Here’s a draft of the new section on Writing an Abstract, to be published in the fourth edition of my Academic Legal Writing book. There’s still plenty of time to improve it, so I’d love to get feedback. (By the way, the abstracts I give as examples are my own, but I’d prefer to use someone else’s abstracts, especially if they are very effective. So if you have any recommendations for very good abstracts, please pass them along.)
An abstract is a short summary — one to three paragraphs — of an article. Some journals include an abstract at the start of the article, or put all the abstracts from an issue on the issue’s table of contents, or put the abstracts on the journal’s Web site. These journals will either require you to write the abstract, or will offer to write it for you. Reject their offer, and write the abstract yourself: It’s your article, and you’ll know better how to summarize it effectively.
But even if the journal doesn’t publish an abstract, you should write one anyway. Services such as the Social Science Research Network (see p. 265) maintain e-mail distribution lists through which hundreds or thousands subscribers get abstracts of forthcoming articles. These distribution lists are invaluable tools for you to get readers for your work.
Whether in a law review or on a distribution list, the abstract is an advertisement for your article. True, you don’t want money from your “customers” (the audience) — you want their time and attention. But their attention is scarce, and lots of authors are competing for it. You want readers to “buy” your article in one of two ways:
- by reading the article (or at least the Introduction) right away, or
- by remembering it (even if just vaguely) for the future, so that when the underlying issue becomes important to them, they can find and read the article then.
And the audience for your advertisement is quite demanding. They’ve generally found the abstract just through a quick skim of an SSRN e-mail or a law review table of contents. (People who find the article through a citation or a Westlaw or Lexis search are probably more likely to skim the Introduction, which is immediately available to them, rather than starting with the abstract.) Readers of your abstract therefore aren’t at all sure the article will be of any value to them.
You need to quickly show them this value. You need to clearly and tersely tell the reader (1) what problem the article is trying to solve, and (2) what valuable original observations the article offers. Naturally, the abstract can’t go into much detail. But it has to at least give the reader a general idea of what the article contributes.
Here, for instance, is an adequate abstract, adequate because it quickly captures the essence of the value added by the article:
People often argue that symbolic expression — especially flag burning — isn’t really “speech” or “press,” and that the Court’s decisions protecting symbolic expression are thus illegitimate.
But it turns out that the original meaning of the First Amendment likely includes symbolic expression. Speech restrictions of the Framing era routinely treated symbolic expression the same as literal “speech” and “press.” Constitutional speech protections of that era did so as well, though the evidence on this is slimmer. And the drafting history of the phrase “the freedom of speech, or of the press,” coupled with the views of leading commentators from the early 1800s, suggests that the First Amendment’s text was understood as protecting “publishing,” a term that at the time covered communication of symbolic expression and not just printing. Though the Court has never relied on this evidence, even originalists ought to accept the Court’s bottom line conclusion that the First Amendment covers symbolic expression.
The first sentence does three things. First, it notes the general topic of the article — the First Amendment and symbolic expression generally. Second, the sentence identifies the specific focus of the article, which is whether the text of the First Amendment must be read as protecting only “speech” and “press” and not symbolic expression. Third, the sentence very quickly provides a concrete illustration (flag burning) for the abstraction (symbolic expression).
The second sentence explains the article’s claim: The original meaning of the First Amendment likely covers symbolic expression. Readers who stop reading there will at least remember something like “There’s an article that says that even originalists should approve of the Court’s flagburning decisions.”
That would be an oversimplification of the article’s claim, but that’s fine — any one-sentence summary that lingers in people’s minds will inevitably be an oversimplification. The important thing is that if the issue comes up for readers in the future, they might well search for the article, find it, read it, and use it. And, if the author is lucky, maybe some readers will be interested enough to actually read the article right away, or at least move from reading the abstract to reading the Introduction.
The next three sentences quickly summarize the main arguments that the article uses to support its claim. These arguments — here, historical assertions, though for another article they might be normative arguments or empirical findings — are part of the contribution that the article offers. Again, the summary is an oversimplification, and as a result may not be entirely clear to all readers. But it should at least give the reader a glimpse of the observations that the article makes.
Finally, the last sentence ties the argument to the caselaw: The sentence explains that this is an article that offers historical support for the Court’s precedents, rather than arguing against the Court’s precedents.
Many authors try to fit an abstract into one paragraph, and some journals seem to prefer that. I advise against this, unless the abstract is very short. Shorter paragraphs tend to be more readable, and longer paragraphs tend to be alienating to many readers. And the reader of the abstract will likely be the sort of reader who is especially unmotivated to read further. The more you can do to make the abstract appealing, within the space constraints you’re given, the better.
Likewise, I like including numbering, for instance in this abstract:
How should state and federal constitutional rights to keep and bear arms be turned into workable constitutional doctrine? I argue that unitary tests such as “strict scrutiny,” “intermediate scrutiny,” “undue burden,” and the like don’t make sense here, just as they don’t fully describe the rules applied to most other constitutional rights.Rather, courts should separately consider four different categories of justifications for restricting rights: (1) Scope justifications, which derive from constitutional text, original meaning, tradition, or background principles; (2) burden justifications, which rest on the claim that a particular law doesn’t impose a substantial burden on the right, and thus doesn’t unconstitutionally infringe it; (3) danger reduction justifications, which rest on the claim that some particular exercise of the right is so unusually dangerous that it might justify restricting the right; and (4) government as proprietor justifications, which rest on the government’s special role as property owner, employer, or subsidizer.
I suggest where the constitutional thresholds for determining the adequacy of these justifications might be set, and I use this framework to analyze a wide range of restrictions: “what” restrictions (such as bans on machine guns, so-called “assault weapons,” or unpersonalized handguns), “who” restrictions (such as bans on possession by felons, misdemeanants, noncitizens, or 18-to-20-year-olds), “where” restrictions (such as bans on carrying in public, in places that serve alcohol, or in parks, or bans on possessing in public housing projects), “how” restrictions (such as storage regulations), “when” restrictions (such as waiting periods), “who knows” regulations (such as licensing or registration requirements), and taxes and other expenses.
Though it’s unusual to number individual clauses in normal prose, here the numbering quickly shows the hurried reader how the sentence is structured, and what the four elements of the proposed framework are. It might have even been helpful to do something similar in the last paragraph. But on the other hand too much numbering might have annoyed readers — a bit of departure from standard prose style is fine, but too much would make the abstract look odd. And the quotation marks surrounding the key items in the last paragraph probably provide some internal delimiters that can serve as alternatives to numbering.
Acquisitions Committee says:
I’ve found that a strong abstract really helps me sell an article I like to the decision-makers. Consider expanding your definition of audience to include the prepublication audience of law review editors.
February 8, 2010, 6:26 pmOrin Kerr says:
I actually prefer a short one-paragraph summary for most abstracts over the slightly longer multi-paragraph version. (If it can be done — more true with some articles than others.)
This is implicit in what you’re saying, but I can’t stress enough the importance of not just saying that the article “discusses”, “investigates,” or “analyzes” an issue: The abstract needs to include the key argument, finding, claim, etc. that follows from the discussion, investigation, and analysis.
February 8, 2010, 6:41 pmEugene Volokh says:
Acquisitions Committee: Very interesting — I had assumed that law review editors would focus more on the Introduction, which is short enough to read with a bit of care, but long enough to offer a reasonable amount of detail (more than the abstract could). Did you find that the abstract really added a lot to it, or led you to read Introductions that you otherwise wouldn’t have?
Also, how many of the articles that you received already included abstracts? And, finally, were most of them set in all-italics (not uncommon for abstracts, I think), and did you find this made them harder to read?
February 8, 2010, 6:42 pmlgm says:
This is great advice, mostly. Here are some comments from a non-lawyer professor.
1. The abstract is distinct from the first paragraphs of the paper. The abstract describes for experts what the paper accomplishes, giving them a reason to read the paper. The beginning of the introduction gives the setting for the paper in its larger context, often for the benefit of people less expert. Full scale jargon is OK for the abstract, but not for the beginning of the introduction.
2. Your first paragraph could benefit from some “Strunk&White”ing, particularly the … or … or … or part. You are trying to set an example of good writing.
3. I particularly agree with your description of the inpatient reader. My time is limited. If I have trouble with an abstract, I will spend my time reading somebody else’s article.
February 8, 2010, 6:59 pmKent Scheidegger says:
The odd thing is that the legal profession is so far behind all the others in putting abstracts on its articles.
February 8, 2010, 7:07 pmEugene Volokh says:
lgm: Thanks for the suggestions! I should note that I think the multiple “or” locution (“a, or b, or c”) is generally fine; but I’ll look over it again and see if it strikes me as annoying or odd.
February 8, 2010, 7:09 pmAcquisitions Committee says:
I don’t have an initial gatekeeper role, so I don’t know how everything comes in. My sense is that abstracts are standard.
We usually circulate the abstracts of all the submissions we are serious about to the entire acquisitions committee. So if I am trying to get support for an article I like and have read in full, I have to plan my pitch around the fact that everyone has seen the abstract. If the abstract is strong, my job pitching your article has become that much easier. The abstract should also give me some good talking points. If the abstract is weak or the gatekeeper had to write a summary for you, well, you get the picture.
Maybe we are the only law review doing things this way. I doubt it.
Italics don’t really matter, since we can easily change the fonts when we copy the abstract into a circular email on today’s hot submissions.
February 8, 2010, 7:39 pmpublic_defender says:
As a practitioner who sometimes uses law review articles, I don’t care whether it’s an introduction or an abstract, but I need something in the Westlaw/Lexis version of the article that tells me what the author saying. Usually, when I’m looking at law review articles, it’s be cause I’m researching first principles, and I need to know if this specific article will address whatever issue is in my case. All too often, the introduction or abstract just doesn’t help. I’m willing to wade through imperfect prose and bad organization to find helpful information, but I’d rather not. (My guess is that most judges and law clerks approach lawyers’ briefs the same way.)
Maybe the abstract/introduction distinction is the best way to reach multiple audiences. If some readers want to slowly wade in, they can meander through the introduction. If they want a well-written practitioner-style summary, they can start with the abstract.
February 8, 2010, 7:56 pmBrianMac says:
As my old boss used to say, an abstract should cover four things:
1) What did you do
2) Why did you do it
3) What were the results, and
4) What do they mean
I’ve found that works pretty well for the “sciences,” but I’m not so sure how applicable it is for law review articles.
February 8, 2010, 8:42 pmrarb says:
If you do end up recommending that abstracts be submitted with the article (assuming a sufficient number of law review editors find them helpful, either in screening the selections or in pitching articles within the editorial board), then I’d suggest that you include some explanation about how to include the abstract when you submit the piece. Should it be a freestanding document separate from the article and the cover letter? If it’s part of the article and comes after the title and just before the introduction, I’d worry that people would read it and make a judgment based solely on the abstract, before reading the introduction.
Isn’t the introduction, not the abstract, the place where you really attempt to draw a new reader into reading the piece?
February 8, 2010, 8:43 pmAcquisitions Committee says:
I used to worry that people judged books by their covers. Then I put the world’s best cover on my book. Now I hope they do.
If I am reading the introduction, I’ve already decided to read the article, though I suppose I could stop if the introduction convinces me I’ve made a mistake. Don’t most people actually read law review articles through Wexis? Is there even an introduction field?
February 8, 2010, 9:56 pmbillb says:
Do law reviews really set their abstracts in all italics? Really? OUCH! That’s a great way to get readers to completely skip reading an abstract.
February 8, 2010, 11:14 pmEric Rasmusen says:
Even in economics, an abstract is very much like legal pleadings. You’re telling your reader what you claim you will show, without giving him the evidence yet. The most important thing for students (and professors) to keep in mind is that, as with pleadings, it isn’t correct just to tell the reader the subject you’re going to talk about. No “In this article I discuss how concentration affects profits in Bulgaria,”; rather, “In this article I show that concentration and profits are positively correlated in Bulgaria.”
You might like my
February 8, 2010, 11:30 pm“Aphorisms on Writing, Speaking, and Listening”. Tho designed for economists, it has a lot of good tips for academic law.
public_defender says:
Even better:
“
February 9, 2010, 5:55 amIn this article I show thatConcentration and profits are positively correlated in Bulgaria.”Writing an Abstract for a Law Review Article « Daniel Joseph Smith says:
[...] Writing an Abstract for a Law Review Article [...]
February 9, 2010, 8:26 amEric Rasmusen says:
Yes, your version is much improved on mine, Public_Defender! Strunk and White strike again.
February 9, 2010, 9:23 am