and highlight the unexpected elements. Frame new questions and look for new insights. Make them clear in your writing.
• In science, the key to highlighting the unexpected is through the knowledge gap theory of curiosity.
• By highlighting that unknow, identifying ignorance in the midst of knowledge, you create unexpectedness and engage a reader's curiosity.
• Science doesn't advance by great leaps but by many small steps.
• In any event, it is better to write about a small knowledge gap than about no knowledge gap at all.
• We make a good story by identifying the knowledge gap we will fill.
• Identifying a knowledge gap creates curiosity. Filling that gap creates novelty.

3.3 Concrete

• Science lives with this tension between concrete data and abstract ideas.
• In fact, being able to convert the concrete into the abstract is part of what makes someone an expert.
• The more we learn, the more we are able to think about a topic at a higher level of abstraction.
• Abstract and concrete, however, are not a dichotomy but a continuum. At the top of the ladder are the widest abstractions. At the bottom are the physical facts.
• The danger zone is in the middle--small-scale abstractions that are neither concrete details nor high-level schemas.
• These middle-level concepts are what outsiders consider jargon.
• You can't avoid the middle rungs, but you can minimize the damage--you can ground and define your specific concepts either in widely understood schemas or in the details that explain the bastractions. I discuss how to do this later in the book (particularly in chapters 11 and 14).
• By linking a concept to a concrete example, the concept itself becomes concrete--a new schema you can work with.

3.4 Credible

• We build a chain that extends from past work into future directions. A break anywhere in that chain makes the whole endeavour lose credibility.
• The proposal only became credible when it became concrete. That's what convinced me it was worthwhile and converted me from a skeptic to a supporter.

3.5 Emotional

• Curiosity
• To engage us in your work, you need to engage our curiosity. You do that by asking a novel question.
• The E element of the SUCCES formula is thus closely aligned with U. Unexpected things create curiosity, so use that link to your benefit. You engage emotion by shifting your focus from "what information do I have to offer?" Phrased differently, shift from "what's my answer?" to "what's my question?"
• You must excite the reviewers. Excitement is the therefore the second acceptable emotion in science, and it grows from cuiosity. We get excited about work that engages and then satisfies our curiosity.

3.6. Stories

• But stories are modular; a single large story is crafted from a collection of smaller story units, threaded together. To write a good paper, you need to think about internal structure and how to integrate story modules.
• For example, in chapter 2, I told a story about the role of sotrytelling in science. I built it from three modules, each its own story with its own characters. The first focused on Elizabeth Kolbert and her perception that scientists don't tell stories. The central characters were Kolbert, scientists, and , importantly, the idea of "story" as a character itself. In the second module, to discuss the idea that science goes from data to understanding, I used the story of the discovery of the structure of DNA. Finally, to describe how "listening to your characters" can enhance science, I used the stories of Bill Dietrich's doctoral work and that of my own. I hope that each of these short stories was sticky in its own right, and that together they created a sticky overall story.
• Find units that you can package into coherent modules. Readers will be able to assimilate each piece, and it will be easier for them to see how they add up to create the whole.
• Before you start writing, take the time to figure out how you are going to weave them into your work. Particularly, take the time to figure out the simple story. Build it around the key questions that will engage U and E. These will guide you in selecting the material you need to present to make the story concrete and credible.

4 Story Structure

• Unerstanding the common elements, the ways you can put them together, and when each structure works provides a powerful tool for approaching different writing challenges.
• There are four elements that underlie the structure of all stories, including thoses we write in science:

Opening (O): Whom is the story about? Who are the characters? Where does it take place? What do you need to understand about the situation to follow the story? what is the larger problem you are addressing?
Challenge (C): What do your characters need to accomplish? What specific question do you propose to answer?
Action (A): What happens to address the challenge? In a paper, this describes the work you did; in a proposal, it describes the work you hope to do.
Resolution (R): Howe have the characters and their world changed as a result of the action? This is your conclusion--what did you learn from your work?

• Unerstanding how to manage the OCAR elements is the heart of successful writing. A story lacking any element from it will be unsatisfying, ineffective, and slippery, rather than sticky.

4.1 The four core story structures

• Are readers willing to wait to get the point of the story, or do they want to see it right away?
• The following structures span from targeting the most to the least patient of audiences.

OCAR structure

• OCAR is the structure we use most frequently in science because readers are patient
• Thus, a paper's challenge is presented at the end of the introduction, and its conclusion comes at end.

ABDCE Structure

• For those a step less patient, a structure known as ABDCE works well.

• Action (A): Start with a dramatic action to immediately engage readers and entice them to keep reading.

• Background (B): Fill the readers in on the characters and setting so they can understand the story as it develops.

• Development (D): Follow the action as the story develops to the climax.

• Climax (C): Bring all the threads of the story together and address them.

• Ending (E): What happened to the characters after the climax? (This is the same as the resolution.)

• The difference between ABDCE and OCAR is that ABDCE front-loads the story by moving the challenge up and collapsing it into the opening to create the initial "action" --an exciting starts to grab your attention. Thus, A and B together comprise the O and C elements.

• ABDEC gets the reader into the story faster by launching directly into the challenge, so it is good with an impatient audience, such as proposal reviewers. But it is less efficient than OCAR in moving the story forward.

• ABDCE reaches its apex in mystery and adventure stories.

• One aspect to both the OCAR and ABDCE structures is that they have a resolution that shows how overcoming the challege has chaged the characters and their world.

• The resolution makes sense out of that action. The resolution wraps up the story that was introduced in the opening; it closes the circle.

• The importance of closing the loop means that a good story is circular; at the end, it must com back to the beginning.

• Thus, a story isn't truly a circle, but a spiral (see figure 4.1).

• Highlighting this spiral structure is key to making an OCAR or ABDCE story powerful.

LD Structure

• Reporters use a structure that they call the "inverted pyramid"; I call it Lead/Development (LD) to highlight its key functional elements.
• In LD structure, the core of the story is in the first sentences (the lead, L) and the rest fills out and develops the story (the development, D). In LD structure, the lead collapses the opening, challenge, and resolution into a single short section, possibly as little as a single sentence.
• Of such a progression of sentences, each tugging the reader forward until he is hooked, a writer constructs that fateful unit, the "lead."
• The point must go at the beginning.

LDR Structure

• Magazine writers can afford to worry about ending well with an effective resolution.
• Thus, magazine journalists use a story structure I describe as Lead/Development/Resolution (LDR).
• These structures give us a continuum, based on readers' patience:

• OCAR: Slowest--take your time working into the story.

• ABDCE: Faster--get right into the action.
• LDR: Faster yet--but people will read to the end.
• LD: Fastest--the whole story is up front.

• You sholud be able to read the O,C, and R of a paper, and still get its key points. If you know the problem, the specific questions, the general approach to answering them, and the conclusions, you may have gotten all you need from that paper. You'll certainly know whether you need to go back and read it fully.

4.2. Applying story structure to science writing

• A good paper or proposal describes the larger problem and central "characters" (O); it frames an interseting question (C); it presents your research plan and results, developing the action (A); and it leaves the reader with an important conclusion about how our understanding of the world has changed as a result of the work (R).
• Papers for specialist journals generally use straight OCAR, framing the challenge at the end of the introduction.
• In fact, in these journals, a lead-based paper might be considered suspect. Front-loading the story with conclusions could make it seem like you knew the story you wanted to tell and were simply forcing the data into that story--that is, you are trying to prove rather than test your ideas--a no-no going all the way back to the foundations of the philosophy of science.
• In contrast, generalist journals, such as Nature or Science, need a faster structure, often closer to LDR.
• A famous use of a lead-based structure in science is Francis Watson and James Crick's famous paper "Molecular Structure of Nucleic Acids: A Structure for Deoxyribose Nucleic Acid," in which the opening was: "We wish to suggest a structure for the salt of deoxyribose nucleic acid (D.N.A.). This structure has novel features which are of considerable biological interest." The rest of the paper describes this structure and resolves with the statement, "It has not escaped our notice that the specific pairing we have postulated immediately suggests a possible copying mechanism for the genetic material." They didn't bother to elaborate on that mechanism--it was obvious enough that they didn't need to.
• A proposal must convince reviewers that the topic identified in the opening is important and then compel them with the excitement of the questions posed in the challenge. If it fails to do this, it is dead.
• When I review proposals I make a "no/maybe" cut by the end of the introduction, and if it's a "no," that is irrevocable. I then only read the rest to be able to give feedback on how to improve the proposal for resubmission. A good experimental design can never compensate for boring questions. A "maybe" at that first cut means the questions are exciting, in which case I read the rest to see whether the experimental design is adequate to answer them.
• "If you haven't told them in the first two pages, you haven't told them." To "tell them" in the first two pages requires a front-loaded structure: either ABDCE or LDR.

4.3. Mapping OCAR onto IMRaD

• We usually write papers using some variation of IMRaD: Introduction, Methods, Results, and Discussion.
• While IMRaD is a rule, OCAR is a principle.
  Introduction :This has three subsections, although they are rarely broken out as such:

• Opening: This is typically the first paragraph that introduces the larger problem the paper is targeting. What is the context, and what are the characters we are studying?

• Background: What information does the reader need to understand the specific work the authors did, why it is important, and what it will contribute to the larger issue? I consider this an extension of the O, as it fleshes out introducing the characters.

• Challenge: What are the specific hypotheses/questions/goals of the current work?

Materials and Methods :This begins describing the action--what did you do?
Results :This continues the action by describing your findings.
Discussion : This develops to the climax and resolution. What did it all mean, and what have you learned? It often ends with a conclusions subsection that is the resolution.

• Thus, opening and challenge block the beginniing and the end of the Introduction. Action encompasses the M&M, Results, and much of the Discussion. The resolution is the last section of the Discussion.
• For science, at the end, we want to know how our understanding of the world has changed as a result of your work, or in the case of a proposal, how you think our understanding of the world will change. The resolution must map back to the opening. It must say something about the larger problem you identified there.
• A closely related aspect of mapping OCAR onto IMRaD is that scientific papers have an hourglass shape to their content (see figure 4.2).
• 加图片4.2

5 The opening

• The most important sentence in any article is the first one.
• Initial impressions are strong and lasting. Your first words have great leverage, making the beginning of a paper a "power position." You must use that power to accomplish three goals: identify the problem that drives the research, introduce the characters, and target an audience.
• You must start well. Your first sentences get readers moving and set the direction; you establish their expectations and generate momentum.
• The opening begins with a single sentence but typically encompasses the first paragraph, and sometimes several more. In a short paper or one for a narrow audience of experts, you can quickly remind people of a problem they already know. When you target a broader audience, one made up of people who hold different schemas than you do, you may need a longer and more complex opening.

5.1. Examples of good openings

• The fundamental question for each is: does it achieve the three goals? Is it clear what the paper is about? Does it frame the problem? Does it introduce the critical characters? Read the openings and answer these questions, before going on to my analysis. Do you agree with my assessment?

Example 5.1
Since the late 1800s, N mineralization has been the perceived center point of the soil N cycle and the process that controls N availability to plants.

• The key word in this sentence is perceived, a distinctive and unusual work that draws your attention. Clearly, this paper is going to challenge that perception. Additionally, there is going to be a historical element--evaluating how the perception has changed since the late 1800s.

Example 5.2
Current public health guidelines in the United States, the United Kingdom, and Australia recommend that women consume a supplemental dose of 400 μg of folic acid per day in the month preceding and during the first trimester of pregnancy to reduce the risk of neural tube defects in children.

• you can infer the entire story: folic acid supplements during pregnancy may increase the risk of childhood asthma.
• In these first two examples, the opening sentences are dramatic and launch quickly into the story. Frequently, however, openings require several steps to develop the issue.

Example 5.3
The topography of mountainous landscapes is created by the interaction of rock uplift and erosion. River incision into bedrock is the key erosional process that controls the rate of landscape response to changes in rock uplift rate and climate.

• The first frames the focus of the story. The second sentence picks up the idea of erosion and develops a specific focus.
• Sometimes the opening needs to be longer and can include the entire first paragraph.

5.2. Bad openings

• Good openings all identifiy a problem of broad interest and give the reader a sense of where the story is going.
• There are two obvious ways to fail: provide either misdirection or no direction.

5.2.1 Misdirection

Example 5.5
Plants are a critical control of CH₄ dynamics in wetland ecosystems. They supply C [carbon] to the soil methanogenic community both through production of soil organic matter, and as fresh exudates and residues. Fresh plant material may be an important CH₄ precursor even in an organic matter–rich peat soil. Strong correlations between net primary productivity and systemlevel CH₄ fluxes across a wide range of ecosystems highlight the importance of plant C inputs.
Vascular plants, however, also transport CH₄ out of soil and sediment, effectively bypassing the aerobic zone of CH₄ oxidation.

• I started in one direction, but then struck out in a different one; that is misdirection, and it's confusing.
• I could have written this better and avoided any potential confusion by changing the first sentence, making it a broader positioning statement.

“Plants control CH₄ dynamics in wetland ecosystems by two mechanisms. The first is to supply C to the soil methanogenic community . . .”

• This would have let you know that the paper is about both mechanisms and might imply that it evaluates the balance between them.
• Even though the first paragraph is about substrate supply, you would know that there is more coming, so the second paragraph would not feel like it was changing direction but completing the direction I had started.

5.2.2 No Direction

• The other common error in the opening is giving no direction.

Example 5.6
In meiosis, genes that are always transmitted together are described as showing “linkage.” Linkage, however, can be incomplete, due to the exchange of segments of DNA when chromosomes are paired. This incomplete linkage can lead to the creation of new pairings of alleles, creating new lineages with distinct sets of traits.

• Rather, it goes over basic, textbook material about eukaryotic genetics that should be second nature to most readers. It explains a schema that scholars in this field don't need explained.
• If you do this, though, when you revise, figure out where the real story starts and delete everything before that. At a writers’ conference my wife attended, a well-known author said that he sometimes has to delete several chapters to get to where the story begins.

5.3. Targeting your audience

• The way you introduce your problem and your characters affects the audience’s attitude toward the work and maybe whether they continue reading. You must know the intended audience to tailor the writing to them.

• Example 5.7
  For a specialist journal: Epifluorescence microscopy and direct viable counting methods have shown that only 0.01 to 0.1 % of all the microbial cells from marine environments form colonies on standard agar plates. Much of the discrepancy between direct counts and plate counts has been explained by measurements of microbial diversity that employed 16S rRNA gene sequencing without cultivation. The present consensus is that many of the most abundant marine microbial groups are not yet cultivated.
• Example 5.8
  For a generalist journal: Antonie van Leeuwenhoek (1632–1723), the first observer of bacteria, would be surprised that over 99 % of microbes in the sea remained unseen until after Viking Lander (1976) set out to seek microbial life on Mars.

• This opening says something similar to example 5.7, but these authors (Farooq Azam and Alexandra Worden) were targeting the editors and readers of Science, a group with limited interest in methods for culturing bacteria. So, they opened with a short story whose characters are Antonie van Leeuwenhoek, the Viking Lander, and microbes.

• To make their story engaging, Azam and Worden pulled strongly on the SUCCES elements. It is simple, and it is unexpected — we are searching Mars for life when we haven’t found 99 percent or
  more of the life on this planet. It is concrete and credible, backed up by specifics. It is also emotional, pulling on your curiosity and amazement — we’ve been at this for 300 years and have seen at best 1 percent of the bacteria that exist!?

• We submitted similar proposals to two agencies. We used different openings:

• Example 5.9:
  The influence of fog on ecological and hydrological processes in coastal zones has long intrigued scientists.
• Example 5.10:
  California’s coastal forests are among its most distinctive and treasured natural resources.

• How we framed the problem here was critical. An effective first sentence might open the door to funding. An ineffective one could close it.

5.4. Opening for a Broader audience: The two-step opening

• When you target experts in your field, you can open quickly, building off the discipline’s core schemas. Sometimes, though, you need to target a broader audience.
• To do this, you need to open with an issue that engages your target audience, but then modulate it to one you want to work with. That requires a multistep opening in which you take time to introduce and then redefine the focus.
• An example of this two-step approach. Though the work was narrow, the opening was wide:

Example 5.11:
The Arctic has become a focus of attention because global warming is expected to be the most severe at extreme latitudes. The thick organic soils of the tundra contain large stocks of carbon (C), and these soils may act as either a source or a sink for atmospheric carbon dioxide (CO2). It has been suggested that as the climate warms, increased organic matter decomposition will release CO2 to the atmosphere, contributing to warming and creating a positive feedback that results in further increases in atmospheric CO2. Alternatively, it has been argued that increased decomposition will release bound nitrogen (N) and other nutrients in the soil and thereby enhance plant growth, since plant growth is nutrient-limited in arctic tundra. Increased plant growth would allow the tundra to be a sink for atmospheric C because plant material has a wider C/N ratio than soil organic matter. Thus, the direction the C balance of the arctic will shift with warming is unclear and depends on interactions between soil C and N cycling that we still do not understand in the tundra.

• He was writing for an audience of global change scientists, trying to convince them that this paper was something they should read, rather than targeting tundra soil ecologists.
• The award was a result of effective storytelling, rather than inherently interdisciplinary
  measurements.
• In proposals, quickly engaging the reviewers is critical, so you may need to use this two-step approach.

Example 5.12
Succession has been a central theme in ecological research for almost a hundred years. Two questions have directed much of that research:
What causes the shifts in communities?
How do ecological processes change as a result of these community shifts?
These questions are linked through a feedback loop: plants affect soil processes which in turn affect plant community structure.

• Although soil microbes and the processes they carry out were the central characters in my story, I did not introduce them in the first sentence or even the first few sentences. I did that deliberately — I submitted the proposal to the ecology program, and I knew the reviewers were likely to be plant (rather than microbial) ecologists.
• I call this a two-step opening for two reasons. One is to highlight that it does take two steps, but also to highlight that, like the dance, it is must be quick — if you take more than two steps, you will stumble.

5.5. Changing style for different audiences

• It is a principle of effective communication that you need to adapt your language, style, and approach to deal with different media and different audiences

Example 5.13:
Larry Pomeroy’s seminal paper revolutionized our concepts of the ocean's food web by proposing that microorganisms mediate a large fraction of the energy flow in pelagic marine ecosystems. Before 1974, bacteria and protozoa were not included as significant components of food web models. Pomeroy argued forcefully that heterotrophic microorganisms, the “unseen strands in the ocean’s food web,” must be incorporated into ecosystem models.

• In contrast to example 5.8, this uses more technical language and targets an audience of marine ecologists.
• This is a strong opening that effectively engages SUCCES elements. It is concrete, giving dates and directly attributing Pomeroy’s paper. It is emotional, pulling on words like revolutionary and argued forcefully to create a sense of conflict. It even draws on the U factor by setting up the contrast between the thinking before and after 1974. This opening frames the story.
• The authors introduce key characters, and so establish the starting point.
• Skilled writers know their audiences and think carefully about what works for them.
• The opening is critical to that answer.

5.6. How wide should your opening be?

• Remember — getting published is not the ultimate goal; getting cited is.
• So you should set your opening, the top of the hourglass, to draw in as broad a readership as you can manage.
• 加图片5.1

Example 5.14:
Opening : The Arctic is important in the global climate system because tundra soils store a large amount of carbon that may be released to the atmosphere as CO₂. An important recent discovery is that wintertime CO₂ fluxes from soil are large.
Resolution 1: Developing a reliable model of CO₂ fluxes in the Arctic therefore requires a better model of winter C cycling processes.
Resolution 2: In the arctic tundra, microbial community composition changes little through the winter.

• Resolution 1 is framed at roughly the same “width” as the opening.

• But if the story ended up with resolution 2, readers would be dissatisfied. It ended up being about soil microbial communities — a bait and switch that wastes readers' time.

• For example, imagine if the story were about modeling CO₂ fluxes in the tundra (resolution 2), but you opened the paper this way:

• "Bacteria living in tundra soils are well acclimated to surviving the cold conditions of the Arctic winter."

• This promises a story about the physiology of tundra bacteria, not something that would interest someone focused on the global C cycle.

• If you err, though, it’s better to err slightly on the wide side. If you oversell in the immediate opening, you can still filter down quickly.

• the paper has been well cited in journals ranging from microbial ecology to global biogeochemistry.

• If you frame too narrowly, you lose readers immediately, and once lost, you can't get them back.

5.7. Positioning statements: Pawn-pushes versus Queen-launches

• Inexperienced writers often imitate opening lines and come up with platitudes. But in a well-written paper, the sentence may be more — it may be a careful positioning statement that is critical to building the story.
• They used a two-sentence opening. The first was an undramatic pawn push, but it was carefully designed, allowing them to introduce erosion at the end of the sentence — a power position.
• In contrast, in example 5.8, Azam and Worden launched a queen with their opening about van Leeuwenhoek and Viking.
• When you write a straight OCAR story, as is common for specialist journals, you can use a pawn push — an opening that unfolds for a patient audience. If you’re writing for Nature or the National Institutes of Health, however, you are likely using an ABDCE or LDR structure that start with action, so you had better launch a queen.
• To write well, you need to learn how to use the power of the opening. Learn when to use a pawn push and when to launch a queen. Learn to push a pawn like a chess master — as the first step in a strategy to develop your argument and take control of the game. Remember the words of Aristotle, “Well begun is half done.”

6 The Funnel: Connecting O and C

• It forms the funnel in the hourglass; it narrows the focus and leads readers from the general to the specific, drawing them along the story and framing in the knowledge gap.
• When you frame the knowledge gap, you provide the background information necessary to understand the story.
• Framing the knowledge gap taps into core elements of the SUCCES formula for a sticky story, particularly the U and E elements, unexpectedness and emotion. By defining a knowledge gap, unmasking a hole in the wall of knowledge, you create unexpectedness: I didn’t realize that we didn’t know that! By closing with a question, you create curiosity: what is the answer? Then you can tell us how you solve the problem and satisfy our curiosity.
• If you do this well, you can bridge from very large problems to very narrow questions.

6.1. Example of the funnel at work

Example 6.1 (见62页）

• After clearly framing the knowledge gap and its importance, the authors stated their challenge by saying “We studied Reaction 1 . . .” It's obvious that the question is “What is the value of k 1?” But it would have defined the knowledge gap more concretely to say, “We measured the value of k1 at temperatures down to 230 K using an experimental . . .
• This was a Nature paper and so quite condensed — it had to narrow quickly with broad strokes. In papers for specialist journals, you have more space to develop the Introduction and can do the narrowing more gently and thoroughly.
• Make sure that you aren’t telling us everything you know about a topic but developing the logical connections between each step to frame the knowledge gap.

6.2. Bad introductions: Failing to define the problem

• A good introduction defines a problem and narrows to an interesting question. A weak or poor Introduction, in contrast, either fails to define the problem or tries to sell a solution before defining the problem, and so fails on curiosity.
• It’s not very convincing to say “little is known about X” for scientific, logical, and literary reasons.
• Scientifically, it is unconvincing because it’s probably false.
• So when someone says, “little is known about X,” we often feel that the author either doesn’t know the literature or is overstating the case.
• Logically, it’s unconvincing because after saying “little is known,” the authors describe a lot that is known.
• We don’t see the “little.”
• “Little is known” is fuzzy — how little is little?
• To make it convincing, it needs to be concrete — what specifically do we not know?
• You must explicitly define the problem.
• Rather, they said “sources are not clearly defined,” which is tighter language that implies something closer to “while the broad patterns are known, important details are not,” a true description of the state of knowledge at the time and enough to get the paper into Nature.

6.2.2. Offering a Solution before Defi ning a Problem

• a door-to-door salesman: “Hi, ma’am, I’m selling the new Buzco Bizzwidget. The Bizzwidget is the most amazing tool you’ve ever seen — why, I don’t know how you’ve ever lived without it! So here, let me show you some of the wonderful things it does.”
• If you are trying to sell us a bizzwidget solution, first convince us we have a problem: “Hi, ma’am — have you experienced problem X? You have? Do you have a solution? You don’t? Well I do — let me show it to you; we call it the Buzco Bizzwidget.”
• It does this by opening with a concern many people share (defining the audience in the opening), and then showing us why we need a Bizzwidget (the body of the Introduction), before introducing the specific product (the challenge).

Example 6.2:
Addressing complex interactions among chemistry, physics, and biology in climate systems requires an interdisciplinary approach. We propose to address this challenge by using Complex Systems Modeling Theory (CSMT). CSMT has been used in chemical systems to model molecular reaction
mechanisms and in cell biology to model physiological pathways. It has been used . . .

• It doesn’t describe the complex interactions, how other approaches have struggled with them, what the CSMT approach is, or why it is better than other approaches.
• To sell us a solution, first sell us a problem.

6.3. Introduction versus literature review

• An effective Introduction cannot be merely a literature review that synopsizes what we know about a topic. Instead, because you must convince us of the importance of the problem, you must show us what we don't know and why it is important.
• The difference between a literature review and an Introduction can be subtle, because they both use the existing literature to discuss the state of knowledge.
• The distinction between them is that a literature review builds a solid wall — describing
  knowledge — whereas an Introduction focuses on the hole in that wall — describing ignorance.
• An Introduction focuses on the publications that define the edges, rather than the core of knowledge.
• How do tell when you are writing a literature review rather than an Introduction? See whether you are focusing on telling us what we know or what we don’t. When you describe something we know, do you use it to identify the boundaries of that knowledge? If so, you’re writing an Introduction; if not, you’re probably creating a literature review.
• Do you write: “Smith (2003) found X” or do you write: “X occurs (Smith 2003)”?
• The important information is almost never that Smith found it; rather, it is almost always what she found. So why make Smith the subject of the sentence?
• This highlights that there is no agreed-on truth but a collection of individual opinions. If there is an accepted dogma that one researcher is challenging, however, you would write something like: “While most reports suggest X (e.g., Smith, 2003, Xu 2004), Jones (2005) found the opposite,
  arguing . . .”
• Most OCAR papers use a simple O → C flow in the Introduction, with a smooth funnel from the opening to the challenge to define the knowledge gap.

7 The Challenge

• In the challenge, you describe the specific knowledge you hope to gain. This starts with the question that drove you to do the research.
• Some authors only pose the question, whereas others do all three, offering a question, framing it into a hypothesis, and then describing specific research goals.Each approach has its place, but the question is the core of it all. If you don’t have a question, you are not doing good science. If readers can’t tell what it is, you are not writing good science.

7.1. Questions versus hypotheses

• But a hypothesis merely takes your question and makes into a falsifiable prediction. The question, defining the knowledge gap, is still the key.
• I can’t give any global advice as to whether to pose a question or to transform it into a formal hypothesis; you need to know the culture of your field. Just remember that the question comes first and must be clear.

7.2. Questions versus objectives

• Despite the importance of the question, many authors define their challenge by stating “Our objectives were” rather than by saying “Our question was.” That is, they focus on the information they will collect, rather than the knowledge they hope to gain.
• Focusing on objectives instead of questions is weak science and weak storytelling. If you leave the question unstated and implicit, and jump straight to specific data-collection goals, the reader has to figure out what your question was and whether you even had one. You leave it to them to figure out how the work will advance knowledge.
• Focusing on objectives also doesn’t engage SUCCES. It doesn’t create unexpectedness or curiosity — at least not the curiosity you want. A reader will wonder about your objectives. Why did you do this work? What was the purpose? What was the underlying question these tasks address? Is it possible that you were just aping experiments published by others without a real question of your own? Those are not questions about your science but about you and your motivation, and there is an underlying criticism embedded — why are you wasting my time with this?

7.3. What comes after stating the questions?

• After posing the question, a good challenge briefly lays out the research approach. This is where you tell us about specific objectives and the information you will generate.
• Some papers also provide a brief overview of the Conclusions, usually starting with language like “In this paper we show . . .”
• This telegraphed OCAR works well for an impatient audience but one that still wants to see how arguments develop.

7.4. Good Challenges

• In papers for specialist journals, a good challenge almost always condenses conceptually to “to learn X, we did Y.” That is, they present the question and lay out an approach to answering it, as illustrated in example 7.1.

Example 7.1
Our goal in this study was twofold. First, we tested whether an animal’s physical environment would affect hippocampal attributes. Specifically, we tested whether food-caching mountain chickadees (Poecile gambeli) housed in captivity differed in hippocampal volume, hippocampal neuron number and neuronal density as compared with fully developed wild-caught conspecifics. We predicted that captivity, with reduced environmental complexity and restricted memory-based experiences (compared with memory-based experiences afforded in the natural environment), would reduce hippocampal volume, neuron number and, potentially, neuron density.

• The authors, however, do several things well. First, they remind us of the overall issue — fundamentally a question about controls on brain development, a topic of wide interest and import, even though the specific question is narrow.
• The authors did an excellent job of connecting their specific question to the larger problem. They go further, though — even after posing a tight question, they frame a hypothesis that defines their measurements and the data that would falsify or support it. As you read the rest of the paper, you know what the authors think and what they did; you can easily follow along as they assess their results and develop their Conclusions. Nicely done.

Example 7.2
Despite the tantalizing evidence for DAG and/or its downstream products in visual transduction and the synergistic role of calcium, in no instance has application of such chemical stimuli fully reproduced the remarkable size and speed of the photocurrent. This may imply that yet another signal may be missing from the proposed schemes. In other systems PIP2 has been shown to possess signaling functions of its own, independent from those of its hydrolysis products. . . . These observations prompted the conjecture that in microvillar photoreceptors PIP2 may help keep the channels closed and its hydrolysis could promote their opening. In the present report, we examined the consequences of manipulating PIP2 on membrane currents and light responsiveness in is qolated photoreceptors from Pecten and Lima.

• The authors clearly lay out the problem — DAG can’t explain existing observations. They hypothesize a new mechanism that involves PIP2 and briefly describe the experiments they did to test this hypothesis — “to learn X, we did Y.”

Example 7.4:
However, three decades of work in the gas phase have explored how the specifics of the forces between atoms involved in isolated chemical reactions determine the final energy partitioning as the reaction moves from the transition state. Is knowledge of these specifics completely immaterial to reaction dynamics in solution?

• This taps into SUCCES: it draws on simple by asking a clean and straightforward question, and it draws on unexpected and emotion by using the highly charged phrase “completely immaterial” to challenge decades of high-quality science.

7.5. Bad Challenges

• If the challenge is unclear, readers will be left adrift.
• A challenge is ineffective if it doesn’t concretely state the question or hypothesis or if gives the reader the wrong impression as to what it is.
• The most common type of unclear challenge is where authors focus on the information, rather than the knowledge they are trying to acquire; they leave off the “to learn X . . .” and just say “we did Y.” They focus on the objectives, rather than the question.
• You must make the question clear. If you fail to do this, your papers will lack power, and your proposals will likely lack funding.

Example 7.5
Some T-cells may be anergic — that is, unable to proliferate after being restimulated with an antigen. Some anergic T-cells are unable to link to the T-cell– antigen presenting cell (APC) interface. Here we examined the structural characteristics of anergic mouse T-cells and we tested their functional response to being rechallenged with antigen-loaded APCs.

• Here the authors tell us which data they will collect, but they don’t specify the knowledge gap. What is the question?
• This would have been much stronger if they had clarified the “to do X” part of the challenge, perhaps like this

To determine what causes mouse T-cells to be anergic, we evaluated the structural characteristics of T-cells and how they responded to being rechallenged with antigen-loaded APCs.

Example 7.7
The study had two goals. First, we aimed to constrain our estimates of grassland plant production by comparing measurements based on two techniques: maximum biomass at the end of the season and periodic measurements of photosynthesis. Second, we examined the response of grass growth to a combination of elevated CO 2 and increased temperatures, conditions that are expected to occur with climate warming.

• Why is this challenge weak? The authors presented the goals in an order that is more chronological than intellectual. But we expect the most important objective to come first, defining a study’s overall thrust.
• To fix this problem, we need to switch the order of the objectives and highlight the core question:

The primary goal of this study was to evaluate how grass growth responds to a combination of elevated CO2 and increased temperatures, conditions that are expected to occur with climate warming. To validate our plant growth measures, we used two approaches to estimate plant production: maximum biomass at the end of the season and periodic measurements of photosynthesis.

• Remember that the critical part of the challenge is not “we did Y” but “to learn X.”

8 Action

• You are not just presenting your results, you are telling a story.
• In a paper, this includes the Materials and Methods, the Results, and most of the Discussion.
• I focus on how to integrate these sections into the overall story and how to use story structure to present them most effectively.
• You are not just presenting your results, you are telling a story.
• What is the point of all that work? What do these results mean? Do they answer your question? Do they support your hypotheses and conclusions?
• As you write each section, think about it as a mini-story that has its own OCAR elements. Consider why you are telling us any particular piece of information — how it contributes to the larger story — and frame the presentation to highlight why you did something as well as what you did.
• The action sections of a paper can be separated into two distinct parts: describing what you did (Materials and Methods) and what came of it (Results and Discussion).

8.1 Methods

• More often we evaluate the methods to assess the credibility of the data and conclusions.
• To serve the needs of all possible readers, the best way to describe a method is use a lead/development (LD) structure, providing an initial overview for all and then the details for those who need them.

• Example 8.1
  Enzyme Inactivation following 3-HPAA Metabolism
  Enzyme inactivation associated with 3-HPAA metabolism was measured by the method of Turman et al. (2008).

• Example 8.2
  Enzyme Inactivation following 3-HPAA Metabolism
  PGHS-1 or PGHS-2 was incubated with 25 μM 3-HPAA. When oxygen uptake was complete, arachidonic acid (25 μM) was added, and the maximal rate was determined as described above and normalized to the DMSO control. The concentration dependence of PGHS-2 inactivation was analyzed in a similar manner with varying concentrations of 3-HPAA (from 10 nM to 25 μM).

• Example 8.3
  Enzyme Inactivation following 3-HPAA Metabolism
  To characterize the extent of enzyme inactivation associated with 3-HPAA metabolism, PGHS-1 or PGHS-2 was incubated with 25 μM 3-HPAA. When oxygen uptake was complete, arachidonic acid (25 μM) was added, and the maximal rate was determined as described above and normalized to the DMSO control. The concentration dependence of PGHS-2 inactivation was analyzed in a similar manner with varying concentrations of 3-HPAA (from 10 nM to 25 μM).

• Example 8.1 is truly miserable — it doesn’t say anything about how the measurement was done. Because the specific method can affect how we interpret results, we could not evaluate this work without tracking down the original paper. Talk about making the reader’s job hard!

• Example 8.2 is imperfect — it goes straight into detail without an opening to guide us in what those details are about.

• Example 8.3 is the best and the one Turman et al. used — it adds a few extra words to provide a brief overview of the goal before describing the approach.

• Many methods and techniques are well known in their field, and some have specific names. Use them when available. For example:

We measured protein by the Lowry approach.
We measured microbial biomass by the chloroform slurry approach as described by Fierer and Schimel (2003).
We amplified DNA by hot-start PCR.

• These start by describing what the authors did (measure protein, biomass, etc.) and then tell us how by naming a well-known approach.
• From that base, the author provides the experimental details, but may only need to highlight differences from the standard approach:

We measured protein by a modification of the Lowry protein assay, in which sodium citrate, instead of sodium tartrate, is used in reagent A.

8.2. Results and discussion

8.2.1. To sparate, or not to separate: That is the question

• Separate Results from Discussion or not:
  1. Make the reader's job easy (our principle no.1): present results and interpretations in a way that best develops the story.
  2. Readers must be able to distinguish what you found from what you think
     • Data: Your actual results.
     • Inference: These are the clear and robust interpretations of the data that almost any practitioner in the field would draw; these are sometimes so obvious that we treat them as data themselves.
     • Interpretation: Your thoughts, hypotheses, and speculation about what the results may mean for the larger problem you identified.
  • Deciding how to structure the presentation is equally a balancing act.

Example 8.5
Paragraph Integrating Results and Inference
Notably, the spectra of Rv0899-B (Figure 3B, red) and Rv0899-C (Figure 3B, blue) form perfect complementary subsets of the spectrum from Rv0899-BC (Figure 3A), spanning both domains, with the exception of some peaks from residues in the BC connecting region. This demonstrates that the B and C domains constitute independently folded modules, as suggested by sequence homology. The resonance line widths measured in the three spectra are very similar, further indicating that all three polypeptides exist as monomeric species in solution. Since the line widths are not appreciably larger in the spectra of the BC polypeptide, the B and C domains may be significantly dynamically decoupled.

• Because the first two sections integrated data and inference, calling them Results would have mislabeled the material. Pulling all the analysis out, however, would have disrupted the paper’s overall flow. Instead, the authors used a partially integrated structure that flows from more Result-like to more Discussion-like material. Within this hybrid structure, they distinguished results from interpretation by presenting the data first and by using words like demonstrates and indicating to identify inferences. This was nicely done.

Example 8.6
The B and C domains constitute independently folded modules, as indicated by the fact that their spectra form perfect complementary subsets of the spectrum from Rv0899-BC.

• I would characterize this sentence as pure Discussion. Though it refers to the data, it does so to support an argument and the data come after the argument. Whenever the data come first, it feels like you are drawing your interpretations from them, and so are obeying Lamott’s dictum of “listen to your characters.” When conclusions come before the data, it feels like you are imposing plot. That is true even within a single sentence, as illustrated by examples 8.5 and 8.6.

• The important point here is that while it is always essential to distinguish results from discussion of them, it isn’t critical to separate them physically.

8.2.2. Choosing Data to present

• The most important decision in describing results is not how to present your data but which data to present. We often collect a lot of data, not all of which is needed to build the story.
• The natural temptation was to present all of it. But that was a huge data set, while the simple story could be collapsed to two words: “they covaried.” The reader didn’t need all the data to get that point; in fact, it would have been a distraction.
• 加图片8.1
• Writers talk about having to “murder your darlings” — the parts you adore but that don’t contribute sufficiently to the overall piece. It’s good advice for scientists, too.

8.2.3. Presenting Data

• After deciding which results to present, you need to figure out how to present them.
• You need to synthesize them into a pattern and fit them into the larger story to provide context.
• Most results call for an LD structure: first frame the major point or pattern, then flesh out the detail. Don’t present all the details and then synopsize them, or worse, present them without synopsizing or synthesizing at all. Without a framework, readers struggle with details.
• This was a very clear LD structure — you get the entire story in the first two sentences, even though the paragraph goes on for an additional five.

8.2.4. Statistics and Stories

• Although statistics are essential for establishing the credibility of your conclusions, remember that the story is not in the statistics — it is in the data themselves. When you tell the story through the lens of the statistics, by focusing on the statistical analysis rather than on the data, you steal both clarity and power from the story.
• Is it adequate to say “warming significantly increased methane emissions (p < 0.05)”? Again, no — the story is in the amount.
• A better way to describe these data would be “warming increased methane emissions by a factor of 3.4 (p < 0.05).” By focusing on the data and making the statistics supporting information, you can tell a story that says more about nature and is more engaging without forgoing rigor.
• You could describe both of these graphs by saying, “The treatment significantly increased the response (p = 0.02).
• I would describe panel A by saying, “The treatment increased the response by a factor of 2.3 (p = 0.02)”; for panel B, I might write, “The treatment increased the response by only 30 percent, but this increase was statistically significant (p = 0.02).
• 加图片8.3
• When you describe the data solely in terms of whether the difference was significant, you present an interpretation of the data as the data, which violates an important principle of science. Any specific threshold for significance is an arbitrary choice with no fundamental basis in either science or statistics.
• I might approach panel C by describing the results as they are: “The response in the treatment was 2.3 times higher than in the control (p = 0.07)” or, to be conservative, as “The response in the treatment was 2.3 times higher than in the control, but the difference was only significant at p = 0.07.”
• The best description of this graph, that is, the most complete and accurate, would be “the relationship between x and y is weak (R2 = 0.10) but statistically significant (p = 0.001) .”
• By focusing on the data, being concrete, and showing the whole story, you effectively and honestly present your results and allow the reader to evaluate them, fulfilling core principles of both writing and science.

8.3. Discussion

• Discussion is where you present your thoughts and interpretations, where you answer the questions you posed in the challenge, and where you show your contribution to the larger problem framed in the opening.
• Writing a good Discussion is the critical act of creativity in science that no book can teach.
• As you sit down to write the Discussion section, however, the first decision you have to make is which internal story structure to use.
• Some writers use an OCAR structure, opening by reminding readers of the challenge and the question, and then working through to the resolution. Other papers use an LDR structure, opening the Discussion by framing the conclusion — what they showed — and then using the rest of the Discussion to support that argument, building to the overall resolution.

Example 8.8
It is well-known that factors such as the nature of the nucleophile, solvent, and leaving group directly affect the rate of the bimolecular nucleophilic substitution (SN2) reactions; yet, in the case of carbanions, little has been documented with absolute rate constants. . . .
Photoinduced decarboxylation of suitable substituted carbanions provides a route for the formation of substituted cycloalkanes that proceeds in high yields in nonhydroxylic solvents and with good leaving groups such as bromide and iodide.

• In this case, the opening of the Discussion reads like a reiteration of the challenge, reminding the reader of what it was, but without posing any conclusion.

Example 8.9
The extracellular forms of viruses face formidable challenges. The virion itself must be sufficiently stable to protect the viral genome during the passage from host to host and cell to cell, and yet, upon reaching the target cell and encountering the appropriate trigger, the virion must initiate programmed steps that result in the release of the viral genome into the appropriate compartment of the cell. For nonenveloped viruses, the conceptually simple mechanism of membrane fusion is not an option. . . .
After binding the cell surface, the virus is internalized through a clathrin-, caveolin-, and flotillin-independent, but actin- and tyrosine kinase–dependent, pathway. After internalization (and only after internalization), the virus releases its RNA rapidly from vesicles that are located within 100–200 nm of the plasma membrane without requiring endocytic acidification or microtubule-dependent transport. Our results have settled the long-lasting debate of whether PV [poliovirus] directly breaks the plasma membrane barrier or relies on endocytosis to deliver its genome into the cell. These results have also opened interesting questions for this important virus that await further investigation, including what characteristic of these endocytic vesicles near the plasma membrane triggers RNA release; and after release near the cell surface, how is the released RNA transported to replication sites.

• This Discussion starts by reiterating the problem and reenergizing curiosity before going on to nailing down the answer with the resolution.
• Contrast these OCAR Discussions to some that use LDR structure.

Example 8.10
We have identified a novel class of GGT inhibitors that are not glutamine analogues. Kinetic studies of the lead compound OU749 revealed that the mechanism of inhibition was uncompetitive relative to the γ -glutamyl substrate, indicating that the inhibitor bound the enzyme-substrate complex. In contrast to competitive inhibitors, which lose potency as substrate concentration builds, uncompetitive inhibitors become more potent as the substrate concentration rises in an inhibited open system. . . .
Development of less toxic GGT inhibitors, such as OU749, holds great promise for enhanced cancer therapy.

• The first sentence addresses the paper’s specific challenge and identifies the important result from the research: a new class of inhibitors. The Discussion elaborates on the properties and benefits of this group of chemicals. This builds to the resolution, which closes the circle back to the paper’s opening, which was about developing new chemotherapeutic agents.

Example 8.11:
In this paper, we have shown how to construct various D-branes in the type IIB plane-wave background that preserve half the dynamical supersymmetries of the background. . . .
The connection of the instantonic branes we have constructed with instantonic branes in AdS5 × S5 is more obscure. Understanding this could lead to an understanding of the relation between the D-instanton and instanton effects in the dual Yang-Mills field theory. It would also be interesting to understand the effect of the D-instanton on the plane-wave dynamics. Finally, one should be able to analyse the D-instanton contributions by considering the effects of the R4 and related terms in the effective low energy IIB action in this background.

• This Discussion starts by identifying the main contribution of the paper: constructing D-branes. It then develops and elaborates that result. The paper resolves by framing interesting new questions that grow from the work.
• Both OCAR and LDR work well for the Discussion — they each provide a coherent structure that allows you to develop a clear and compelling story.
• With any story you have a choice of structure, and the Discussion should form a story within itself.

9 The Resolution

• People remember the last thing you say. The resolution should be "take-home message," your strongest and most memorable words.
• A good resolution achieves this by stepping backward through OCAR: it reiterates the action, answers the questions raised in the challenge, and demonstrates how those answers contribute to the larger problem.
• Because last words are so pwoerful, people will accept whatever you put there as the take-home message. If you are not careful, some weak or extraneous thought that finds itself in the closing position can come across as your most important.

9.1. Good resolutions

• The first example of a good resolution is straightforward, walking backward through the OCAR steps without distraction or complication.

Example 9.1
{1} In conclusion, {2} our data suggest that Y-phosphorylated p27 can inhibit cyclin D-cdk4 complexes by two independent mechanisms: blocking access to the T-loop and disrupting the cdk4 active site directly. {3} Our model suggests that p27 Y phosphorylation is a molecular “switch” that would help turn cdk4 activity on or off. {4} Modulation of Y kinase activity would permit activation of preformed, inactive p27-cyclin D-cdk4 complexes by cdk7 and may be used to regulate cdk4 activity throughout the cell cycle.

• This resolution does a number of things well.
  • {1} The statement “In conclusion” is a flag, telling the reader that what follows is the resolution. Such road signs make it easier to navigate through a paper.
  • {2} This states that two mechanisms of inhibition are involved. This is the key result of this work, and it answers the question posed in the challenge.
  • {3} This statement interprets that result and synthesizes it into the idea that p27 Y phosphorylation is a “molecular ‘switch.’” That creates a simple message and an accessible intellectual model for how this compound works — switches turning on and off the processes that drive the cell cycle. This starts “widening the hourglass” by moving away from the specifics of how p27 inhibits, to what that means for cell cycle regulation.
  • {4} This finishes opening the hourglass by bringing the story back to issue the paper opened with — what regulates the cell cycle. It even puts the phrase “the cell cycle” at the end of the concluding sentence, closing the circle back to the opening sentence of the paper, which was: “Cyclin– cyclin-dependent kinase (cyclin-cdk) complexes drive progression through the different phases of the cell cycle by acquiring catalytic activity only at specific points.”

Example 9.2
{1} These templated nanostructured frameworks thus hold several advantages for the design and synthesis of devices. {2} Films can be selectively deposited through solution phase routes using the chalcogenide affinity to bind to gold. {3} Furthermore, the ability to control the elemental compositions of the nanostructured films allows the band structure of the inorganic framework to be tailored for specific applications. {4} Current research is underway to create composite materials using an organic semiconductor as the structure directing agent. Such materials would make good candidates for device applications such as photovoltaics. Moreover, it is likely that these same band energy trends will hold for nontemplated versions of chalcogenide glass semiconductors synthesized using Zintl cluster precursors. {5} As a result, the data presented here provide a basis to predicatively synthesize a broad range of semiconductors with desired band properties using Zintl cluster precursors and simple solution phase methods.

• {1} This is a statement of the overall accomplishment — the authors created a useful material. It gives a clear sense that this is the resolution and that they will flesh out this point in the rest of the paragraph.
• {2} In this second sentence, they state the key result from the work: “films could be selectively deposited.”
• {3} Here they start expanding back out, with a more general interpretation of that result.
• {4} The authors continue the widening process, going beyond their specific research and discussing the implications of the work. They expand it by pointing out that “these same band energy trends will hold for nontemplated versions.”
• {5} Here in their final wrap-up statement, the authors give the most general application of their work: “a basis to predicatively synthesize a broad range of semiconductors.”
• It then, however, carries out the same functions — it identifies the key result, opens up the hourglass, and resolves by tying back to the big picture of the paper’s opening. This paragraph creates a complete story within itself — opening, developing, and resolving, a strong approach for a longer paper and a longer resolution.

9.1.1 Concluding with a Question

• In each of the foregoing examples, the authors identified a result and explained its significance. Sometimes, however, the most important thing you discover is that there is a new question, one you hadn’t anticipated, that you want to pose to the community. Fine. Do it, but make the question concrete, and be clear about how it grew from your work — you didn’t fail to fill one knowledge gap but identified a new one. Ending with a concrete new question engages a reader’s curiosity and can be a powerful way to resolve a paper.

Example 9.3
{1} There was an extraordinarily large amount of ice bottom melting in the Beaufort Sea region in the summer of 2007. Solar radiation absorbed in the upper ocean provided more than adequate heat for this melting. An increase in the open water fraction resulted in a 500 % positive anomaly in solar heat input to the upper ocean, triggering an ice–albedo feedback and contributing to the accelerating ice retreat. The melting in the Beaufort Sea has elements of a classic ice–albedo feedback signature: more open water leads to more solar heat absorbed, which results in more melting and more open water. The positive ice–albedo feedback can accelerate the observed reduction in Arctic sea ice. {2} Questions remain regarding how widespread this extreme bottom melting was, what initially triggered the increase in area of open water, and what the summer of 2007 portends for 2008 and beyond.

• {1} The main part of this resolution states both the findings and conclusions using clear strong language: “solar radiation . . . provided more than adequate heat for the melting” and “The positive ice–albedo feedback can accelerate the observed reduction in Arctic sea ice.” There is no hesitation or weakness.

• {2} The resolution goes further to frame a series of questions about both the mechanisms involved and the implications for the future. However, rather than undermining the conclusions, these questions actually reinforce and extend them; they point the direction forward. They engage a reader’s interest. Even using a word like portend emphasizes the new question — it’s an ominous word.

• However, rather than undermining the conclusions, these questions actually reinforce and extend them; they point the direction forward.

9.2. Bad resolutions

• Now let’s consider bad ones — resolutions that fail in those core functions.
• You can be weak, distracting, or, at worst, you can actively undermine your conclusions.

9.2.1 Weak

• Weak resolutions fail to frame the conclusions. In this type of ending, authors usually synopsize their results and then tell you that they are important, but don’t clarify how:

Example 9.5
A proteomic evaluation of hummingbirds under simulated migratory conditions revealed evidence of several stress-associated processes: protein degradation in wing muscle tissues, depletion of metabolic cofactors, and enhancement of stress-response proteins. These results suggest that changes in the hummingbird proteome may provide new insights into the complex physiology of avian systems biology.

• This paragraph does a good job of synopsizing the results, but then it stumbles. Rather than synthesizing new knowledge, it skips that step. Instead, it simply tells us that the research is important and has implications beyond hummingbirds.
• They tell us that it “may provide new insights . . . into avian systems biology,” but they don’t tell us what those insights are! What did these authors contribute to the wider field of bird physiology and ecology? We’re left to figure it out for ourselves.
• To fix a resolution like this, you need to identify the new insights.

A proteomic evaluation of hummingbirds under simulated migratory conditions revealed evidence of several stress-associated processes: protein degradation in wing muscle tissues, depletion of metabolic cofactors, and enhancement of stress-response proteins. While hummingbirds migrate long distances over water without feeding or resting, it is physiologically stressful, and the birds’ ability to manage this stress may limit the distance they can migrate.

• Here, rather than trying to make a methodological but largely meaningless suggestion about how to study birds in general, the paper ends with a clear conclusion about what these data mean — migrating is stressful — and a suggestion for what they say about hummingbird biology and behavior, suggestions that clearly relate to other birds. This resolution says something concrete — it resolves.
• If the authors wanted to open the hourglass wider to explicitly encompass other migratory birds, they could modify the last sentence to make hummingbirds a member of that larger group:

“While many birds, such as hummingbirds, migrate long distances without feeding or resting, it is physiologically stressful, and birds’ ability to manage such stress may limit the distance they can migrate.”

• This adds “such as hummingbirds” to make it clear that they are an example; it also condenses “and the birds’” to “and birds’,” a subtle change that shifts it from referring to specific birds to birds in general.

Example 9.6
In summary, we show that X7 alters the expression pattern of extracellular proteases in the “flesh-eating bacterium” Streptococcus pyogenes, which causes necrotizing fasciitis. If the function of X7 can be fully established, it would likely deepen our understanding of this destructive disease.

• There's nothing wrong with not fully establishing its function, but don't end a paper by telling your readers what you didn't achieve.
• the fix is to make a concrete conclusion that synthesizes the results into knowledge and provides a meaningful take-home message:

“In summary, we show that X7 alters the expression pattern of extracellular proteases in the ‘flesh-eating bacterium’ Streptococcus pyogenes, which causes necrotizing fasciitis. This research may offer a route to developing therapeutic agents that would minimize tissue damage while antibiotic treatments were directly attacking the bacterium itself.”

• Note that this resolution, while ending with a strong message, is carefully constrained. It doesn't say that X7 is necessarily going to be that new therapeutic agent, and it only says that this “may” offer a route — it might not work in vivo. You can make a strong statement without overselling.

9.2.2 Distracting

• Some papers conclude with material that is distracting — ideas that should be in the Introduction or is already in textbooks and that neither synopsizes nor synthesizes the results.

Example 9.7
The mycorrhizal fungal hyphae extending out from tree roots can comprise more than 1/3 of the total biomass of microbes in the soil. They greatly extend the absorptive surface area of the root system and enhance total nutrient uptake by the trees. Additional work, however, is required to assess how much mycorrhizal fungi enhance the uptake of organic N forms in forest soils.

• These first two sentences are truisms that have been known for decades — textbook material, rather than results of this particular study.
• A second way a resolution can be distracting is by introducing new information at the end. The following might appear to be a strong resolution:

Example 9.8:
In arid environments such as East Africa, termites are critical “ecosystem engineers.” They collect resources such as nitrogen and phosphorus from far afield and accumulate it in and near their mounds, creating nutrient hot-spots on the landscape. These hot-spots may be sites for colonization by new seedlings of both the native savanna trees and for novel invasive plant species.

• The problem here is that invasive plants were never mentioned in the Introduction. The idea that termite mounds create invasion sites is interesting and important, but it must not be a new idea, first raised in the resolution.
• The resolution must close the circle back to the opening. Instead of closing the circle, however, this resolution goes haring off in a new direction.
• That’s great; developing new ideas while you are writing is exactly what Montgomery meant when he said that “clear thinking can emerge from clear writing.” Never close your mind to new insights about your work and its implications. But when you have them, go back and weave them into the opening and Introduction.
• From the perspective of getting your message out to the widest possible audience, surprise resolutions are a disaster.
• Everyone loses — plant ecologists miss useful information, and the authors lose citations.

9.2.3. Undermining your conclusions

• The worst possible way to end a paper is to actively undermine your conclusions, and yet this may be the most common way to end scientific papers. Many end by saying “more research is needed to clarify our findings.” Resolving a paper this way focuses on what you haven’t accomplished.
• we know our work isn’t perfect and that there are still questions about both the big issue of our opening and the small issue of our challenge. Uncertainties remain. But the resolution is not the place to discuss them.

Example 9.9
To conclude, 3-methyl-ambrosia offers a new approach for thyroid carcinoma therapy. Our data provide evidence on safety and in vivo activity of this compound in patients with this condition, although the proof for clinical benefit remains to be established in future clinical trials.

• In the first part of this passage, the authors tell us that they have a new therapy that appears safe and effective. Their take-home message, though, is that they don’t know whether it really works!
• Talk about destroying the story. This would have been much better as:

“While further clinical trials will be necessary to establish the full benefits of 3-methyl-ambrosia as a therapeutic agent, our data provide evidence that it is safe and shows in vivo activity against thyroid tumors. 3-Methyl-ambrosia therefore may offer a new approach for treating patients with thyroid carcinoma.”

• This version says the same things as the original, but strongly and positively.
• But it ends by highlighting the authors’ intended message: 3-methyl-ambrosia may be an effective new anticancer drug.
• This kind of “more research is needed to clarify our results” statement is fundamentally different from “concluding with a question” that I illustrated in examples 9.3 and 9.4. Those resolved by posing concrete questions that grew from the work. “More research is needed” poses questions about the work and makes it sound like the author didn’t complete the research.
• There are many ways to undermine your results, including expressions such as “but the importance of this has yet to be assessed,” “we hope that this review will simulate further research to answer the many unanswered questions,” “this topic deserves more research,” and so on. All of these use fuzzy expressions that suggest weaknesses in the existing work, rather than expressing substantive conclusions or pointing out clear new questions.

9.3. How to fix a bad resolution

• The solution is to first pare away the dead tissue — the fluff, the detractions, and the new ideas. When those are important, move them elsewhere. Then, condense your resolution to do three things: (1) synopsize the key results, (2) synthesize those results — show us how they answer your question, and (3) show us what this contributes to solving the larger problem. If you achieve those three objectives, each clearly and concretely, you will have a strong resolution that ends your paper with maximum punch.

9.4. Resolutions in proposals

• But many go all the way to pure LD structure. After describing the proposed experiments, they end, having said everything that seems to need saying. There is no synthesis, no wrap-up, no resolution.
• That's a mistake. Make space for a resolution paragraph that encapsulates the proposal, reiteratesthe big issue and explains how the components work together to address it — make the final pitch for why the proposal should be funded.
• They are thinking about how the pieces fit together, whether the experiments will work, whether this will really solve the problem, and, importantly, what to write in their review. This is your last opportunity to give them the words. To convince them to check “excellent;” or as the program officer holds your fate in her hand, hovering over the line on the whiteboard, asking “which side does this go on?,” to say “must fund.”
• We’ll never know whether that resolution paragraph was what gave it the million-dollar nudge, but I do know that a reviewer’s opinion is sometimes not solidified until the end. So end strong.