Ejusdem Generis: What Is It Good For?


Vol. 100 No. 2 (2016) | A Judge in Public Life | Download PDF Version of Article
Cartoon of a lawyer with his head in his hand

In my view, this canon of interpretation is so fraught with uncertainty of various kinds that courts should give it little weight. Better yet, drafters should not unwittingly bring it into play. If, for instance, the landlord wanted to prohibit all pets, why even mention dogs and cats? The drafter might say pets of any kind, including dogs and cats, but adding specifics always creates some risk of the canon’s rearing its head.

Still, drafters so routinely mix specifics with a more general term that courts often have to divine whether some other specific is included. A Westlaw search for “ejusdem generis” in April 2016 produced 2,848 federal citations and 6,154 state citations, or about 9,000 total. In 2015 alone, the canon was cited in 68 federal cases and 109 state cases. Courts consider it, then, in scores of cases every year, and its use will probably only increase after the U.S. Supreme Court’s decision in Yates v. United States. 1 I’ll return to Yates at the end of this article.

Do drafters know the canon and rely on it?

In their noteworthy book Reading Law, the late Justice Antonin Scalia and Bryan Garner answer that question like this:

The truly knowledgeable interpreter (and drafter) knows the ejusdem generis canon; it has become part of the accepted terminology of legal documents. Any lawyer or legislative drafter who writes two or more specifics followed by a general residual term without the intention that the residual term be limited may be guilty of malpractice.2

What a “truly knowledgeable” drafter knows is one thing. What typical drafters know and trust is another, and we now have evidence from two studies about actual real-world practice. In the first, a large study of 137 congressional staffers, 65 percent could not name the canon, although 71 percent use the underlying assumption.3 A second, similar study of 128 federal-agency rule drafters produced somewhat lower numbers: 47 percent could not name it, but 60 percent use it in drafting.4 At best, then, the canon is an “accurate judicial approximation[] of the way that [most?] drafters put language together.”5

But the numbers should still give us pause. Roughly a third of professional drafters — at the highest level — apparently do not rely on ejusdem generis in their own work. Imagine what those numbers would be for a cross section of practicing lawyers, who are generally not known for their drafting acumen.6 Perhaps you could argue that regardless of what drafters profess to rely on, the principle behind ejusdem generis is somehow in their linguistic bones. I wonder, though, whether that supposition can justify a lot of confidence that drafters and courts — the writers and readers — are in tune on the canon’s value.

The view of commentators

Reading Law frankly acknowledges that “[c]ommentators sometimes dispute whether the ejusdem generis canon is beneficial.”7

On the positive, supportive side is Sutherland’s Statutes and Statutory Construction: “Ejusdem generis expresses a meaningful insight about language usage that can be a relevant aid, if not a simple and certain exponent, to resolve that issue [of how broadly to apply the general term].”8 The treatise does generally endorse the canon, with “its specific, doctrinal limitations.”9

On the opposite side, though, are many commentators with a shared skepticism about the canon’s worth:

• The redoubtable Max Radin compared it with the canon expressio unius est exclusio alterius, which he castigated as being “in direct contradiction to the habits of speech of most persons” and as “illustrat[ing] one of the most fatuously simple of logical fallacies.”10 He then said that ejusdem generis “has a little greater foundation in logic and in ordinary habits of speech.”11 That is, a little more than none at all. He was damning with faint praise. And Radin went on to argue, as others have, that the canon “is of limited importance in discovering the determinates covered by the statute.”12

• Reed Dickerson, the father of modern-day legal drafting, questioned whether it is “lexicographically accurate” and called it a “convenient fiction[] for deciding specific controversies” when “the search for meaning and reliance on the methods of reasoned elaboration fail.”13

• Robert Benson, an influential early proponent of plain legal language, denounced its vagueness and manipulability: “The canon doesn’t help you identify the class. You’re limited only by your imagination, and by what you can persuade others is reasonable.”14

• Jack Landau, an Oregon justice and professor, also made that point when reviewing opinions in his state: “[S]ome individual canons are themselves so malleable that they are worthless. A good example is the canon ejusdem generis . . . . [The supreme court’s] continued reliance on a number of fictions . . . such as ejusdem generis . . . undercuts the persuasiveness of its opinions and the credibility of its decisions.”15 The common criticism about squishiness is borne out by a large sampling of cases involving the term other minerals in legal documents. The author of the study concluded that in using ejusdem generis, courts had to “arbitrarily emphasize one characteristic over another”; “the doctrine fosters uncertainty and instability in land titles”; and it “fails to meet the goal of fairness.”16

• Likewise the conclusion from a review of “a harvest of litigation” over the term other casualty in Internal Revenue Code § 165(c) (3) [26 U.S.C. § 165(c)(3)]: “[T] he courts have had great difficulty applying the rule of ejusdem generis in difficult factual situations. As a result, the developed case law is in such a state of confusion that it caused the Tax Court to remark: ‘There are numerous cases involving casualty losses, some of them difficult to reconcile with others, either in result, theory or language.’”17 The author added that the statute’s legislative history “seems to negate the applicability of . . . ejusdem generis.”18

The Many Qualifiers

The canon is beset with conditions, qualifications, and contradictions — and ultimately with considerable uncertainty and malleability. The first six items below appear in Reading Law (five are called “caveats”), but the list is longer than that.

One: “[E]jusdem generis generally requires at least two words to establish a genus . . . .”19 A sensible limitation, as the authors clearly explain. They cite the United States Supreme Court case Ali v. Federal Bureau of Prisons, which did reject an ejusdem generis claim because the phrase in question seemed to involve only “one specific and one general category.”20 Yet the Court had suggested otherwise in an earlier case involving the phrase the antitrust law and . . . all other law: the canon was said to potentially apply “when a general term follows a specific one.”21

Two (a big one): You have to decide “how broadly or narrowly to define the class.”22 The authors acknowledge, again, that this indeterminacy is what makes the canon unpopular with many commentators. They also concede that their advice “must be a generalization: Consider the listed elements, as well as the broad term at the end, and ask what category would come into the reasonable person’s mind.”23 The authors say, “Often the evident purpose of the provision makes the choice clear.”24 But just as often, or more often, the choice won’t be clear.25

Three: “[S]ometimes the specifics do not fit any kind of definable category . . . . With this type of wording, the canon does not apply.”26

Four: “[W]hen the specifics exhaust the class and there is nothing left besides what has been enumerated, the follow-on general term must be read literally.”27

Five: “[T]he general word will not be treated as applying to persons or things of a higher quality, dignity, or worth than those specifically listed.”28

Six (another big one): “When the context argues . . . strongly against limiting the general provision, the canon will not be dispositive.”29 The example given is a sign at the entrance to a butcher shop: “No dogs, cats, or other animals allowed.” The authors say that “no one would think that only domestic pets were excluded, and that farm animals or wild animals were welcome.”30 The trouble lies in the word context. The authors can’t be talking about the purely linguistic or textual context, as they are at other places in the book. They seem to be talking about the situational context, the meaning in light of the circumstances. The same sign would no doubt be read differently at a livestock auction. So what’s the difference between asking readers to consider the situational context and inviting them to use their intuition and common sense about the sign-writer’s intent? As Judge Richard Posner points out, with this butcher-shop example the authors seem to do “an aboutface. . . . [They are] correct not by virtue of anything textual but because of the principle that meaning includes what ‘would come into the reasonable person’s mind’ or what we know an author has ‘in mind’ in writing something.”31

Below are some other possible qualifiers that may or may not be widely accepted. The point is that they are out there and available for courts and advocates to use.

Seven: The canon doesn’t apply so as to limit the general term “if the [specific] terms supplementing the general term are themselves general.”32 That seems like potentially fertile ground for argument.

Eight: According to one authority, “[E] jusdem generis may apply only to help clarify ambiguous legislative intent or statutory meaning.”33 According to another, “Judges who refuse to apply the . . . canon[] in the absence of ambiguity are simply wrong.”34 That’s obviously a major disagreement, and a recent Fifth Circuit case concedes that precedent from the Supreme Court and from the Fifth Circuit itself “is not entirely clear on this point.”35

Nine: “By adding catch-alls, legislatures do not intend to narrow lists, but rather include a catch-all because they cannot identify, in advance, everything they would want included in the list. Thus, legislatures would likely prefer that . . . general terms be interpreted broadly, not narrowly, to cover all possible, but similar, contingencies.”36 Arguably, courts should keep this in mind when faced with a choice between classes.

Ten: There’s at least some suggestion that the canon may have less sway if the general term is modified by any or every or all. 37 And what about an even stronger modifier, as in dogs, cats, or other pets of any kind? At that point, doesn’t ejusdem generis give way?

Another twist: reversing the pattern

Now for a fundamental question: Does ejusdem generis still apply if the general term comes first? Reading Law insists that the answer is no: The canon applies only if the general term follows the specifics.38

The same question was addressed in an article in volume 11 of The Scribes Journal of Legal Writing. 39 The article noted that, beginning with the fourth edition, Sutherland Statutes and Statutory Construction has stated that the doctrine applies to both patterns: “Where the opposite sequence is found, i.e., specific words following general ones, the doctrine is equally applicable, and restricts application of the general term to things that are similar to those enumerated.”40

The article then took the following example of general-to-specific language from a case: earth movement, including but not limited to earthquake, landslide, mudflow, earth sinking, earth rising, or shifting. 41 And the article asked:

Will the courts still apply ejusdem generis in cases like this? I found more cases saying yes than no. . . . [A]pplying ejusdem generis when the general term comes first appears to be the majority view.42

That is, of the cases that have considered the more uncommon general–specific pattern, most appear to have applied ejusdem generis. Regardless of what the actual percentages are, a number of courts have limited the general term because of the specifics that follow it.43 And why not? The logic is the same either way, regardless of the pattern.

In rebuttal, Reading Law contends that the general–specific pattern is different because it “can be thought to perform the belt-and-suspenders function”:

Following the general term with specifics can serve the function of making doubly sure that the broad (and intended-to-be-broad) general term is taken to include the specifics. Some formulations suggest or even specifically provide this belt-and-suspenders function by introducing the specifics with a term such as including or even including without limitation . . . . Enumerating the specifics before the general, on the other hand, cannot reasonably be interpreted as having such a function.44

Well, why not? Michigan’s recreational user’s statute refers to a person who is on another’s land “for the purpose of fishing, hunting, trapping, camping, hiking, sightseeing, motorcycling, snowmobiling, or any other outdoor recreational use or trail use.”45 That looks like belt-andsuspenders to me. It’s not readily apparent how the specific examples limit the general terms.

Reading Law offers a second argument for accepting the traditional pattern (only):

When the genus comes first (“all buildings, assembly houses, courthouses, jails, police stations, and government offices”) it is a stranger that arrives, so to speak, without an introduction saying it is limited; one is invited to take it at its broadest face value.46

I’m not sure I get the point here. Are the authors talking about a simple miscue? (Not a good thing, but not fatal.) How likely is it that a court would perceive a class of buildings yet reject ejusdem generis for lack of an introducer? At any rate, the example seems contrived. Surely, if a drafter wanted to limit the general term and decided to put it first, he or she would introduce the specifics with buildings similar to or all buildings like the following. (Using buildings such as, even without a comma, is possible but risks being read as introducing examples rather than a class.47)

At bottom, ejusdem generis is a construct from a simple combination of a general term and specific terms. Probably the general term comes last more naturally and thus more often, and perhaps that explains how the canon developed in its traditional formulation. But tradition is one thing; the underlying logic or principle is another. Which should prevail? Why should it make a difference whether a drafter writes dogs, cats, and other pets or pets, including dogs and cats?

A last twist: the canon as redundant

There’s another, deeper reason that may account for Reading Law’s insistence on the traditional formulation: canonical redundancy. If ejusdem generis works both ways, then it adds little or nothing to the associated-words canon (noscitur a sociis): “Associated words bear on one another’s meaning.”48 In fact, ejusdem is described as a “subset” of noscitur49 or “an instance of [the] more general maxim.”50 But are we supposed to preserve two canons that are based on the same essential linguistic and logical principle — just for the sake of preserving them? As rationales go, that one seems pretty arbitrary.

We might ask whether it matters to have two canons — one broader (noscitur) and one narrower (ejusdem) — for the same principle. The Scribes Journal article cited one case dealing with the identical earth movement language to which other courts had applied ejusdem generis. 51 This one court did not, because the general term came first.52 But the associated words were all natural occurrences, so applying the noscitur canon might well have narrowed the general term and thus changed the result. That’s the trouble with having both canons: It makes for inefficiency, confusion, and conflict.

Scalia and Garner advise that “the most common effect of the [noscitur] canon is . . . to limit a general term to a subset of all the things or actions that it covers . . . .”53 But that’s exactly what ejusdem generis — applied both ways — does. Ejusdem, it seems, is expendable, at least in theory.

There’s no easy way out of this interpretive mess. When listed specifics are not accompanied by a general term, noscitur a sociis may apply. When they are accompanied by a general term, then either canon may apply — for what they’re worth. As for ejusdem generis, it is heavily qualified and has led to uneven results in groups of similar cases. At best, even when the class is more or less clear, ejusdem should only create a presumption that is “perhaps not too difficult to overcome.”54

The drafting quandary

In all of this, there’s a deeper question for drafters: whether to list specific items to begin with.55 Doing so, even carefully, always presents some danger that a court will limit the general term when no limitation was intended or will limit it to a class that wasn’t intended. The better choice, as I suggested at the outset, may well be to drop the specifics and sharpen the general term as needed. In the buildings example on this page (middle column), for instance, if the drafter means all buildings or buildings of any kind, perhaps the drafter should say just that. Or if the drafter has in mind a certain kind of building, then just describe it as accurately as possible.

Or take the case and insurance language that the article in the Scribes Journal used as its jumping-off point: earth movement, including but not limited to earthquake, landslide, mudflow, earth sinking, earth rising, or shifting. 56 The plaintiff’s buildings were damaged by the underground construction of a rapid-transit system. Applying ejusdem generis, the court agreed with several other courts and limited the exclusion to “natural” phenomena or occurrences like the specific examples — those that “would cause such widespread damage to so large a number of policyholders as would otherwise destroy the probabilities upon which insurance companies operate and for which a policy exclusion would be expected.”57

Note: The court did not treat including but not limited to as the belt-and-suspenders language of specific examples only.58 The specifics betrayed the drafter if he or she intended a broad reading of earth movement. So again, the drafter might have been better off to say earth movement of any kind if indeed that was the intended meaning. If not, what exactly was it?

Finally, let’s put the drafting quandary in the context of the most common rationale for the ejusdem generis canon:

[W]hen the tagalong general term is given its broadest application, it renders the prior enumeration superfluous. If the testator really wished the devisee to receive all his property [in the example “my furniture, clothes, cooking utensils, housewares, motor vehicles, and all other property”], he could simply have said “all my property”; why set forth a detailed enumeration and then render it irrelevant by the concluding phrase all other property?59

This rationale is open to question, suggesting as it does that the drafter consciously used the imprecise term — all other property instead of all other personal property — because he or she knew all about ejusdem generis and was confident that a reader would apply it. But it seems just as likely that the interpretive problem — whether the drafter meant to include real estate — arose from loose drafting. One more time: If the drafter had used all my personal property, then arguably the specifics would have been unnecessary to begin with.

The Kaluza Case

Let’s examine two recent high-profile cases — one in this section and one in the next — with the preceding discussion in mind. I’ve already cited both of them in passing (endnotes 25, 35, and 37).

The first is the Fifth Circuit case of United States v. Kaluza. 60 The defendants were the highest-ranking BP employees — the well-site leaders — on the Deepwater Horizon oil-drilling rig that exploded in the Gulf of Mexico in April 2010, killing 11 men. The court described the Deepwater Horizon as a “mobile offshore . . . drilling vessel.”61 The drill crew that the defendants led was separate from the marine crew; once the rig arrived at the site, the marine crew was responsible only for keeping it in position over the wellhead.62

The defendants were indicted for seaman’s manslaughter under 18 U.S.C. § 1115. But the district court found that they did not fall under the statute’s description of persons who could be prosecuted for negligence: “Every captain, engineer, pilot, or other person employed on any steamboat or vessel . . . .”

The appeal centered on ejusdem generis. The parties first argued about the need to find ambiguity before applying canons of construction. (See page 51, item 8.) But the court sidestepped that question by finding ambiguity. On the one hand, [e]very . . . other person employed on any . . . vessel is all-encompassing. (The court felt the need to consult dictionary definitions to so conclude.) On the other hand, a broad reading would render the specific terms — captain, engineer, pilot — superfluous (sound familiar?). “This ambiguity necessitates the use [of] the canon of construction of ejusdem generis.”63

Consider the implications here: Whenever the general term is modified by an expansive word like any, every, or all (see page 51, item 10), the phrase is ambiguous?

• dogs, cats, and all other pets [ambiguous?]

• dogs, cats, and other pets [not ambiguous?]

So in every contract that uses an expansive variant, the phrase must be construed against the drafter? And in every criminal statute that uses one, the rule of lenity may come into play?

No, the uncertainty that inheres in the expansive variant is not ambiguity in the typical or strict sense. The reader is not choosing between two quite different meanings of a word or phrase;64 the reader is deciding whether to take literally the expansive term — any, every, all [whatever] — and treat the specific terms as nothing more than prominent examples (not members of some class). Ironically, the court in Kaluza found ambiguity before proceeding to invoke ejusdem generis when it might just as well have asked whether ejusdem applied to begin with. That’s always a question with the expansive variant: whether to effectively ignore the specific items.

At any rate, having decided that ejusdem generis applied, the court next wrestled with how best to define the common attribute of captain, engineer, [and] pilot. 65 (See page 51, item 2.) ) The court rejected three descriptions suggested by the government, including the broader first and second ones below, in favor of the narrower last one:

• Someone with a substantial degree of responsibility for the vessel’s safety.

• Someone responsible for the operation, equipment, or navigation of the vessel.

• Someone responsible for the marine operations, maintenance, or navigation of the vessel.66

The difference between the last two choices is in the word marine, a qualification that the court partly derived from dictionary definitions of the specific terms. The court said that the narrower description involved responsibility for the vessel’s use as a “means of transportation on water.”67 The defendants were in charge of drilling; they had no transportation-related duties. So even though the statute applied, it did not apply to the defendants, because they fell outside the narrowest of four possible classes.

After its extended discussion of ejusdem generis, the court turned to legislative history — and determined that it supported the conclusion reached under the canon. But the court had earlier said that “legislative history can only be a guide after the application of canons of construction.”68 I’ve argued elsewhere that canons as a group have no superior claim to legitimacy or reliability.69 Used sensibly, legislative history may sometimes trump or counter a canon, sometimes reinforce it, and sometimes suggest how broadly or narrowly to apply it. Legislative history should not automatically play a secondary role.

At the end, in the opinion’s next-to-last paragraph, the court made these telling observations:

Counterarguments in favor of interpreting § 1115 to cover Defendants have purchase. Yet we are left with textual indeterminacy, as well as the incongruity of applying a statute originally developed to prevent steamboat explosions and collisions on inland waters to offshore oil and gas operations — all approaching a bridge too far. . . . At some point, and we think it here, the doctrine of lenity takes hold and dismissing this part of the indictment was not error.70

The court seemed to concede, despite its earlier analysis, that it was left with some measure of “textual indeterminacy.” But that’s commonly, or usually, the case with ejusdem generis. So why don’t courts treat it more skeptically to begin with? Yes, you get the standard qualifications about its being “only an instrumentality” and “not . . . conclusive.”71 But even that overstates its validity.

The quotation above also suggests that the court was never comfortable with the “incongruity” of applying the statute to a drilling rig. That’s an expression of policy, of the court’s intuitive sense that this criminal statute should not apply to an atypical situation that the drafters in 1838 could not have imagined. The problem is that the statute used the modifying phrase on any steamboat or vessel, and the drilling rig was unquestionably a vessel under the Dictionary Act.72 So the court was stuck, textually, with ejusdem generis.

I have a hunch, though, that the court’s sense of whether the statute should apply came before it went about parsing the different possible classes. Rather than putting so much stock in dictionary definitions of captain, engineer, and pilot, the court could have conceded the uncertainty and chosen the narrowest possible class in light of the legislative history. The court did say that the legislative history “supports a narrow reading of the statute.”73 Or did the history drive the reading? This may seem like a niggling difference, but it gives less credence to a shaky canon of construction.

And remember — if every . . . other person had been read literally, the court could have concluded that the three specific terms were no more than examples. That same issue came up again in our second case.

The Yates case

Just two weeks before Kaluza was handed down, the Supreme Court decided Yates v. United States. 74 Yates had caught undersized fish, red grouper, in violation of federal regulations and had ordered a crew member to throw the catch overboard. He was charged under 18 U.S.C. § 1519 — originally part of the Sarbanes–Oxley Act — for destroying or concealing “any record, document, or tangible object with the intent to impede” a federal investigation.

For our purposes, the disagreement among the justices can be boiled down to this: whether to (1) give any and tangible object their ordinary, or dictionary, definitions or (2) restrict the words to a narrower class of items like a record and a document. Five justices, including one who wrote separately, voted in favor of (2): The statute covered only the kinds of objects that “one can use to record or preserve information.”75 Justice Kagan wrote the dissent. She argued that even if (contrary to the ordinary meaning of any and tangible object) you were trying to identify the common traits of a class, records and documents are “things that provide information, and thus potentially serve as evidence.”76 That’s what the fish would have done.

The opinions argued a number of other points as well:

• The importance and import of the context in assessing whether the statutory terms were ambiguous.

• The meaning of tangible object in Federal Rule of Criminal Procedure 16.

• The heading to § 1519 and to the section of the Sarbanes–Oxley Act that included the provision.

• The placement of § 1519 immediately after other U.S. Code sections that have a limited reach.

• The possible redundancy of § 1519, read broadly, with another section of the Sarbanes–Oxley Act.

• The noscitur a sociis canon.

• The use of record, document or thing in the Model Penal Code.

• The rule of lenity and whether the statute was ambiguous.

• The possible significance of the 20-year maximum sentence.

Five observations about Yates, the last four of which cement points that we’ve already covered.

First, the ejusdem generis canon will often conflict with the ordinary-meaning canon. Take our original example: The ordinary meaning of pet is “a domesticated animal kept for pleasure rather than utility.”77 Even a snake can fit that definition. So you can add this conflict to the potential conflict between other textual canons: last antecedent vs. series qualifier, for instance, and scope of subparts vs. series qualifier.78

Second, the conflict between ejusdem generis and ordinary meaning is more pronounced when an expansive term like any or all is included. In fact, the dissenting opinion in Yates said, “This Court has time and again recognized that ‘any’ has ‘an expansive meaning,’ bringing within a statute’s reach all types of the item (here, ‘tangible object’) to which the law refers.”79

Third, you can just about define the class to your liking. At the least, there will always be different possibilities that are plausible.

Fourth, ejusdem generis doesn’t (or shouldn’t) add anything to noscitur a sociis. The plurality opinion described them as “related.”80 The dissent even referred to the plurality’s “noscitur/ejusdem argument,” as well as both canons’ requirement of “identifying a common trait.”81

Fifth, if the congressional drafters had wanted to cover any tangible object, why didn’t they say just that — or otherwise just try to accurately describe the class?

Any record? Or if document adds something, any record or document? Or were they concerned about business documents only? What’s needed is more careful thought and less imitation of old patterns (assuming that the drafters are not constrained by the need for consistency with a related or parallel provision).

Now, just as in Kaluza, one paragraph seems to reflect the plurality’s intuitive sense of the right result in light of the statute’s purpose:

Section 1519 was enacted as part of the Sarbanes–Oxley Act of 2002, 116 Stat. 745, legislation designed to protect investors and restore trust in financial markets following the collapse of Enron Corporation. A fish is no doubt an object that is tangible; fish can be seen, caught, and handled, and a catch, as this case illustrates, is vulnerable to destruction. But it would cut § 1519 loose from its financial-fraud mooring to hold that it encompasses any and all objects, whatever their size or significance, destroyed with obstructive intent. Mindful that in Sarbanes– Oxley, Congress trained its attention on corporate and accounting deception and cover-ups, we conclude that a matching construction of § 1519 is in order: A tangible object captured by § 1519, we hold, must be one used to record or preserve information.82

After this, it seems fair to ask how much was gained by engaging in an extended battle of competing canons.83

But that’s the textualist exercise. It purports to unearth a law’s fixed, objective meaning through the application of an elaborate and often contradictory set of canons. On the other side, of course, are those who believe that judges are doing more when faced with vague language or with novel, anticipated facts, or both. Descriptions vary: Judges in these circumstances are said to “impose an interpretation,”84 or “derive meaning,”85 or even “give [a text] meaning.”86 Surely the judges in Kaluza and the justices in Yates were at least elaborating the meaning of those statutes to arrive at a decision.87

What it all means

Here’s a brief summary.

First, ejusdem generis is highly malleable, not to say manipulable — just as textualism itself is.88 The canon has generated significant criticism and produced variable results in similar cases.

Second, the canon’s logic works both ways, regardless of whether the general term comes first or last. So with either a specific–general pattern or a general– specific pattern, the canon essentially duplicates noscitur a sociis.

Third, drafters must take great care:

• They should think twice about listing specifics if they don’t want to run the risk of limiting the general term. They might rather take pains to describe the class accurately and let that general term stand alone.

• If they nevertheless choose to list specifics intended to be examples only, drafters must make that as clear as possible: X, Y, Z, and other buildings of any kind; buildings of any kind, including X, Y, and Z. Even then, there’s no guarantee that a court won’t apply ejusdem generis, as courts did in the earth movement cases.

• If they wish to risk the vicissitudes of ejusdem and list specifics intended to be members of a class, drafters must be clear about that as well: X, Y, Z, and other similar buildings; buildings similar to X, Y, and Z. (Or perhaps use of the same kind.)

• Drafters should be wary of terms like any, every, and all if they do intend to create a class.

Finally, courts should try to ground their decisions in something more solid than ejusdem generis — or at least cite other reasons for deciding as they do.

Joseph Kimble is professor emeritus at Western Michigan University–Cooley Law School, where he taught legal writing for more than 30 years. He is the author of Writing for Dollars, Writing to Please: The Case for Plain Language in Business, Government, and Law and Lifting the Fog of Legalese: Essays on Plain Language; senior editor of The Scribes Journal of Legal Writing; and editor of the “Plain Language” column in the Michigan Bar Journal. He served as president of the international organization Clarity and was a founding director of the Center for Plain Language. Since 1999, he has been a drafting consultant on all federal court rules.

1 135 S. Ct. 1074 (2015).
2 Antonin Scalia & Bryan A. Garner, Reading Law:
The Interpretation of Legal Texts 212 (2012).
3 Abbe R. Gluck & Lisa Schultz Bressman, Statutory
Interpretation from the Inside—An Empirical Study
of Congressional Drafting, Delegation, and the
Canons: Part 1, 65 Stan. L. Rev. 901, 932–33
4 Christopher J. Walker, Inside Agency Statutory
Interpretation, 67 Stan. L. Rev. 999, 1019, 1020
figs.1, 2 (2015).
5 Gluck & Bressman, supra note 3, at 952.
6 See Joseph Kimble, You Think Lawyers Are Good
Drafters?, 17 Green Bag 2d 41 (Autumn 2014).
7 Scalia & Garner, supra note 2, at 211.
8 Norman J. Singer & Shambie Singer,
2A Statutes and Statutory Construction
§ 47:18, at 393 (7th ed., rev. 2014).
9 Id. at 392; cf. Joel R. Cornwell, Smoking Canons:
A Guide to Some Favorite Rules of Construction,
CBA Record, May 1996, at 43, 45 (observing
that “[t]he canon appears beautiful in its simplicity, a gem of common sense,” but cautioning
about its “elasticity” and the need for an
advocate to “calibrate a genus” that produces the
desired result (emphasis added)).
10 Max Radin, Statutory Interpretation, 43 Harv.
L. Rev. 863, 873, 874 (1930).
11 Id. at 874.
12 Id.
13 Reed Dickerson, The Interpretation and
Application of Statutes 234, 249 (1975).
14 Robert Benson, The Interpretation Game: How
Judges and Lawyers Make the Law 40 (2008).
15 Jack L. Landau, Some Observations About Statutory
Construction in Oregon, 32 Willamette L. Rev. 1,
28, 67 (1996).
16 Brant M. Laue, Interpretation of “Other Minerals”
in a Grant or Reservation of a Mineral Interest, 71
Cornell L. Rev. 618, 622, 623, 624 (1986); see
also Daniel B. Kostrub & Roger S. Christenson II,
Canons of Construction for the Interpretation of
Mineral Conveyances, Severances, Exceptions, and
Reservations in Producing States, 88 N.D. L. Rev.
649, 661 (2012) (“The drawbacks to applying ejusdem generis [to and all other minerals] are readily
apparent. . . . [C] ourts lack an objective standard
for guidance in choosing the qualities that the
named substances share, and which of those will be
determinative to define the class . . . .”).
17 Stephen L. Golding, The Scope of “Other Casualty”:
A Rejection of the Rule of Ejusdem Generis,
51 Taxes 525, 525, 527 (1973) (quoting Harry
Heyn v. Comm’r, 46 T.C. 302, 309 (1966)). For
recent commentary on the same provision, see
Mark Lee Levine & Libbi Levine Segev, Real
Estate Transactions: Tax Planning and Consequences § 187, at 288 (2015 ed.) (“It appears that
[ejusdem generis], as a canon of construction, many
times operates to ‘answer’ a question superficially,
and obscure the function, purpose and reason of
the statute.”).
18 Golding, supra note 17, at 527.
19 Scalia & Garner, supra note 2, at 206.
20 552 U.S. 214, 225 (2008) (per Thomas, J.).
21 Norfolk & Western Ry. Co. v. Am. Train Dispatchers Ass’n, 499 U.S. 117, 127, 129 (1991) (per
Kennedy, J.).
22 Scalia & Garner, supra note 2, at 207.
23 Id. at 208.
24 Id.
25 See, e.g., Yates v. United States, 135 S. Ct. 1074
(2015) (in which the Court split 5–4 on whether any
record, document, or tangible object included fish).
26 Scalia & Garner, supra note 2, at 209 (citing
a case in which the general term all manner
of merchandise was not limited by a preceding
enumeration of fruit, fodder, farm produce, insecticides, pumps, nails, tools, and wagons).
27 Id. (citing a case in which the final term was not
limited in the phrase used in the preparation . . .
of soft drinks or other beverages, or for any other
28 Id. at 210 (citing a case in which a duty imposed
on copper, brass, pewter, tin, and “all other metals
not enumerated” did not apply to gold and silver).
29 Id. at 212.
30 Id.
31 Richard A. Posner, Reflections on Judging
211 (2013) (citations to Reading Law omitted).
32 Singer & Singer, supra note 8, § 47:20, at 395–96
(citing five cases, including one involving the
phrase [by] any false token or writing, or by any other
false pretense).
33 Id. § 47:18, at 387–88 (citing 14 cases).
34 Linda D. Jellum, Mastering Statutory Interpretation 138 (2d ed. 2013).
35 United States v. Kaluza, 780 F.3d 647, 658 n.34
(5th Cir. 2015).
36 Jellum, supra note 34, at 138.
37 See, e.g., Harrison v. PPG Indus., Inc., 446 U.S.
578, 589 (1980) (rejecting ejusdem generis mainly
because “the phrase ‘any other final action,’ in the
absence of legislative history to the contrary, must
be construed to mean exactly what it says, namely,
any other final action”); see also Yates v. United
States, 135 S. Ct. 1074, 1092 (2015) (Kagan, J.,
dissenting) (citing several other Supreme Court
cases giving any “an expansive meaning”).
38 Scalia & Garner, supra note 2, at 203–05.
39 Gregory R. Englert, The Other Side of Ejusdem
Generis, 11 Scribes J. Legal Writing 51 (2007).
40 2A Sutherland Statutes and Statutory
Construction § 47:17 (C. Dallas Sands ed., 4th
ed. 1973) (the same sentence appears in all later
41 Holy Angels Acad. v. Hartford Ins. Grp., 487
N.Y.S.2d 1005, 1006 (Sup. Ct. 1985).
42 Englert, supra note 39, at 53–54.
43 Id. at 54 nn.5–6 (citing nine cases); see also Jellum,
supra note 34, at 138 (“In [a case under discussion],
the general term preceded the specific listed items.
For ejusdem generis purposes, placement is irrelevant.”); Singer & Singer, supra note 8, § 47:17, at
378–80 n.4 (citing over 60 cases, although a review
shows that about 20 of them support applying
ejusdem to the opposite pattern).
44 Scalia & Garner, supra note 2, at 204.
45 Mich. Comp. Laws § 324.73301(1).
46 Scalia & Garner, supra note 2, at 205.
47 See Kenneth A. Adams, A Manual of Style for
Contract Drafting §§ 13.624–.634 (3d ed.
2013) (discussing the ambiguity of such as and
advising drafters not to use it).
48 Scalia & Garner, supra note 2, at 195.
49 Jellum, supra note 34, at 136.
50 David L. Shapiro, Continuity and Change in Statutory Interpretation, 67 N.Y.U. L. Rev. 921, 929
51 Jones v. St. Paul Ins. Co., 725 S.W.2d 291 (Tex.
App. Corpus Christi 1987).
52 Id. at 292.
53 Scalia & Garner, supra note 2, at 196.

54 Shapiro, supra note 50, at 930.
55 See Adams, supra note 47, §§ 13.280–.288
(advising against listing common or obvious
specifics regardless of whether the general term
would come first or last).
56 Holy Angels Acad. v. Hartford Ins. Grp., 487
N.Y.S.2d 1005,1006 (Sup. Ct. 1985).
57 Id. at 1007.
58 For other cases that did the same, see Adams, supra
note 47, §§ 13.275–.277.
59 Scalia & Garner, supra note 2, at 199–200.
60 780 F.3d 647 (5th Cir. 2015).
61 Id. at 650–51.
62 Id. at 651.
63 Id. at 659.
64 See Bryan A. Garner, Garner’s Dictionary of
Legal Usage 50 (3d ed. 2011).
65 Kaluza, 780 F.3d at 661.
66 Id. at 661–62.
67 Id. at 662.
68 Id. at 658.
69 Joseph Kimble, The Doctrine of the Last Antecedent, the Example in Barnhart, Why Both Are Weak,
and How Textualism Postures, 16 Scribes J. Legal
Writing 5, 37–41 (2014–2015).
70 Kaluza, 780 F.3d at 669.
71 Id. at 661.
72 Id. at 659 (citing 1 U.S.C. § 3, which defines vessel
as “includ[ing] every description of watercraft or
other artificial contrivance used, or capable of being
used, as a means of transportation on water”).
73 Id. at 667.
74 135 S. Ct. 1074 (2015).
75 Id. at 1081.
76 Id. at 1097 (Kagan, J., dissenting).
77 Merriam-Webster’s Collegiate Dictionary 926
(11th ed. 2003).
78 See Joseph Kimble, The Puzzle of Trailing Modifiers,
Mich. B.J., Jan. 2016, at 38; see also Lockhart v.
United States, 136 S. Ct. 958 (2016) (dramatically
illustrating the conflict between last antecedent and
series qualifier).
79 Yates, 135 S. Ct. at 1092 (citations omitted).
80 Id. at 1085.
81 Id. at 1097, 1098.
82 Id. at 1079.
83 Cf. Richard A. Posner, Divergent Paths: The
Academy and the Judiciary 116, 117 (criticizing
Yates as “distended” and “canons-riddled” and
asserting that an opinion explaining the decision
“need not have exceeded three or four pages”).
84 Id. at 104.
85 Id. at 105.
86 Benson, supra note 14, at xv.
87 For the view that statutory interpretation is always
both textual and purposive, see William N.
Eskridge Jr., Interpreting Law: A Primer on
How to Read Statutes and the Constitution
(forthcoming 2016).
88 See Kimble, supra note 69, at 28–37 (citing six
empirical studies of Justice Scalia’s opinions, along
with extensive commentary, for the contention that
textualism as practiced has little claim to objectivity
or political neutrality).