Category Archives: Legal Drafting

Should You Use “Shall”?

In legal drafting, should you worry about using shall? Consider three examples:

  • Subcontractor shall comply with all Contractor safety rules.
  • “Seller” as used in this contract shall mean Oaker Services.
  • Members of the panel shall be selected by the commissioner.

Although shall is generally harmless in those examples, taking more care when using shall would be a good thing for three reasons.

1. Shall can be ambiguous.

Courts have construed shall to mean both “must” (mandatory) and “may.” Here are two statements about the meaning of shall from the Texas Supreme Court:

  • We agree, of course, that ‘shall’ is mandatory language.” C. v. M.B., 650 S.W.3d 428, 443 (Tex. 2022) .
  • “The word ‘shall’ in a statute may be and often is held as merely directory and as having been used in the sense of ‘may.’” Thomas v. Groebl, 212 S.W.2d 625, 630 (Tex. 1948).

These conflicting meanings of shall have prompted scholarly criticism in articles like “The Many Misuses of Shall,” “Shall We Proceed?” and “Shall Must Go.”[1]

2. Shall is used in multiple senses in the same document.

The preferred meaning of shall in legal drafting is to create a duty and to impose that duty on an actor:

Subcontractor shall comply with all Contractor safety rules.

  • Here, shall imposes a duty on the Subcontractor.

But shall is also used to mean that something will or must occur or be treated or viewed in a certain way:

“Seller” as used in this contract shall mean Oaker Services.

  • Here, shall is not imposing a duty; it is saying that the term “Seller” refers to Oaker Services.

This usage isn’t ideal: legal drafters shouldn’t use the same word to meaning different things—no citation needed. By the way, Black’s Law Dictionary gives five senses of shall in legal writing: has a duty to, should, may, will, and is entitled to.[2] One leading expert on contract drafting, Kenneth Adams, says that as to being used with differing meanings, “the word most abused in that regard is shall.”[3]

3. Shall is archaic.

Shall is an outdated word that, in the U.S., is rare in written or spoken English. Bryan Garner explains that today shall is usually used in only two situations: “(1) interrogative sentences requesting permission … <shall we all go outside?>; and (2) legal documents ….”[4]

Recommendations

You could limit shall to one meaning in documents by scrutinizing every shall to ensure that it imposes a duty on an actor. One good test: try substituting the phrase “has a duty to” for shall.[5] If it makes sense, you’ve used shall correctly:

Subcontractor shall [has a duty to] comply with all Contractor safety rules.

  • Correct use of shall.

“Seller” as used in this contract shall [has a duty to] mean Oaker Services.

  • This usage doesn’t make sense: How do you impose a duty on the term “Seller”?
  • Drop shall: “Seller” as used in this contract means Oaker Services.

Members of the panel shall [have a duty to] be selected by the commissioner.

  • This usage is not correct: It’s the commissioner who has a duty, not the panel members.
  • Will works: Members of the panel will be selected by the commissioner.
  • Active voice works: The commissioner shall [has a duty to] select members of the panel.

If scrutinizing every use of shall in your document would be too time-consuming and costly for your client, you could reasonably decide to leave the shalls as they are—especially if you’re drafting contracts. Although there are hundreds of cases of courts construing shall, nearly all of them concern statutes, not contracts.[6]

Finally, if you’re drafting simple agreements for nonlawyers (residential leases, consumer notices, website disclaimers), steer away from shall. For those, you could simply use “agrees to,” and if “agrees to” doesn’t make sense in a certain sentence, then shall likely wouldn’t have been correct, either.

_____

[1] Joseph Kimble, The Many Misuses of Shall, 3 Scribes J. Legal Writing 61 (1992); Alex MacDonald, Shall We Proceed? Ebbs, Flows, and Futility in the Debate over Words of Authority, 20 Scribes J. Legal Writing 81 (2022); Michele Asprey, Shall Must Go, 3 Scribes J. Legal Writing 79 (1992).

[2] Shall, Black’s Law Dictionary (11th ed. 2019).

[3] Kenneth A. Adams, A Manual of Style for Contract Drafting 9 (2013).

[4] Bryan A. Garner, Garner’s Modern English Usage 825 (4th ed. 2016).

[5] Bryan A. Garner, Garner’s Dictionary of Legal Usage 952 (3d ed. 2011).

[6] 39 Words & Phrases 173-239 (2006).

Readable Contracts Part 4

Archaic legalese

My books: Legal Writing Nerd and Plain Legal Writing

This is part 4 of a series based on the study, Poor Writing, Not Specialized Concepts, Drives Processing Difficulty in Legal Language,[1] in which the authors compared contract language with everyday written English. The series concludes with my comments on a few words found in a 1.1 million-word corpus of commercial contracts. But first, I’ll acknowledge reality.

Lawyers prepare commercial contracts using forms and templates, and that saves time and money. It also provides assurance—risk avoidance. Suppose the form contract has been used in 20 or 30 or 50 other transactions, all of which closed and were performed without litigation. By relying on that form, you avoid risk, reassuring yourself and your client that this transaction, too, will be performed without serious problems. So retaining and re-using forms is a good practice, even if the forms use some archaic legalese.

But may I offer a few suggestions?

The following words are unnecessary because there are everyday equivalents, and some of them cause problems—albeit rarely—so I recommend deleting and replacing them. Parentheses show the number of appearances in the contracts corpus.

aforementioned (15), aforesaid (49)

The main problem with aforementioned is not that it’s a multi-syllabic monster; the problem is that it’s vague. As I said of aforementioned in 2008: “Why use this outdated word when its shorter cousin, aforesaid, is available? I’m kidding. Eliminate them both and specify what you’re referring to.”[2] In addition, the meaning of aforesaid has been construed in reported appellate decisions at least five times.[3]

herein (1093), hereinabove (7), hereinbefore (10), hereinafter (120)

Again, the problem is vagueness. As the legal-language expert David Mellinkoff put it, “Where? This sentence, this paragraph, this contract, this statute? Herein is the start of a treasure hunt rather than a helpful reference. The traditional additives are equally vague: hereinabovehereinbeforehereinafter ….”[4] Also, herein’s meaning has been litigated in at least 11 reported cases.[5]

said (214)

When used as a demonstrative pronoun or “pointing word,” said adds no precision, only a legalistic tone. As the contract-drafting expert Tina Stark says, “Said and such are pointing words. They refer to something previously stated. Replace them with the, a, that, or those.”[6] If “that party” is vague, changing it to “said party” won’t clear it up. And said’s meaning has been litigated 30 times.[7]

whereas (224)

This word appears in the formal, archaic recitals that proceed with a series of paragraphs beginning with “WHEREAS” and conclude with “NOW THEREFORE …” But Kenneth Adams, a leading expert on contract language, doesn’t like whereas: “The recitals serve a storytelling function. They’re the one part of a contract that calls for simple narrative prose. So don’t begin recitals with whereas, as this meaning of whereas—‘in view of the fact that; seeing that’—is archaic.”[8]

witnesseth (21)

At first, I found only 8 occurrences of witnesseth in the contracts corpus, and I was surprised but happy to think that its use was declining. Then I searched for it with a space between each letter: W I T N E S S E T H, and found 13 more. I think it needs to go, and so does the legal-language expert Bryan Garner: “This archaism is a traditional but worthless flourish. … There’s absolutely no reason to retain witnesseth. It’s best deleted in modern contracts.”[9]

Ultimately, retaining these words is probably harmless, but removing them is, too.

My books: Legal Writing Nerd and Plain Legal Writing

_____

[1] Eric Martinez, Francis Mollica, & Edward Gibson, Poor Writing, Not Specialized Concepts, Drives Processing Difficulty in Legal Language, Cognition 224 (2022).

[2] Wayne Schiess, Ten Legal Words We Can Do Without, Austin Lawyer 6 (May 2008).

[3] 2D Words & Phrases 294 (2020).

[4] David Mellinkoff, Mellinkoff’s Dictionary of American Legal Usage 283 (1992).

[5] 19A Words & Phrases 36-37 (2007 & Supp. 2021).

[6] Tina L. Stark, Drafting Contracts: How and Why Lawyers Do What They Do 257 (2d ed. 2014).

[7] 38 Words & Phrases 29-31 (2002 & Supp. 2021).

[8] Kenneth A. Adams, A Manual of Style for Contract Drafting 20 (3d ed. 2013).

[9] Bryan A. Garner, Garner’s Guidelines for Drafting & Editing Contracts 454 (2019).

Readable Contracts Part 3

Readability, grade level, and sentence length

This is part 3 of a series discussing the study, “Poor Writing, Not Specialized Concepts, Drives Processing Difficulty in Legal Language,”[1] in which the authors compared contract language with everyday written English. I was able to contact the authors and access the corpora they used, and I conducted my own assessments. (Note: the corpus of contracts I used had 837,000 words; the corpus of everyday written English had more than a million words.)

I assessed the text for average sentence length, Flesch Reading Ease, and Flesch-Kincaid Grade Level, and also included those averages from my last 10 columns that have appeared here. The results:

These results support the conclusions of the original study’s authors, but I’ll say a bit more about them here.

 

 

 

Avg. Sentence Length

Flesch Reading Ease

Flesch- Kincaid Grade

Everyday written language

17 56 9

Contract language

42 20 19

Schiess’s last 10 columns

17 52 10

Average sentence length

The average for the everyday English—17 words—is short but typical: everyday-English sentences average 15 to 20 words. This corpus included blogs, web pages, and TV and movie scripts, so the low number makes sense.

The 42-word average for the contracts is, well, huge. As I pointed out last month, these are commercial contracts entered by sophisticated parties represented by counsel, so the long sentences aren’t as troubling as they might be if the contracts were apartment leases, credit-card agreements, or software-user agreements.

The 42-word average means that there are some really long sentences, and even experienced transactional lawyers might find reading those long sentences difficult.

Flesch Reading Ease

This formula, included in Microsoft Word, was finalized in 1948 by Rudolf Flesch, an Austrian lawyer who fled the Nazis in 1938 and earned a Ph.D. in education in the United States. It assesses the number of syllables and sentences per each 100 words and uses that assessment to produce a score from 0 to 100: 30 is difficult, and 60 is plain English.[2]

At 56, the everyday English text comes close to Flesch’s standard for plain English—as it should. And as we might have predicted, the Reading Ease score for the contract language is low—what Flesch labels “very difficult.”[3] The long average sentence length doubtless contributes to this low score, but the average number of syllables per word surely does, too.

Flesch-Kincaid Grade Level

This scoring system was derived from the Flesch Reading Ease score by J.P. Kincaid[4] and reports the number of years of formal education a reader needs in order to understand the text. The everyday English scored a 9, meaning one who has completed the ninth grade should be able to read and understand it. My own writing—which is mostly about writing—tends to hover around the tenth-grade level.

The Flesch-Kincaid grade level for the contract language is high at 19, although I once read a decision from an administrative-hearing appeal that scored a 20. But grade level 19 is, unsurprisingly, the equivalent of a reader with a high-school education (12), a college degree (16), and a law degree (19).

Thus, the grade level is appropriate given the context: these contracts were prepared by and for attorneys.

Recommendations

Still, the 42-word average sentence length is taxing at best and on the border of impenetrable. Anything we can do to reduce that average will make the contract easier to read and understand and, therefore, easier to draft, easier to review, and easier to explain to the client.

Next month: a report of additional findings based on the two corpora.

_____

[1] Eric Martinez, Francis Mollica, & Edward Gibson, Poor Writing, Not Specialized Concepts, Drives Processing Difficulty in Legal Language, Cognition 224 (2022).

[2] William H. DuBay, Smart Language: Readers, Readability, and the Grading of Text 56 (2007); Rudolf Flesch, How to Write Plain English 25 (1979).

[3] Flesch, How to Write Plain English at 25.

[4] DuBay, Smart Language at 90-91.

Readable Contracts, part 2

Three more findings—with advice

This is part 2 of a series discussing the findings of a study, Poor Writing, Not Specialized Concepts, Drives Processing Difficulty in Legal Language,[1] in which the authors compared contract language with everyday written English. Here are three more findings: Contract language has higher frequencies of—

  • center-embedded sentence structure
  • words used rarely in everyday English
  • words with a higher-frequency synonym (fancy words that could’ve been simpler)

Center embedding means inserting a phrase or clause within another phrase or clause. Here’s an example from the authors:

In the event that any payment or benefit by the Company (all such payments and benefits, including the payments and benefits under Section 3(a) hereof, being hereinafter referred to as the “Total Payments”), would be subject to excise tax, then the cash severance payments shall be reduced.[2]

In a sentence of 47 words, placing another clause of 22 words in the middle makes for difficult reading. An easy fix is to place the embedded definition clause in a separate sentence:

In the event that any payment or benefit by the Company would be subject to excise tax, then the cash severance payments shall be reduced. All such payments and benefits, including the payments and benefits under Section 3(a) hereof, are hereinafter referred to as the “Total Payments.”

Now we approach a thornier topic: Even without this study, any lawyer and anyone who has read a contract could’ve told you that contracts use words that are rare in everyday English as well as words that have simpler or more-readable synonyms. But is that a problem?

On this topic, I won’t offer recommendations for using shorter or simpler words in contracts. Yes, doing so could make contracts more readable, but I’ll propose three reasons that doing so might not be ideal.

  1. The studied contracts were commercial contracts entered by sophisticated parties represented by counsel.

In my review of the contracts in the corpus, I didn’t see a single consumer contract: apartment lease, credit-card agreement, software-user agreement, car insurance.

So the studied contracts don’t necessarily need to be read and understood by someone without legal training. For the contracts in the corpus, those who need to read and understand the language are lawyers, and those lawyers—we hope—can explain the contract language to those who need to understand it.

  1. Contracts contain some legal terms that either cannot or should not be simplified.

Yes, contracts use words and phrases unique to legal language or with a different meaning from the everyday-English meaning. But replacing those words and phrases to enhance readability could introduce risk, or it could require the drafter to use even more words to explain what the legal term means.

But note: the number of terms of art or unique legal terms is fairly small—smaller than some legal drafters claim. In one study, only 3% of the words found in a standard real-estate-purchase agreement had ever been construed or defined by a court.[3]

  1. Revising lengthy, complex contracts for readability might not be cost effective.

Given that the commercial contracts in the study were prepared by, reviewed by, and could (I assume) be explained by transactional lawyers, how would we justify the cost of revising them? We’d need to replace rare words with everyday words and replace or explain legal terms. Who’s going to pay for it?

These contracts were certainly based on forms or precedents from previous transactions, a practice that saves time and money. Add to that fact the reality that very few contracts result in “disputes” (as high as 9% according to one commercial source[4]) and that even fewer end up in litigation (fewer than 0.1% by one estimate[5]).

Next month: my own findings based on the corpora from the study.

__________

[1] Eric Martinez, Francis Mollica, & Edward Gibson, Poor Writing, Not Specialized Concepts, Drives Processing Difficulty in Legal Language, Cognition 224 (2022).

[2] Id. at

[3] Benson Barr, George Hathaway, Nancy Omichinski, & Diana Pratt, Legalese and the Myth of Case Precedent, 64 Mich. Bar. J. 1136, 1137 (Oct. 1985).

[4] Contract Assistant, Not Good: Average of 9 Percent of Contracts Result in Dispute, https://contractassistant.com/​not-good-9-percent-of-contracts-result-in-dispute/ (accessed June 23, 2022).

[5] Id.

Readable contracts Part 1

ALL CAPS and passive voice

This is part 1 of a series on contract drafting; I’ll report on a study of contract language and offer comments and recommendations. The study is Poor Writing, Not Specialized Concepts, Drives Processing Difficulty in Legal Language.[1]

The three authors (two linguists and a lawyer/linguist) used corpus analysis to discover why contract language “remains notoriously inaccessible” to nonlawyers. They asked which of these two causes could account for the difficulty:

  • Is it the specialized and complex content?
  • Is it the form of expression—the way contracts are written?

They compared a corpus (think “database”) of contracts with a corpus of standard English—newspapers, magazines, blogs, web pages, and TV and movie scripts. The two corpora[2] contained more than 10 million words, and the authors assessed five variables: frequency of all caps, frequency of passive voice, frequency of center-embedded sentence structure, frequency of everyday words, and frequency of words with a higher-frequency synonym (fancy words that could’ve been simple).

The authors found that “all of the metrics we looked at were prevalent to a greater degree in contracts than in the standard-English corpus.”[3] Let’s start with the first two variables: contracts use all caps and passive voice more than everyday writing.

I’m not surprised that contracts use more all caps than other writing, but the questions is, why? Typically, all caps are used to draw attention, to make text conspicuous and noticeable. In fact, there are Texas statutes that require certain contract language to be “conspicuous.”[4] But those statutes don’t expressly require all-capitals text. For example, here’s one definition of conspicuousness: “Required information in a document is conspicuous if the font used for the information is capitalized, boldfaced, italicized, or underlined or is larger or of a different color than the remainder of the document.”[5]

So all caps is one option, but not the only one, although I did find two Texas regulations requiring all-caps.[6]

Still, the question remains: why does all-capitals text persist in contracts? Three possible reasons and a recommendation: First, all-caps is a vestige of the typewriter, which couldn’t produce boldface or italics, so some form contracts retain all-caps because they’ve never been updated. Second, all caps really do stand out if the rest of the contract is in regular type. Third, some lawyers mistakenly assume that statutes require all caps for conspicuousness; but although some statutes mention all-caps, they almost always give other options. (See the example quoted above.)

Recommendation: Convert all-caps to boldface—and maybe even increase the font size. Blocks of all-caps text are difficult to read[7] and are nowadays perceived as shouting.

Why do contracts have more passive voice than other ordinary writing? Two possible reasons: Passive voice is preserved because “it’s in the form.” Many contract drafters are wary of changing form language, especially if that form was the basis for numerous contracts that have closed and been performed without a glitch.

And passive-voice just sounds more formal—more lawyerly. But if that’s a source of excessive passive voice, we can let it go. Consider these examples of passive voice that are clearer in the active:

1. Permits must be secured before work commences.

  • By whom? Better in active voice:

1a. The owner or contractor must secure permits before work commences.

2. The Purchase Price shall be paid by wire transfer of immediately available funds.

  • Certainly, the buyer’s obligation to pay the Purchase Price was stated earlier in the contract, but I still prefer this active-voice version:

2a. The Buyer shall pay the Purchase Price by wire transfer of immediately available funds.

Next month: additional findings of the contract-language study.

_____

[1] Eric Martinez, Francis Mollica, & Edward Gibson, Poor Writing, Not Specialized Concepts, Drives Processing Difficulty in Legal Language, Cognition 224 (2022).

[2] Yes, that’s the plural of “corpus,” which I had to look up,

[3] Poor Writing, Cognition 224, at 3.

[4] Tex. Bus. & Com. Code Ann. § 2.316.

[5] Tex. Bus. Orgs. Code Ann. § 1.005

[6] 7 Tex. Admin. Code § 84.808; 16 Tex. Admin. Code § 22.71.

[7] Matthew Butterick, all caps, Butterick’s Practical Typography, https://practicaltypography.com/all-caps.html (last visited June 1, 2022).