17

There is a pattern I keep seeing and I'm very curious about the reason behind it.

Let's say you're going to read a paper on astronomy: In order to understand it, you need to know some basics about the topic. Similarly, a paper on electronics will require some basic knowledge of the topic if the reader wants to have any chance of understanding it.

But I keep seeing research where, at the introduction the prerequisite knowledge is defined, as if the reader was unaware. At first, I was assuming that it might be similar to legal documents where you will define the terms when it comes to their meaning, coverage and exclusion, but the more documents I read, the more I realize this is not the case.

Let me take some simple examples here, about sleep apnea:

(1) Obstructive sleep apnea-hypopnea (OSAH) is characterized by episodes of complete or partial upper airway obstruction during sleep.

(2) OSA occurs when there is obstruction of the upper airway causing cessation of airflow

(3) ... is characterized by momentary cessations in breathing (apnea) or significant reductions in breath- ing amplitude (hypopnea) caused by an obstructed or collapsed upper airway

I can find many many links exhibiting this and it seems a lot more prevalent in medical literature.

For example, when reading a paper about technology, you don't find a paragraph starting with "electricity results from the existence of charged particles...": there is an assumption you know what electricity is.

Content space, and the reader's time, are wasted to define the basics even though the contents are usually incomprehensible to anyone that wouldn't know the basics anyways.

Can anyone explain why this seems to happen more in some disciplines than others?

Azor Ahai -him-
  • 30,111
  • 9
  • 88
  • 115
Thomas
  • 327
  • 1
  • 2
  • 5

10 Answers10

106

That's what an introduction is. A good introduction will (1) start with the basics that everyone knows, (2) present the problem that we're trying to solve, and then (3) set up the rest of the paper (how this is done may be field dependent -- it could be a summary, or just an overview of the paper's organization).

These introductions should be concise, and they are not necessarily intended to be comprehensible to a general audience. For example, your first article defines central sleep apnea as follows:

Central sleep apnea (CSA) syndrome is characterised by recurrent episodes of apnea in the absence of upper airway obstruction during sleep which can occur in association with alveolar hypoventilation (i.e., hypercapnic), or it can be normocapnic or hypocapnic

Anyone interested in this study likely knows what CSA is, but seeing this definition stated gives me a good orientation -- I now know exactly what we're talking about and am ready to "drill down" to the specific things this study considered. It's also helpful to see the definition stated using the same language that the paper is going to use. But note that this definition is not really intended for non-experts: "normocapnic" and "hypocapnic" are not words that most people know.

Next you say:

For example, when reading a paper about technology, you don't find a paragraph starting with "electricity results from the existence of charged particles...": there is an assumption you know what electricity is....Can anyone explain why this seems to happen more in some disciplines than others?

I disagree with your premise. Sure, physicists probably don't define electricity just like the doctors in the above study didn't define sleep. But a well-written introduction does start with "the basics that everyone knows." For example, I pulled up the first link from the first open-access electronics journal that google returned, which led me to this article, the introduction of which begins:

The increasing global energy demand over the last 20 years means that electric energy is gaining more and more importance in our daily lives [1–3]. Furthermore, when taking into account the imminent emergence of the electric vehicle (EV) [4–6], there is no doubt that the electric energy sector will be of great importance in the immediate future. Inside the electric energy sector, the power converter is a key device whose design and behavior directly affect its efficiency, its cost, and the size of the final solution

All of this is definitely "the basics that everyone knows" -- in fact, this seems even more accessible than the sleep apnea introduction.

One last comment: not all introductions are well-written, so occasionally, your observations can be explained by "this author did a bad job." One example that comes to mind is papers about neural networks (especially deep convolutional neural networks, and especially in journals that are not "about machine learning"). Giving a clear explanation of what these neural networks are and how they work is very difficult, and most authors do not want to allocate many lines to the task. So, the result is exactly as you say: they give a thoroughly useless explanation which is incomprehensible if you don't already know how such networks work, and not needed if you do already know. One must make judicious choices in deciding what material to present in the paper versus what material should be left to references.

cag51
  • 67,924
  • 25
  • 181
  • 247
  • 33
    +1. A common advice is that introductions should be "funnel-shaped": start broad, then narrow things down to the concrete problem statement and an overview of the solution. – lighthouse keeper Feb 07 '22 at 10:52
  • 6
    Good answer. I'll also note that many methodologies are widely applicable regardless of the domain area, whether working in medicine, business, pure science, etc. A statistician or someone designing an experiment may well be interested in the approaches taken in the example manuscript, even if they know absolutely nothing about CSA specifically. The domain background can be very useful for those who work across domains. – Nuclear Hoagie Feb 07 '22 at 21:03
  • 3
    I think this definition "handshake" is useful in ANY kind of technical conversation, not just papers. Half the job for any good IT helpdesk engineer is making sure they're speaking the same language as the caller – Indigenuity Feb 07 '22 at 22:57
  • Incidentally, let me suggest that common sense sentences as those related to "increasing global energy demand" are OK in the Introduction but definitely out of line in the abstract, where space is precious. – Miguel Feb 08 '22 at 14:57
  • Well said. I'd add that the introduction is a key part of the paper in that you should be following the "tell them what you're going to tell them; tell them; then tell them what you told them" model. – Karl Wolfschtagg Feb 08 '22 at 23:05
  • I was thinking: do definition change over time? If this is not so uncommon in medicine it might be good to include it. – lalala Feb 09 '22 at 08:52
44

Content space, and the reader's time, are wasted to define the basics even though the contents are usually incomprehensible to anyone that wouldn't know the basics anyways.

Do you remember when you were an undergraduate (or even worse, high school student) reading your first research paper? If your experience was anything like mine then you might have had thoughts such as "I have no idea what this means" or "I understand this paper up to exactly this point since I have seen that equation in my textbook, but I don't understand anything beyond that at all", or even "Why do authors use such complicated sentences instead of plain English?"

Have some mercy on people less expert than you - you can skip past things that are obvious much faster than they can learn things which are not obvious.

Allure
  • 127,528
  • 50
  • 325
  • 493
  • 1
    Similar experience even for experts, especially when reading outside their area(s) of expertise. I thoroughly appreciate earnest introductions that don't assume knowledge and which provide enough for a layperson to comprehend what is discussed. As Einstein (apocryphally?) said: "If you can't explain it to a six year old, you don't understand it yourself". – stevec Feb 08 '22 at 02:13
  • 6
    This is particularly true in medicine, where conditions frequently straddle the line between multiple specialties and require doctors to occasionally dip their toe into new areas. They learned these terms/concepts in school but don't encounter them every day, so a quick reminder of basic terms/concepts can be enough to jog the brain and get them on the same track as the writer. – bta Feb 09 '22 at 03:22
  • That'd be the entire damn undergrad. And usually it goes like "Consider 2+2=4" me, confidence boosted: "I'm with you so far". The paper would then continue "This can be viewed as a factorization of a pseudo-Riemannian manifold using Ruse-Lanczos identity (a bunch more jargon goes in there)". Me, a poor undergrad: .... It is not too bad if this change happens in two paragraphs and not just two sentences - at least then you get a shot at extracting some keywords and looking them up. – Lodinn Feb 11 '22 at 04:41
24

Understanding a paper is not a binary. I challenge the assumption of the question that someone who doesn't already know the "obvious" facts won't understand a good part of the study.

I'm no medic, and before reading I knew very little about sleep apnea ("it's some problem that makes people sleep badly and snore I think...") but I opened one of the papers you link, and after just a quick scan I have a reasonable idea what they were trying to do and what they found. If pressed I could probably offer a summary that could go in a funding request or something.

I don't know the context of the study in the wider field. I wouldn't spot any shortcomings, aside from perhaps botched statistical analysis. I would only have a basic idea about how to compare it with another study. And certainly, nobody is going to rely on me to review it. However, all that said, it's not hieroglyphs to me, and I learnt something reading it.

In addition, if I found I wanted to know a bit more, the author has given me some terms that I could search to fill out my knowledge further. So that was helpful.

As mentioned in the other answers, these statements can serve to ensure other expert readers know what basic definitions you are using. But they do also expand the group of people who may read the paper and make some sense of it. You don't need to understand every last detail of a paper to get something useful from it. What with most of us being publicly funded, perhaps we have a duty to take some steps to make research accessible, and basic definitions are easy to provide.

Clumsy cat
  • 6,926
  • 2
  • 27
  • 37
  • This would likely be very different if you presented a paper on say, knot theory. But I'd be interested to see if papers on knot theory did the "broad intro" thing at all.... – Clumsy cat Feb 07 '22 at 11:12
  • 1
    Despite my years as a boy scout and a sailor, I'd never heard of knot theory as an academic field. So I looked it up. The first couple of paragraphs in Wikipedia (https://en.wikipedia.org/wiki/Knot_theory) do an excellent job of providing a very quick overview/introduction to the subject, understandable to my sailor/boy scout mind. The broad intro thing is possible – Flydog57 Feb 07 '22 at 22:40
  • 1
    @Flydog57 yes, that is a very nice intro. The point of the comment was that papers on the subject don't often even try; https://arxiv.org/pdf/1806.05168.pdf Perhaps because even if they did, I don't think I'd get much from the rest. In this regard, knot theory is knot like medicine..... – Clumsy cat Feb 08 '22 at 09:48
17

The definitions of medical terms (and terms in other disciplines) are reasonably likely to change over time, so even if every reader today knows the meaning of a term, it does not follow that every reader in the future will know the same meaning.

That said, any paper (or really any verbal explanation of anything) should include some statements of things the explainee already knows, because the way humans learn is by associating new ideas with things we already understand, and so an explanation needs to tell you which old ideas the new idea is (or should be) associated with.

Consider a mathematical proof: these often contain a statement of something the reader already knows (e.g. a definition or theorem), because although the reader knows it already, the reader doesn't know a priori that that particular definition or theorem is the one that justifies some later step in the proof. And that's not just true in mathematics; every paper is practically an argument for why its conclusion is correct, and arguments should state their premises, even if the premises are well-known facts, so that the reader can follow the reasoning of the argument.

kaya3
  • 2,149
  • 5
  • 16
  • Consider a mathematical proof: you start by stating definitions - A lot of presentations of proofs I see in math papers don't start with definitions. Is this what you meant? – Kimball Feb 07 '22 at 14:39
  • @Kimball: As I'm sure you know, the definitions are usually not just before proofs, but incorporated into the introductory material (examples for a recent notion), although for what it's worth, older papers (I'm talking about 50 or more years) often sprinkled definitions of terms throughout the paper as needed, which sometimes made it difficult to find such a definition when encountering the term in a later section of a paper that you're only browsing through for something or other. – Dave L Renfro Feb 07 '22 at 15:27
  • 1
    (+1) especially for first sentence. I was thinking this myself when I read the question (yesterday), especially given the OP's medical example, as I suspect these types of terms are not decades old and are also likely to be different in a few decades, although I suppose in this field (as opposed to mathematics, for example) people aren't going to be paying much attention to papers several decades old. My guess is that it's mostly inertia from conventions in the field, rather than explicit considerations given in the comments/answers here, that mostly drive what one finds in publications. – Dave L Renfro Feb 07 '22 at 15:37
  • 1
    @Kimball Indeed, many mathematical proofs don't begin by stating any definitions, and I'm not specifically thinking of proofs in research papers, where space is more limited. Nonetheless, it's very typical for mathematical proofs to begin by stating the definitions which the later steps in the proof derive from. – kaya3 Feb 07 '22 at 15:57
  • My point was the question is about writing papers, which means how to present something. So if you're comparing writing a paper to presenting a proof, I think you need to explicate the analogy a little more to explain why the introduction of a paper should start with (or just contain?) definitions, at least from this mathematician's point of view. (FYI many math papers don't start with definitions in the intro, though many do.) – Kimball Feb 07 '22 at 17:46
  • @Kimball The question is about why people do write this way, not about how they should write; I'm not proposing that mathematical proofs should begin by stating definitions, just observing that they typically do. Surely not all of them do. If a definition isn't used in a proof then there is no need to state it; imagine a proof of the form "all foos are bars, and no bars are bazzes, therefore no foos are bazzes"; the reasoning does not depend on the definition of a foo, a bar or a baz, so there is no need to state those definitions, but nonetheless the premises are stated even if everyone ... – kaya3 Feb 07 '22 at 17:49
  • ... knows that all foos are bars. I have said that arguments "should" state their premises and explanations "should" include statements of things the reader already knows, but I only mention mathematical proofs as an example of this being done (where definitions are the things the reader knows), not to say that definitions "should" be the things the reader already knows in a mathematical proof. – kaya3 Feb 07 '22 at 17:53
  • I've just edited that paragraph to hopefully make it clearer what I'm trying to say. – kaya3 Feb 07 '22 at 18:01
10

Simply, it is that it is both standard practice and wise to define the technical terms you intend to use in a paper so that there is less chance of misinterpretation. What is obvious to you will not be obvious to others and, you need to admit, that in come cases, what you think is "obvious" is actually wrong.

Each element of the medical term you describe is essential in some way and should be stated. Alternatively, such terminology can be introduced by citing another work in which the definitions occur.

To do otherwise is probably malpractice.

I'll also note that some things are "common knowledge", such as most things taught in an undergraduate curriculum and certainly anything taught in most secondary schools. These things don't don't normally need definition when used as the definitions are both standard and easily available to anyone. There are exceptional cases, of course, in which even the definition of electricity needs to be specified, such as when it is used in some esoteric (non common knowledge) way. Scholars are familiar with the concept of common knowledge because such doesn't normally need referencing or citation in a paper.

Buffy
  • 363,966
  • 84
  • 956
  • 1,406
  • 6
    I disagree that things taught in undergraduate curriculum are "common knowledge". Or more accurately, I disagree that the conventions, names and notations regarding things taught in undergraduate curriculum are common knowledge. Researchers assume too often that the way they were taught things is the only way. It's never the only way. There is always a different convention used somewhere else, and not stating the convention at the beginning of a paper is always bad and makes things confusing for no reason. – Stef Feb 07 '22 at 11:55
  • 3
    Along similar lines to @Stef: which undergraduate curriculum? Careers aren't always linear and people can drift e.g. chemistry<->physics<->engineering. Also even if the definitions are common, they can still be necessary to establish that the writer is using the common definition and not a technically correct but confusing one. This is especially true where industrial usage conflicts with research usage. (+1) – Chris H Feb 07 '22 at 15:25
  • 1
    @Stef, "common knowledge" doesn't mean that everyone knows it, just that it is widely available and generally agreed upon. If you can capture it in a textbook then is is almost always common knowledge. – Buffy Feb 07 '22 at 15:30
  • @Buffy I suspect most of us have seen at least one textbook apparently written with an intended audience of 20 students. In that case I suspect the contents would only be common knowledge in that classroom. – origimbo Feb 07 '22 at 16:35
8

I see this from a slightly different perspective than many of the other responses (from my cursory glance at them). The three examples you chose each has a slightly different definition. That's important. Though it seems they are repeating the same thing, they actually aren't. Why is this important? Because the exact definition may be the determining factor in what the study is trying to resolve. As mentioned by another, they are not defining "sleep." The kick is, they COULD have defined sleep if they had a specific reason for the type/kind of "sleep" they were concentrating on for the purposes of the study.

Consider this: Write a paper regarding something in your field of work or study. Write it as you wish papers were written, skipping all the "common sense" terms and concepts. Now, give that paper to someone who may have a general idea about the topic, but no where near what you have. Or better, give it to someone who has NO idea. Get their reaction. Did they understand it? Did they want/need foundational information to help with comprehension?

davidinark
  • 111
  • 1
7

When papers "state the obvious" it often appears weird and unhelpful, as you noticed. Yet this awkwardness does not come from "stating the obvious" but rather from immediately thereafter assuming the non-obvious. Let's consider your first example, along with the sentences that follow it:

Obstructive sleep apnea-hypopnea (OSAH) is characterized by episodes of complete or partial upper airway obstruction during sleep. Conditions associated with worsening severity of events include supine body position and rapid eye movement (REM) sleep. The therapeutic response to continuous positive airway pressure (CPAP) is usually complete.

Central sleep apnea (CSA) syndrome is characterised by recurrent episodes of apnea in the absence of upper airway obstruction during sleep which can occur in association with alveolar hypoventilation (i.e., hypercapnic), or it can be normocapnic or hypocapnic.

Merely three sentences after defining sleep apnea, the paper assumes the reader already understands "alveolar hypoventilation". How many people don't know what apnea is, but do know what "alveolar hypoventilation" is?

Introductions should be funnel-shaped, starting out broad and then gradually narrowing. They should provide prerequisite knowledge that a typical reader might not know. But when introductions narrow sharply, they seem faintly ridiculous. And when introductions provide prerequisite knowledge that most high school graduates know, but fail to provide prerequisite knowledge that first-year graduate students might not know, they are rather unhelpful. It is this unfortunately common practice that makes you (and many other readers of academic literature) doubt the value of introductions.

Betterthan Kwora
  • 531
  • 4
  • 12
3

Pro forma introductions to a research topic that are plagiarized from other papers are highly tedious and, as you suggest, a real turn-off towards such papers' findings. I think that journal editors ought to advise such authors to use their own words, write them clearly and avoid overuse of the jargon of the field: this much at least would show that they thought it through rather than simply mealy-mouthed it.

But introductions per se are vital for research papers. The reader's attention must be focussed on the direction of the author's own research work and its purpose in the context of existing knowledge in this field.

When other researchers are overviewing a long list of papers, a well-written intro is far more useful to them than the balder 'bombshell' abstract. Perforce this involves going back to first principles of the science/field. But it is where the author goes from these that is important - and how they go from there. While many will just follow the 'orthodox' path (and not question its many unstated assumptions) beaten by previous publications and enlarge it little further, others will be more radical and/or more thorough - and hopefully find a more fruitful vista.

Trunk
  • 4,418
  • 9
  • 28
1

This is a very old problem. There is ancient Chinese wisdom about it.

"In explaining anything to anyone, one should start with that which is absolutely clear."

It's a bit hopeless, since another equally old Chinese wisdom says:

"It is not possible to have anyone completely understanding what it is you mean, without you telling him what it is you believe in."

Add to that the physical fact that all language is a lie by definition and every use of it religious and you will have read the useless basics of this answer.

The truth is, these basics are usually added by demand. Papers aren't just written, but also published and the publisher has objectives of his or her own. Also the particular discipline a paper relates to may demand it, like in medicine or law. It is actually quite seldom a writer himself who wishes to start a scientific paper with the basics.

Another reason to include basics, is to define the school of thought that the paper follows. Basics, even if commonly accepted, aren't always 100% certain. This makes mentioning the assumptions a paper is based upon unavoidable.

Then there is the fear of being misunderstood and the desperation of not knowing what level of understanding basics to expect of the readers.

Local custom or rule in the discipline may also account for the inclusion of basic information. Sometimes it's just the way the writer learned to write. Believe it or not, some people were actually taught to write as if they are explaining something to a child.

Last but not least, there is the fact that those who write papers, as much as they may be experts in their field, are seldom experts in writing.

As a subscriber to "Nature" I did a lot of interdisciplinary scientific reading over the years. Over time I have come to appreciate that where the point of a story doesn't always get me to understanding its extend, this seemingly ballast information sometimes reveals the true nature of the core message of a paper. It's doesn't just contain 'the' basics, but rather the writers basics, which I by occasion have found to be quite revealing. I prefer a paper with too much basic information over a paper in which I have to fill in the blanks any time.

In general, a well written scientific paper is like a roller coaster ride. You have to climb to the top first, which may not be the most exiting part, but the higher the climb, the wilder the ride.

0

Journals can be odd things. Often, they publish papers in very narrow areas that can still be of interest to their wider general readership. This is particularly true of clinical literature, but it comes up everywhere.

Other times, especially where a paper takes common knowledge in one field and presents it to a broad readership in an adjacent area, basic explanations are quite necessary. For example, while apparently apnea impacts circulation in a way that it might interest the general readership of Circulation Research (one of your examples comes from there) it would be a mistake to assume that the general readership knows what sleep apnea is.

I suggest that what you call "obvious" may not be obvious when you consider a wider readership than what you apparently assume to be the case, necessitating more basic material than what would be required by an expert in a narrow area.

Scott Seidman
  • 31,120
  • 4
  • 52
  • 121