2. What’s Your Problem?

Now that you have questions, the next step is to answer them, right?

Not exactly.

In this chapter, you will begin to find and use primary sources, and you may find the answers to at least some of your questions. But answering questions will not be the primary focus. Educating your questions will be the focus.

The questions you have generated thus far are, by and large, less developed than they could be and will become. This has nothing to do with your abilities as a researcher. Rather, it is a fundamental part of this stage of research: your questions are underdeveloped at this point because you have not yet had the chance to conduct research into your subject matter. It’s to be expected.

Wait a minute! You might protest at this moment. Before, you told me that I need to generate questions in order to do research. But now you’re telling me that I need to do research in order to generate questions? That’s impossible. It’s an infinite loop. It’s a trap!

It’s not a trap. But it is true that it takes a lot of research to arrive at the right questions. And then it takes more research to answer these questions and generate new ones. In this early stage of research, the goal is not, as many assume, the generation of answers. It is about the refinement of your existing questions and the generation of new (and better) ones.

The goal of this chapter is to help you identify and articulate the problem underlying your many research questions. Accomplish this, and you will end up asking better questions, doing more significant research, and carrying it out more effectively.

Don’t Jump to a Question (or You’ll Miss Your Problem)

Over the course of generating, analyzing, refining, and adding to your questions, you may have wondered: How do I know when I’ve found my Problem? Do I really have a “Problem,” or have I merely compiled a random set of questions that don’t really add up to anything? After all, we’re curious about many things, but we don’t launch research projects to satisfy every curiosity. Nor should we.

Here’s a simple way to distinguish a problem from a random set of curiosities: if it changes by the day, week, or month, chances are it’s a passing curiosity. If it endures, it just might be a problem.

A problem is a nagging presence within you—one that disturbs, bothers, and unsettles you, but also attracts, compels, and keeps you coming back. It’s something that generates questions in your mind—questions that, no matter how varied and unrelated they might seem to an outsider, you know to be somehow interrelated, even if you can’t explain why. A problem is something that follows you around. It doesn’t care if you are a historian of France, a sociologist of the Philippines, or a literary scholar of India—it calls out for you to try to solve it. Your job is to give that problem a name, to identify a case of that problem that you will be able to study (given your personal abilities and constraints), and to figure out how to study that case so that you might arrive at a broader solution.

Researching a problem requires asking questions, of course, but (again, to state the obvious) a question is not a problem.

You can think of plenty of questions that have answers, but whose answers do not solve any problem. Asking and answering such idle questions is a waste of time, so you want to make sure that your questions are indeed problem-driven. This is why it is so important not to jump to a question.

A problem has several functions for the researcher, among them the following:

• It motivates you to ask questions about your topic.
• It determines which questions you ask.
• It defines the what/why/when/how of your engagement with your topic.
• It guides the path of your inquiry.
• It shapes the story you tell when the time comes to share your research results.

Up to now, you have been generating “first-draft” questions based on an initial foray into sources. But you want to make sure that you are asking questions that do more than just satisfy a personal curiosity. The next steps in this chapter will help you figure out

• how to improve the questions you have already generated;
• how to use sources to identify the problem motivating your questions;
• how to use your Problem to generate new and better questions.

We all know not to “jump to a conclusion”—an action prompted by prejudicial or hasty thought. We’ve all seen it happen, and we’ve all done it ourselves—we arrive at an argument or thesis about a sure thing even though we haven’t spent sufficient time thinking it over. And we end up being wrong.

What the early-stage researcher has to avoid is jumping to a question. You have generated many questions, and the risk now is that you’ll feel pressured to jump ahead and choose one prematurely.

What is your Research Question? You’ll hear this demand from other people, and eventually from a little voice in your head that tries to trick you into thinking that your project must have only one Research Question, and that you must settle on it early.

The Jump-to-a-Question Trap can be as harmful as the Narrow-Down-Your-Topic Trap.

Jumping to a question is like constructing a home without examining the ground on which it will stand. Your architectural plan might be stunning, the plot of land spacious, and the vista marvelous, but if you build on sand you are going to have serious issues when those sands shift. By the time issues show up, renovations may be costly, and you might find it impossible to relocate.

Stress-Testing Your Questions

Now that you have done the work of producing a multitude of questions—small, factual questions, ideally—you will still need to stress-test, refine, and winnow them out, removing any dead ends, enhancing the rest, and adding additional questions that will better serve your research process.

Think of a question as if it were a car. Before jumping into this vehicle, and certainly before bringing others along, you would want assurance that its steering and brakes have been subjected to rigorous testing. You’d want to know that prototypes underwent crash tests, over and over, until the manufacturer felt certain that the structure of the vehicle was ideally suited to protect the driver and the passengers.

Here are two ways to stress-test your questions and improve their soundness. The first focuses on language; the second is subject-specific and focuses on sources. We recommend that you tackle them in that order.

TRY THIS NOW: Run a Diagnostic Test on Your Questions

The goal: To ensure that the vocabulary, grammar, and phrasing of your questions are specific and unprejudiced so that they do not presume a certain outcome.

Rewrite your research questions with particular attention to the following:

1. Punctuation. Do your questions actually end in a question mark? Or have you phrased them in more general, and vaguer, terms like “This is an examination of . . . ,” “I plan to explore . . . ,” or “My project is about the question of . . . ”? If you find yourself articulating your questions as “I want to examine how” something happened, there is a fair chance that what you have are not really questions at all, but rather topics disguised as questions. Be more specific, and add a question mark.
2. Adjectives and adverbs. Do your questions rely on broad, generic, imprecise, or sweeping adjectives like “modern,” “traditional,” or “Western,” or adverbs like “scientifically,” “rationally,” or “effectively”? Try to cut such adjectives and adverbs out entirely.
3. Collective nouns. Do your questions depend upon collective nouns like “Asians,” “the French,” “students,” “women,” or “North Americans”? If so, do your best to replace these nouns with more precise demographic categories: women of what ages, students living where and when, North Americans of what background, socioeconomic status, race or ethnicity, or family status? You do not need to take into account all possible demographic variables, but you do want to try to include all of those that might make a difference to your project.
4. Verbs. Do your questions contain verbs like “influence,” “affect,” “shape,” or “impact” or passive constructions such as “was affected by,” “responded to,” or “reacted to”? In such cases, chances are high that you are building your questions in such a way that they rule out an entire set of possible answers and outcomes. Rephrase to avoid presumptions that could result in confirmation bias.

By the end of this process, your questions should meet these criteria:

• They should be clear, precise, and jargon-free. If your questions are too hard for a colleague or mentor to understand, this means that you (and not they) still don’t get what your Problem really is. Your shorthand might be hiding your significant specifics. Likewise, if your Problem is hiding behind jargon—words designed to make your questions sound “smart” and “important”—replace it with language that is clear and vulnerable. You and your intended audience should be able to tell what your research is driving at, even if this means your language might be less articulate and refined—and less concise—than you might like.
• They should be rooted in verifiable and falsifiable data. Your research questions should have integrity. This means that they should be inspired by fact, rather than by speculation, prejudice, or opinion. What are the facts that motivated these questions? Are they verifiable? Where and how could these facts be checked? Have you checked them yourself?
• They should be indifferent to the outcome. The best research questions are open, agnostic, unprejudiced. Put another way, a research question should not presume a certain answer. If yours does, rewrite it to eliminate that presumption.
• They should be clear about the subject. Your questions should not be reliant on broad categories of identity, such as “students,” “women,” “Europeans,” “Brazilians,” “Christians,” and so on. Refer to the suggestions above, and be as specific as you can be about the who in your question.
• They should be raw and undisciplined. At least for now. Although we encourage you to make sure that each individual question in your list is as precise, detailed, and grounded as possible, remember that, collectively, your list of questions need not be overly polished or coherent at this point. If the questions seem random to you, let them be random. If they seem unrelated to one another, let them be unrelated.

COMMONLY MADE MISTAKES

• Asking leading questions, which are phrased so as to predetermine the answer. These questions are motivated by unproven assumptions, and result in confirmation bias. The result of leading questions is that you inevitably find what you are looking for. (See the example below of How did X influence Y?)
• Asking advocacy questions, which promote a certain ideology (taken-for-granted worldview) or course of action. These questions take a position and encourage others to adopt it, irrespective of the actual facts of the case or which interpretations the evidence suggests are plausible. Example: “Why is ‘feminism’ a better analytical rubric than ‘romance’ for understanding Joan Didion’s novels?”
• Forcing all your questions to “make sense” or “add up.” Don’t worry. That part will come soon.

Leading questions are so common and so prejudicial to the research process that it’s worth looking at one extended example. Maybe you’ve seen a version of this question before:

How did X affect Y?

Consider this example:

How did the ruinous taxation policies of King Louis XVI during the 1780s erode popular support for the aristocracy and pave the way for the French Revolution?

Wow, that is one “educated” question! In order to pose it in the first place, one would already need to know a considerable amount about French history.

But take another look. See any issues? When we ask How did X affect Y? the implied answer is that X did affect Y, and all that remains are questions of how and to what degree. Building a question this way creates a major weakness. The researcher at this point has not yet established that such influence existed in the first place. The very phrasing of the question rules out the possibility that X didn’t affect Y at all. If it turns out that it didn’t, you’d have one heck of a short paper.

Let’s say your hunch is still that X did affect Y. It might have. However, you couldn’t know at this point, having not yet done the research. What you want to avoid is building your questions in such a way that you actually need this “influence” to exist in order for your questions to be viable. Almost inevitably, you will end up discovering specious “proof” of influence in primary source material, misleading both your readers and yourself.

If you detect shortcomings in your question, try to repair them. If your question is actually a topic masquerading as a question, reword or restructure it. If you are relying on abstract nouns, adjectives, or adverbs, substitute specific words. Articulate your question without using any sweeping words at all. And if your choice of verbs is committing you to research outcomes too early, change them.

TRY THIS NOW: Use Primary Sources to Educate Your Questions

The goal: To learn how to run keyword searches designed to enhance or “educate” the questions you are asking about your topic. These searches uncover primary sources relevant to your research that themselves contain new keywords you were previously unaware of (thereby enabling you to run follow-up searches to reveal even more, and more useful, primary sources).

Running language diagnostics like you did in the previous exercise is a first step that will help you avoid common errors when posing your questions. This next exercise requires you to delve back into your specific subject matter and into primary sources.

Thus far, we’ve been keeping you somewhat at arm’s length from primary sources. In chapter 1, we specifically encouraged you not to delve into them too deeply in the “Search Yourself” exercise. We now encourage you to dive in, but in ways you might not expect. Rather than trying to use primary sources to start answering the questions you’ve come up with, we want you to use them to develop, refine, and expand those questions. You’ll eventually start using primary sources to answer questions, but at this early stage, we suspect, your questions still need more refinement before you start investing large amounts of time and energy in answering them.

How do we use primary sources to strengthen and “educate” our questions? The answer is simple: primary sources alert you the existence of other primary sources, exposure to which helps you ask more mature questions about your subject. By contrast, researchers who “jump to questions,” and then dive headlong into answering those questions using primary sources, run the risk of confining themselves to a kind of intellectual and empirical bubble.

Let’s imagine that you’re interested in one of the following topics:

• turn-of-the-twentieth-century African American literature
• the history of artificial intelligence
• food culture in twentieth-century Hong Kong

Let’s further imagine that you’ve already done the hard work of transforming your initial topic into a set of specific questions, and that you’re now collecting and exploring sources. You start by running searches in a digital repository of historic newspapers containing hundreds of fully text-searchable periodicals from across the world.

But then you hit a stumbling block. Practically all the search results for “food AND Hong Kong” come from the 1950s onward. Or a search for “African American AND literature” yields many articles and reports from the 1980s onward, but almost nothing from earlier periods. A search for “artificial intelligence” returns lots of materials from the 1980s onward, but very little before.

What is going on? Common sense tells you that there were African American writers circa 1900, that food culture in Hong Kong predates the 1950s, and that research into AI began before the 1980s. Why is your search failing?

In this case, the cause is simple: the keywords you’re using are anachronistic. That is to say, these are terms that people use here (in your hypothetical location) and now to describe the identities, places, and subjects you are trying to find primary sources about. But these are not the words that people necessarily used in the past, or in other places. “Artificial intelligence” is a term we use today to describe a branch of computer science. It isn’t necessarily the one used by the scientists who gave rise to this field. They more often used terms like “Systems Thinking,” “Machine Intelligence,” and a number of others. As a place-name, “Hong Kong” has been in usage for a very long time, and yet as an English-language spelling, it has changed a great deal (decades ago, you would have been much more likely to see it spelled “Hong-Kong,” with a hyphen, or “Hongkong,” as one word). Likewise, the term “African American” was popularized only in the 1980s, prior to which one would have encountered terms like “Afro-American,” “Negro,” “Colored,” and other appellations, many of which are deeply offensive today.

In this earliest stage of finding primary sources, then, your main goal is actually not to start answering your questions, but to use the primary sources you do find to reveal new keywords that you did not know existed—keywords that you can then feed back into the search process, in order to uncover more and better primary sources, more and better keywords, and most importantly of all, more and better questions.

This may seem like a daunting recommendation. After all, even if your search terms are “imperfect,” they may return thousands—even tens of thousands—of results. Should you really be expected to read, notate, and cite even more sources?

Not necessarily. And fear not—we’ll get to source management. Right now, your goal is to identify omissions in your search inputs, so as to eliminate false negatives in those searches’ outputs. Big picture, you’re improving your grasp of your topic by eliminating blind spots.

Whenever you do a keyword search, ask yourself: Are there other search terms I should be using? Might there be different spellings of the search terms I already have? You need to be as confident as possible that the search results you are getting are broadly representative and reflective of available primary sources, and not the by-product of narrow or unrefined searches. If your search results all cluster within a narrow time period (as in our example) or were produced in a very small number of places or were written by a very small number of people, chances are that something in your search process is the cause. Phrased differently: Hong Kong existed before the 1950s, as did African American writers before the 1980s, and so the peculiar clustering of your results has nothing to do with “reality” at large, and everything to do with the way you searched. And if you didn’t stop to refine your search, instead racing ahead to read, notate, and cite all of the materials you did find, your overall research project would be woefully incomplete.

Here are some techniques to help you use primary sources to refine your keyword searches.

The Art and Science of Keyword Search: A Few Tricks

Improving a keyword search might sound like a rather straightforward process, but there is a devilish paradox at play: most of the primary sources you discover that contain “present-day keywords” (e.g., “artificial intelligence,” “African American,” “Hong Kong,” . . .) will not contain the other keywords that you need to find (“Hong-kong,” “Hongkong,” “Afro-American,” . . .). In most searches, the situation is all-or-nothing. Either the keyword you used in your search is present in the primary source, and thus the source appears in your search results; or the keyword is not there, and it simply doesn’t. Here’s how to get around that impasse.

Take Advantage of Category Searches

In certain databases, you might be fortunate to come across materials that are accompanied by metadata (data about data), crafted by librarians and archivists whose goal it is to make sources more discoverable to researchers like you. In such cases, you might find a primary source that contains the term “artificial intelligence,” and then discover that it has also been “tagged” with this same keyword in the database. By clicking on this tag, you gain access to all of the other primary sources in that database that were categorized that way—including those that may not contain the term “artificial intelligence” at all! This is one way to get from a source that contains only the keywords you used in your search to another source that contains none of the keywords you used.

Here’s what to do: after you run your search, and receive your results, sort the results chronologically, and then explore only the results that come before the 1980s—the time when, in your preliminary search, they seemed to disappear. As you scan through these titles, take notice: What words show up in the title? If you are able to read the work online, scan the table of contents, the preface, the introduction, and the index. What words, terms, and vocabulary are used? Are there any words or phrases you notice that might, if you were to run them through your database search, yield other hits that your first keyword search did not? These are your new keywords. Write them down.

A caveat: metadata, too, is the product of context and should not be taken as definitive. Categories are cultural constructs, including those created by librarians and archivists. The categories in the metadata they create are shaped by stances and worldviews and protocols, and thus should never be taken as the “final word” on any subject. Always assume there is more, and that no one can do your work for you.

Locate Self-Reflexive Sources

In some cases, you might be fortunate to find a primary source, like a historical dictionary, that explicitly addresses the shifting nomenclature surrounding the very topic the primary source is about, outlining for you the varied ways a given idea, place, community, practice, or the like has been named and renamed across time and space. Moments like this are joyous, since they open countless doors that the researcher need only walk through!

Even in such cases, however, remember that a primary source still is subject to its own limitations. No source will document all of the variants in terminology that might be useful to your searches. No source (as we explain below) should be taken as the final word on the subject. It still falls to you to determine whether or not the primary source in question is empirically accurate. Every source possesses its own stances, worldviews, and perspectives. But for your current purpose of finding more generative keywords, the source can be useful to you whether or not its data or conclusions are accurate, so you can defer judgment on those questions for the time being. The goal for now is to determine if this source will lead you in the direction of further primary sources that you wouldn’t have been able to find otherwise.

Keep Track of Your Keywords and Searches

As you discover and try out more and more keywords—and even a smaller-scale project can produce hundreds—it’s easy to lose track of them and get overwhelmed. The other fundamental aspect of this process involves, alas, the unglamorous world of record-keeping.

Did I try this keyword before? I can’t remember. Did I try this keyword in this database? Not sure. When was the last time I ran this keyword in this database? No idea.

The dangers of missing something are real, since databases are continually updated and expanded, and since some projects can take many months—or years—to complete. You can imagine how many hours you might end up wasting repeating searches you’ve already done.

Fortunately, there’s a simple solution: track your searches using a table. Here’s how, in three steps:

1. In the rows on the left side, enter the keywords you plan to use.
2. In the column headers, enter all of the electronic databases or library catalogues you plan to explore.
3. Inside each cell, keep track of when you ran a particular search. Enter the date of your search, and perhaps also a brief note on the number of results you found.

The result is a huge time-saver, and better research results: you will always know, at a glance, which searches you have run, or still need to.

Table 2. TRACK YOUR KEYWORD SEARCHES

For more keyword search tips, tracking sheets, and a downloadable version of table 2, visit whereresearchbegins.com.

As you continue to use primary sources to further “educate” your questions, two other helpful things will inevitably happen: you will end up answering some of your questions along the way, and you will find that some are not actually worth answering. In other words, you will discover that some of your initial questions can be scrapped. This is precisely what you want to happen.

The process can feel miraculous. As you stress-test your questions, you learn more about your subject matter. And as you learn more, your instincts regarding your subject matter improve. In “educating your questions” you are educating your instincts. When an experienced mechanic tells you, “Sounds like something’s wrong with your transmission,” you listen closely because their ability to detect anomalies is highly refined. To the average car owner, any loud noise might prompt us to ask the generic question “What’s wrong?” Educating your questions will help you to hone in on “real” questions, and you’ll happily discard the ones born of naivete.

TRY THIS NOW: Make Your Assumptions Visible

The goal: To become aware of the assumptions you bring to your research project and use them to identify the problem that motivates your research questions.

Now that you’ve analyzed your questions using the two techniques described above, there is still one more thing you need to do: identify the assumptions that underlie your questions, make them visible, and make peace with them.

You are not a blank slate. You arrived at your topic and your questions with a whole mess of assumptions. This is a natural thing—a good thing, in fact. After all, these are the reasons why you thought the topic is interesting and why you think your question is the right one for you. Everyone brings their own baggage to a research project.

Welcome to Baggage Claim.

Some teachers take it as their mission to “shatter” all your “illusions” about the world.

You believe the Vikings were a horde of marauding savages? Behold as I tear the veil of ignorance from your eyes!

You believe that Japanese society is homogenous? Watch as I reduce your prejudicial views to dust!

Dispelling misconceptions can be useful in many pedagogical and research settings. Yet the dispelling process, however well meaning, can have an inhibitory effect. Watching a fellow researcher or student get “disabused” can make others want to keep quiet to avoid being embarrassed themselves. For the researcher, the “disabuse” model can also lead to the unproductive belief that assumptions are the enemy—that they are shameful things to keep hidden, obstacles to be overcome, or evidence of incompetence.

Self-Centered Research is premised upon a very different approach to assumptions, as follows:

1. Assumptions should be made visible, and thus vulnerable.
2. Assumptions should not, however, be stigmatized, silenced, or driven underground, since this, counterintuitively, encourages holding on to them more tightly.
3. Assumptions are fuel to be consumed. Using them, you can achieve two goals at once: you can move in a new direction, and you can exhaust your assumptions in the process (meaning that you will eventually need new fuel).

Your assumptions about the world—even the most naive or negative—serve you at this point in the research process. To set out on a research quest with no assumptions at all would be like trying to sail on a windless day. Assumptions are the wind in your sails, and you need to channel them to keep your voyage on course.

Before evaluating your assumptions—which you will do shortly—thank them for helping you take note of things. They are the reason that any search results jumped out at you at all. They are what helped you notice a detail in a primary source. It was the gaps between your assumptions and the world as it really is that gave rise to all those specific research questions. Your assumptions shape your expectations about reality. And when those expectations are not met, it’s time to pay attention.

So let’s get to work on making your assumptions visible, and vulnerable. Here’s what to do:

1. Review your most recent set of questions and ask yourself: For each of these questions, what has to be true in advance in order for me to ask this question in the first place?
2. List the small questions/things you noticed, and write down the assumptions you may hold that helped you notice each in the first place.
3. Make a list of the assumptions you bring to this particular question, and sort them into the following categories:
1. a. Assumptions you want to work with, for now
2. b. Assumptions you want to discard right away
3. c. Assumptions you are unsure or ambivalent about
4. Write two lines to justify your choice for putting each assumption in a particular category.
5. Now go back to all of the questions in your list whose underlying assumptions fall into category A. Since these are built on assumptions that you, having reflected on them, feel safe in maintaining, then these questions are good as they are.
6. What about questions whose underlying assumptions fall into category B? Although you might be tempted to, do not throw them away just yet! If you find them to be based on weak, prejudicial, or unfounded assumptions, try to rephrase them so that they aren’t. Can they be rebuilt as more grounded, open-ended questions? Try to improve them before you discard them.
7. As for questions built on category C assumptions, these fall somewhere in between. Most likely you would want to keep them in your list, but perhaps flag them, as reminders to yourself that you want to keep an eye on them, and revisit them as your research deepens.

To keep things organized, try creating a chart like the one in table 3 for each question, in which you identify and analyze underlying assumptions and revise the question as needed.

Table 3. MAKE YOUR ASSUMPTIONS VISIBLE

Here’s an example: Imagine that one of the things you noticed and wrote down was a short quotation that jumped out at you in a letter written between two friends in the year 1944, during World War II. Perhaps there was a particular passage or sentence that jumped out at you—perhaps one of the friends cracked a joke about the war, and it stuck with you.

In this exercise, it’s your goal to brainstorm why the passage or sentence jumped out at you, by contemplating what assumptions you may hold that this quote contradicted. Feel free to speculate. You are not expected to “know yourself completely” right away—that takes time. Perhaps you think that people living during World War II would never have dreamed about making jokes about a conflict that, by the year 1944, had already taken the lives of millions and destroyed the lives of countless more. Or perhaps you assume that, during wartime in general, people are allergic to humor itself, and prefer to carry themselves in a somber manner befitting the gravity of their situation. Or perhaps you assume that there are some episodes and experiences in history that are so horrific—the Holocaust, the Armenian Genocide, the slave trade—that no one would dream of speaking about them comically.

Write down every possible reason why you might think what you think, even if you’re uncertain, without judging them as good or bad. The point here is not to “expose” your assumptions in a negative way. Rather, the goal is to bring to the surface those parts of your thinking that remain invisible yet influence how you think.

COMMONLY MADE MISTAKES

• Not identifying or divulging the assumptions motivating your research questions—for any reason, including embarrassment or self-consciousness. Remember: you’re admitting these assumptions to yourself so that you can improve your own thinking. There is no external judgment here.
• Not attempting to revise or restructure a research question based on category B assumptions.
• Dismissing or throwing out category C assumptions, instead of examining them as a type of self-evidence. Remember: the gaps between your assumptions and the world as it really is can generate useful research questions.

TRY THIS NOW: Identify the Problem That Connects Your Questions

The goal: To identify the problem underlying your multiple draft research questions.

Now you’re ready to take a key step. In chapter 1 you “searched yourself” to find questions within a topic. Now you’ll search yourself again, but this time with more self-evidence. By now, you have completed several exercises to produce a large number of questions about facts related to your project. What you now want to figure out is, What is the problem that connects your questions?

Try to think flexibly but rigorously. What relationships can you find between the different questions and fragments you have created and gathered thus far? What motivates your search for these particular facts? You could have asked any questions about this topic—why these? Which questions are the most compelling to you (and which seem less important)? Figure this out, and you’ll have accomplished a major breakthrough: you’ll have identified the underlying pattern that connects all (or most) of your questions in a coherent whole. In other words, you will have found your Problem.

Try this procedure:

1. Lay all of your questions out in front of you.
2. Do not try to answer all those questions for now. Instead, ask yourself: What are the shared concerns that connect these questions?
3. Step outside yourself. If you were someone else looking at these questions, what might you speculate are the deeper questions that connect these small questions?
4. Write down those questions.
5. If necessary, prioritize your questions by degree of specificity or generality, as medium-level or high-level questions. These questions should be more general than the specific factual questions you generated earlier.

The higher-level questions might not all add up. Don’t force them to. But be creative, and spend some time on this. What are the parent categories that connect two or more of your questions? The connective tissue might not be obvious immediately. Finding it might require thinking counterintuitively.

COMMONLY MADE MISTAKES

• Trying to answer your multiple questions, instead of focusing on identifying the shared concern that underlies them.
• Not thinking beyond the particular topic or case, and ignoring a more fundamental concern.

SOUNDING BOARD: Get Leads on Primary Sources

When you are searching for your Problem, or verifying that the problem you’ve been working with is the right one for you, it might still be too early to talk to a Sounding Board about your assumptions. As we mentioned above, the tendency of experts and authorities to “disabuse” us of “bad” assumptions is so prevalent that you will probably want to delay that conversation for the time being.

What your Sounding Board can help you with at this stage is finding primary sources that you can use to educate your questions. Above, we’ve provided a few examples of databases you can use for the exercises in this chapter. Describe those exercises to your Sounding Board, and ask for suggestions of other databases or archive catalogues or repositories of primary sources you might use.

You Have a Problem (in a Good Way)

You have now taken a close look at your many factual questions and grouped them under parent categories by shared concern. You have formulated higher-level questions motivated by these concerns. The key concern that overshadows all others might have emerged in a flash or intuition. Or perhaps you’re still trying to decide which of them is the most important to you. If you feel like you don’t yet have enough self-evidence, you can of course repeat the exercises in this chapter. But even if you think you do, you might still wonder: How do I know when I’ve truly discovered my Problem?

A problem is never a fleeting thing. Rather, it is something that is sustained and enduring. To you, it can’t be easily dismissed or ignored. Frida Kahlo painted surrealistic self-portraits because she was driven by a problem. In the world of music, John Coltrane worked on A Love Supreme, and Billie Holiday sang “Strange Fruit” because they were driven by problems. Bob Dylan entered a “blue period” because of a problem. Researchers are just the same.

Problems are good things. They are good to have, good to worry about, good to mull over. The problems we carry around with us can be thought of as the productive frictions that happen as we move through, and rub up against, existence itself.

Ultimately, however, the final decision can only come from you. Only you can know whether or not the cluster of fascinating questions you’ve generated thus far add up to a problem, or just a highly sophisticated and interesting set of curiosities.

You may well have multiple problems, but for now let’s just tackle one problem at a time. We’ll discuss what to do with the others in the final chapter.