The Constitution:
        Has Anyone Proofread Our Copies?

Not very well.

        People act partly on what the Constitution says, what attorneys say it says, what the courts say it says — and, it turns out, what top publishers say it says.

        Schoolchildren are taught the Constitution’s words. Write “We the People” and most of the public will recognize your flourish. Cheap almanacs serve it up in full text.

        Professional renditions, however, on which judges, lawyers, and professors of law rely without a second thought, are the epitome. We being duly diligent, we expect these sources to be thoroughly correct. Stoically right.

        Hah.

No less than the United States Code, United States Code Annotated, United States Code Service, United States Statutes at Large, United States Code Congressional and Administrative News, the Federal Register, pamphlets from the Senate and the House of Representatives, Black’s Law Dictionary, the interpretive Constitution Annotated from the Library of Congress, and the Keepsake Edition once autographed by Supreme Court justices — all these had discrepancies.

        I did this research primarily in 2002 and 2003.¹ I plan to leave its updating to other researchers. Perhaps the publishers of these renditions would be pleased to update the research themselves.² One may presume that corrections won’t have occurred without an examination, and corrections are urgently needed.

        Punctuation, capitalization, spelling — and, if that’s not enough, some entire words are wrong. In total, over a thousand discrepancies abound.

        There’s more. Footnotes are oft absent when explanations would turn errors into valid decisions. If an editor believes an element of drafting was meant to convey a certain content and edits the final rendition accordingly, a note is warranted and probably required for accuracy and scholarship. But absent explanation or excuse, errors stay error.

        The only Constitutional rendition that probably is error-free is on the National Archives web site, specifically the set of High Resolution images, showing original handwriting. Even that until recently included only twelve of the twenty-seven amendments, the rest having been displayed until then in a retypeset rendition.³

        I didn’t bother checking the dime-store paperbacks. We expect them to be mistaken in places.⁴ Surely no lawyer would base a brief on one of those. If we want to be right, we know where to turn. Don’t we?

        At least one pocket edition was in use by a High Court justice. Did anyone check it?

        You probably assume these books are correct. You have common sense. These publishers — I don’t know.

        Electronic services, from two publishers of hardbound sets cited above, overlap the books. Cost control theory suggests that one file would generally be used to produce all media, albeit with a tradeoff of propagating any errors widely. And if different files are being used, that risks introducing new errors.⁵

        Vagueness can hide errors. Citations drawn from electronic research reputedly rarely declare their true sources, preference being given to citing prestigious, traditionally stitched books while omitting parallel citations. It’s the modern version of the problem with advance sheets, when judges revise opinions. How many users openly cite advance sheets or update quotations when replacement volumes arrive? Will anyone check whether their online sources are correct, or whether their Constitutions are right? And if hard copy sources are cited when electronics are used in fact, resulting in error, the wrong publisher may be blamed, or, worse, the lead citing author may be faulted. Thus, everything significant citing anything that has parallels online may, for practical purposes, be untraceably and incurably infected with error.

        One hopes that writings have enough textual redundancy that readers catch errors as nonsense. One hopes that all errors look like nonsense. But they don’t. And many writers fairly succeed at concision, reducing redundancy and thus concealing errors.

        But not all is lost. Now the more venerable publishers can peruse a public database of hard copy faults, the same one you can see:

        What’s more, in 1787 humans had foibles. They erred, too.

The framers didn’t write just one master and fax it to the states.

        Four large pages were laden with writing, as was the Bill of Rights.⁶ I doubt any of the 14 sets, 13 plus the Federal, were printed on a press. Virginia’s copy of the Bill of Rights, photographed in Life magazine, was handwritten.

        All that wrist-work left room aplenty for slips of the pen. The original set that remains in Federal custody, and presumably is the most supremely authoritative of all, had errors. The Secretary at the time described several corrections in a lengthy note directly on the fourth page of the original.

        Literacy wasn’t exactly wall-to-wall back then. Authors often misspelled family names, never mind getting everything right. And they, at least, were authors. Three quarters of the country’s adults, give or take, couldn’t sign their own names.⁷ If we’re lucky that a quarter were literate at all, we couldn’t demand a very high standard from those who were. Pretty good would have to do.

        Which raises the question about why the best copy wasn’t left with the Federal government: — Actually, maybe it was.

        When writing 13 copies for the states, upon seeing an error in the middle, would a scribe have mindlessly made more mistakes? Wouldn’t someone have decreed it wiser to cease inking errors before multiplying the damage by 14? If writing all day meant some errors were likely, why add deliberate errors, too? How would it look for a new national Congress to propagate bad writing in its single most vital pronouncement of law?

        You could argue for precision, but that’s where cameras are good, and Kodak and Xerox weren’t in business yet. The scribes would’ve been under pressure, because the easiest way to get the copies to the states would’ve been to give them to the delegates, and they were leaving soon. There was no thought of FedEx-ing revisions later. The discussion would’ve been short, the outcome obvious. Where there’s a quill, there’s a way. We can imagine what they wrote.

        Maybe someone will open states’ archives, and we won’t have to imagine anymore.

Did all the states ratify the same Constitution?

        The later admissions to the union probably did, telecommunications having improved. But in 1787 the long-distance carrier of choice was the horse. That could discourage proofreading.

        Proofreading would’ve been easier said than done. If any newspaper printed the whole Constitution, using that without checking it against a handwritten copy would have left the newspaper’s accuracy in question, especially germane as many publishers were partisan. No; proofing would have to be done with the authentic original, miles away. But anyone proposing that would have had to grapple with the logistics.

        Pity the proofreader. Whoever had that chore needed time and money. If they couldn’t snag a stomach-rattling stagecoach or sail the storm-ripped Atlantic, hoofing it into muddy streams and across slanted mountain scree would have to do.

        Horses weren’t fast when routes were long. Mileage varied with the oat supply. Terrain conditions, autumn and winter weather, riders’ moods, and snakes a-hissing impeded schedules. Horses probably didn’t feel kindly toward whips and spurs and the men who loved them.

        Would a legislator have made an extra copy by hand, proofread it character for character, made decisions about calligraphic ambiguities and other uncertainties, clambered aboard a favorite horse, traveled for up to a fortnight, sought food, shelter, and veterinary care for two along the way, sat in front of the original Constitution, proofread against it character for character, made decisions about calligraphic ambiguities and other uncertainties, collected their notes and analyses, ricocheted atop a saddled spine for up to another two weeks seeking food, shelter, and veterinary care for two along the way back, and, upon discovering part of their notes obliterated by mud and rain, dissembled to their legislative colleagues and leaders about what they purportedly had or had not found?

        Personally, I doubt the Constitution was proofread by horse.

        Legislators fixing to deliberate at their capitol had to come from around the state, but couldn’t leave home right away, if they were to get a feel for constituents’ concerns and avoid abandoning the harvest before the frost, back when harvests were less frequent. Then, they had to get to the capitol in order to debate and decide. They had no time for diddly stuff.

        Did debaters desist while a colleague busily proofreading was beyond hollering range? Who stops debating even today just because their data might be a little short? Doubtless, they were comfortable relying on what they had in hand. Decisions could be made, alliances fashioned, factions hardened, and eggs thrown and left to dry long before any neigh-sayer clippety-clopped into town with news about grammar and spelling, not an electrifying subject. Right now, you can dismiss many a dispute by calling it “just semantics.”

        No legislator would have tolerated missing capitol happenings so he could contemplate commas. Instead, any absentee would have been someone rather inconsequential who could be spared, but who still would have had political credibility with legislators pro and con. Otherwise, upon return, any inconvenient news could be ignored. Secretaries today have had that experience. Cabinet Secretaries have had that experience. Messengers can testify that being ignored is more humane than being shot, but just as effective.

        Headhunters for inconsequential proofreaders doubtless scratched most names off. Women were told to stay home. No frigid three-week journey with both legs on the same side of a galloping horse would do for a wife and mother, when a husband and his manly friends heard about it. Older men over 65 would’ve oft been thought too frail to bounce on boulders. African Americans, called Negroes when slavemasters chose politeness, weren’t supposed to read.⁸ Any politician sponsoring a traveler too contra custom could probably forget about being reelected.

        Rural people were almost all of the population, by nearly 15 to 1,⁹ but the action was at the capitol, and legislators would likely pick someone with whom they had worked. That’d be someone proximate, because long commutes were probably rare and trust would be critical, and trust is scarcer when it depends on communications longer than a stone’s throw. I’m not convinced a descendant of English Mayflower ancestry would have taken up a German, Iberian, Welsh, Scottish, or Scotch descendant’s offer of help, never mind one from a tribal Indian. Regardless of origin, whoever would go had to be literate; most weren’t. Almost any adult still left was probably otherwise occupied.

        At the maximum, Delaware had perhaps 13 acceptable proofreaders; Massachusetts, nearly 6,400 (the most populous city being the capital); New Jersey, just four; Georgia, ten.¹⁰ Would Georgia have trusted a Massachusetts type? With states yielding sovereign rights to a stronger central government, one state wasn’t about to believe what another state spun. No one cared to be robbed of rights. And, presumably, no one would miss the chance to lobby and line up colleagues for their view, or miss a high-profile vote on the Constitution itself.

        Ratification was rather rapid, considering what they were considering. Georgia ratified the original Constitution in three and a half months, Massachusetts nearly a month later, in a close vote. New Hampshire ratified the first ten amendments within four months of submission, South Carolina in slightly less.¹¹ Among elected officials, that wasn’t the time to disappear. Whether to persuade or merely to proofread meant choosing between stardom and tedium.

        It could be that no one went. When few are available, other priorities may claim their time.

        Nonetheless, assume arguendo that at least one state did it. We’d want to know whether anyone filed a report of proofreading discoveries, whether the report was accurate, and whether it was implemented.

        We can only surmise about that history, but we have a modern comparison.¹² Until C-SPAN started telecasting Congressional discourse, transcripts heading into the Congressional Record were reportedly often rewritten by Congressional members’ staffs, whose jobs were to protect their bosses’ jobs. The Congressional Record is essentially determinative for Federal floor debates on intent. If our official records were severely mangled just decades ago, you could wonder whether they were any better centuries ago.

        Any interest legislators had in augmenting archives in those days likely took a back stool to nation-building, public relations, running farms and businesses, and getting reelected. Newspapers were smaller then, carrying less news per month, making for less transparency. And with fewer people back then, fewer could be aides, resulting in fewer records, less attention to detail, and less checking.

        I wouldn’t be sanguine about that.

We could reproofread every twenty years.¹³ Epistemological standards and the content of our knowledge improve over time.¹⁴ Historians review their findings; scientists reconsider hypotheses. Lawyers who specialize find utility in rereading basic law of their fields every year or so, even without specific research goals.

        In olden days, a page of cold type occasionally crashed to the floor. Nowadays, a computer file crashes. Either event could rearrange our Constitution. Gross errors aren’t so bad; they’re seen quickly. Subtle errors slip through. We must catch them ahead of a court’s cryogenic glare in a crucial ruling. Methods of proofreading are thus ripe for review.

        It’s boring work, proofreading.¹⁵ Slough it off one too many times on the same serf’s back and they start counting the days till they get another job.¹⁶ That dissipates any desire to protect the publisher’s long-term interest. This may be unrelated, but one 58-word clause in USCS, just one clause, had 24 divergences.¹⁷

        Notice, in the appended lists, how many errors are homophonous. They sneak through more readily when word-checking is rushed and no one enunciates each character.

        How much you should pay for proofreading depends on how much diligence you want, or how much is due. Two listeners plus one audible reader are more accurate (all else equal) and cost more than one or two people doing it all. Some evidence indicates that the listeners should be separated in order to increase their total accuracy rate.¹⁸ Character-by-character recitation, stating whether capitalized and if italicized, is more accurate and takes longer than word-by-word, maybe thrice as long, maybe much more.¹⁹ In general, avoiding being hasty improves rates of error capture.²⁰ The Constitution, of course, is worth it.

        Imagine, as I have, modern editors toiling away at the words of this supremely august text. They are precisians first; that is why they were hired.

        I see two editors, one above, one below. The assignment is given: Copy the text. So proceeds the underling, inscribing from an unimpeachable publication.²¹ The work product is duly transmitted to the overseer of all scribes.

        The overseer, knowing whom to trust and not having all day, spot-checks here and there. Pulling yet another entirely certifiable authority from a handy shelf, a couple of errors are found, and the supervising editor corrects those with pride, promptitude, and pluck.

        The Editor Superior tells The Peon that the scribing was well done and that only a couple of errors were found, which have now been corrected, and thanks are offered all around. The exact errors are unidentified, thus preventing the lower editor from defending the composition, thus not challenging the supervisor, who’s pleased. Nothing is said about exactly which authorities were relied upon by each editor. That might open the speaker to a question of competence, and why touch that?

        Thus, they fail to coordinate upon what authority they will rely.

        You can confirm this in the appendicular lists. Some content is entirely error-free in all renditions; whereas thirty-six fragments differ in seven to thirteen renditions each.²² Evidently, publishers shared many of the same errors.

        Now, in effect, in our hypothesized situation, we are disclosing the failings of both editors, and they are impressive, indeed. But suppose we didn’t insist on hanging the entire list in public view like laundry.²³ Suppose we quietly slipped our lists to the editors without telling anyone else. What, in our scenario, could we expect to happen next?

        The High One Above would tell the Mere One Below to look into allegations of error. This gives a double problem to the inferior one, who has no tenure. More important than the accuracy of the Constitution (anyone can handle that) is the need to stay employed. Thus inspired, the subordinate will note a quandary: If the editor in the teensy cube reports to the Editor With the Corner Acreage that there are 162 errors not counting captions, won’t it be easier for the Editor In the Clouds simply to fire the editor of no moment? That would show all the powers-that-be that something was done, thereby preserving high majesty despite the facts.²⁴ That’s why public scrutiny is handy.²⁵

        One lesson for proofreading: Editors at all levels should agree — in advance — on precisely which sources constitute proper authority for any text.

        Another lesson: Be transparent with alleged errors and differences of opinion. Supervisors should list them, so editors beneath can challenge or accept each one. That will let editors, both above and below, learn. Learning is a good thing.

        Another: Hesitate to cut corners.

        Conceptual editing also needs review.²⁶ If a judgment call shaped the text, the judgment should be transparent. A note will be apropos. That no competitor among publishers saw fit to attach a note to the same hook is irrelevant. Let the judges and advocates hash out the final decisions.

        Proofreaders, while proofing notes already present, can identify where new notes are needed. So, their managers should listen to their suggestions.

        Here’s where one note would have helped:

That’s in the handwritten original. It’s not a known form of praise, nor of prowess.²⁷ Later publishers thought the Sixth Amendment drafters meant “process”. But how come no one tells us why they transmogrified praeess into process? Footnotes are perfect for explaining such decisions.

        If a linguist ever finds praeess on an Anglo-Saxon tablet, we may need to reconsider what the framers were thinking. Until then, accurate quotation will help level the playing field.

Could a clash among circuits and state courts be due to a textual conflict? Might not finding an error in one court’s quotation of the Constitution and applying its announced reasoning result in a new conclusion, instigating a collision with another court? Old decisions newly colliding might make interesting appeals. The opportunity to distinguish case law, even established case law, justifies compiling misquotations evincing judicial misunderstandings.²⁸

        This isn’t about the intent of the framers, the spirit of anything, or the need for interpretational flexibility to accommodate modernity. This is about plain words and syntax.

        But can you challenge state ratifications for nonidenticalness? Suppose a transcription error impacts a cause today. Wouldn’t the judiciary conclude that it’s a bit late? Perhaps not, but antiquity has weight in law. Any ability to question the text presumably existed since it was written. Compelling indeed would the argument have to be that would persuade five justices to heave two centuries out. And wouldn’t Wyoming, Texas, and Alaska care to be amici, believing that the first thirteen ratifiers were on the same page²⁹ and therefore that those who ratified in the nifty days of telegraph and Boeing were right to rely on the unitary text widely thought to be perennially proper?³⁰

        If enough ratifications are fatally flawed (and almost all could well be), then no Congress, no President, and no Supreme or inferior Court lawfully exists, or ever did.³¹ We can, holding our breaths, refresh our reading of the Articles of Confederation, assuming they’ve been proofread.

        Specific ramifications can be drawn from some textual differences; see an appendix for a few. You might realize more.

Not only inferential differences from punctuation and word choices, but spelling and perhaps capitalization errors can ramify in dialectal semantics.³² E.g., the words control and controul³³ mean the same thing unless there was a subtle difference if one was historically used in contexts in which the other was not, researching which point may require more than just the best dictionaries.

        Language evolves slowly in the mouths, pens, eyes, ears, fancies, and memories of people who live in a society with many overlapping subcommunities, dialectal and substantive, including lawyers’. Even if one spelling is universal, a new spelling is born and gradually becomes commoner, whilst the older gradually disappears from any but historical reference, with changes arising unevenly in subcommunities. Context controls.

        Dictionary compilers seem to solidify the language as of a date of publication. However, the primary ones don’t. They don’t prescribe; they describe. They collect the better evidence still available, accept readers’ submissions of more, and draw expert conclusions.³⁴ Evidence is always incomplete, rendering all definitions tentative, some more so than others. Just as physicists will not guarantee the law of gravity, only that they know no better explanation for the data, linguists can never be absolute about their conclusions, and the evidence for gravity is heavier.

        Linguistic research will rarely be worth your effort, until you have an extraordinary case with not much else on which to peg a win, but that may not stop an opponent who finds an opening.

Are any laws accurately transcribed?³⁵ One hopes so. So strong is the hope that few of us take the major time needed to check. Opponents aren’t checking. Why should we?

        Municipal and state codes might deserve even more scrutiny than Federal collections.³⁶ The smaller the government, all else equal, the less able it will be to afford pursuing accuracy.

        And antiquity might be unavailable as a counterargument, since most statutes and regulations are younger than key Constitutional provisions. Effectively, counsel may choose which set of rights and obligations obtains by choosing to point out that a rendition is discrepant or choosing utter silence, knowing that no judge can keep current on every bit of law without help. How convenient.

        A few years ago, I looked at a municipal statute. The city has people enough to repopulate ten states, and it has contracts and litigation befitting its size. The codified statute did not conform to the amending text previously enacted. One hopes the code publisher heard about the error even though I said nothing.

        Codification judgments, too, may be erroneous. I uncovered a miscodified Federal provision. It was properly codified at its initiation, but part was repealed, and what was left, as potentially very influential, should have been repositioned or cross-referenced. The government agency responsible for its on-the-ground implementation almost certainly forgot about it. And I liked it forgotten.

        Codes enacted as positive law often permit previous statutes to continue in effect for causes pending at codification, preserving rights. Have any rights been forgotten? How about obligations? Other codes may be prima facie evidence of the underlying statutes. But prima facie evidence is rebuttable; and purportedly conclusive evidence can sometimes be gotten around. Given the numbers of Constitutional errors uncovered, how perfect are the tax statutes?

        Uncodified statutes fall through the cracks. In codifications limited to enactments that are deemed to have general and permanent applicability, some notable laws are elided.

        Annual pacing of code publication seems to have bounded the permanence of a law as being a minimum of a year and a day. If a legislature doesn’t specify at least that duration of effectiveness, specify enactment as positive law within a code scheme, or amend an existing code, the provision it enacts probably won’t make it into a code, not even into a note.

        At first blush, this makes sense. There may be too much temporary law for economical publishing, and they do appear in chronological series in print, with (limited) finding aids. Appropriations are often in this category, and they usually have little lexical impact on litigation. Laws to be implemented in short order that then expire of their own weight are also of this ilk.

        But annual reenactment of temporary or seasonal laws may amount to constructive permanence or semipermanence, in that certain parties can’t assume that obligations are entirely past. Although excluding them is not error, what may be error is not revisiting temporal definitions so as to serve the legal profession and the public. While scarcely will one’s opponents try to dig up anything embarrassing in this arena, it’s hard to claim that there’s no point to having these laws available. Publishers can regularly surface them, offering newsletters to bring the more notable provisions to our attention.

        The other edge of the abyss is “general” as a criterion. Were codes to present statutes that applied to only one named person each (such as many of Congress’ private laws), and therefore to include laws that applied to any larger or vaguer range of entities, the codes might well enlarge like tumors. Yet codes already include many laws that apply to subsets of populations. The question is where to draw the line.³⁷

        I don’t know whether the threshold is too restrictive now or how often the question is researched or rethought by code compilers. But it wouldn’t hurt to reevaluate legislative patterns every so many years, to allow different standards for various jurisdictions, and to state in code front matter what standard a publisher generally applies for the code’s particular jurisdiction.

        Nor need competing publishers agree on one inclusiveness standard, as long as they effectively agree on nonconflicting citational schema, which shouldn’t be hard, since citations can be to already-standardized chronological forms, or a legislative office can offer schematic specifics.

        It is error not to review generality thresholds as societal complexity, legal needs, and legislative customs evolve. Not only would lowering a threshold let smaller subpopulations more easily find laws that apply only to them,³⁸ it would encourage consideration of equal protection claims of classificatory irrationality.

Failures of codification and textual accuracy, not just Constitutional, yield issues counselors can exploit after research.³⁹ And niche practitioners can narrow their research to fewer topics, sensing within them more nuances and discrepancies.⁴⁰ Research need not be case-specific.

        The best time is usually between clients, because most leads will fail, although a few may be very useful and it’s more rewarding to know about them before clients walk in the door. Cyclicity of client demand (an “up-and-down” practice load) provides opportunities of time.

        When challenging common knowledge, extra effort is needed to educate adversaries, agency administrators, and judges, making litigation and appeals likelier. A strategic preference for arguments that already enjoy consensus remains valid, but, when needed, alternatives may succeed.

What the people ratify is what we get.

        We alone may override the popular will so expressed. Our ancestors ratified our Constitution as supreme, and, until we amend it, supreme it stays.⁴¹

        Authenticity is thus warranted.

See the appendices at the Internet address cited ante for discrepancies and other analyses.

        A list of all of the appendices is in Appendix 0. Specific comparisons for each of 16 Constitutional renditions are in Appendices 1–16, with details documenting each rendition and parameters of research and interpretation being in Appendix 17. Implications in law about specific discrepancies are suggested in Appendix 18. Statistics about the discrepancies are in Appendix 19. Some errors in other aspects of law are in Appendix 20. In support of the census of people qualified to proofread during preratification debates, population data are in Appendices 21–22.

Notes:

        * Nick Levinson, P.O. Box 8386, New York, N.Y. 10150, is a nonlawyer.

        The author thanks, among others, New York Public Library and Brooklyn Public Library for having resources useful to this project.

        Trademarks and service marks, registered or not, anywhere appearing, are their owners’.

        1. Reading a major print source years ago, I happened upon some errors. Certainly the publisher would want to fix them. I wrote to them. No reply came.

        No changes ensued, either, as far as I knew.

        A new edition came out. I spot-checked. A bevy of errors littered where I looked.

        I wrote another letter, listing some errors. They’ll fix them, yes?

        Nope. Not a word. As of mid-2004, it’s been about three years. A supplement came out, but with no changes to what mattered.

        That inspired me to proofread at length. Time-consuming it was, but not stellar science. You don’t need a Nobel to compare words. Neither do the publishers. And, for some of these publications, they charge enough to cover the cost of their own proofreading quite generously. Too bad they shortchanged that step. I’d be happy to send a bill.

        2. What liability, if any, attaches when a professional publisher of law has misquoted the Constitution or other law without notice, even if corrected in a later edition, especially if an error was old and long uncorrected, and what the damages could be or how one could calculate for the statute of limitations are not matters I have researched. Nor have I researched whether standing belongs to a misled attorney, to the client of said attorney, or to a court misled in its own research or by the research presented by a party or attorney before it. And an attorney preparing a case on behalf of a plaintiff may have a challenge in researching law without incurring extraordinary levels of expenditure, or in explaining why reliant on the very sources challenged in the plaintiff’s complaint.

        3. I asked that the rest be posted, and now, apparently, they have been. While that’s very useful, their utility can be enhanced, because at least two versions of each amendment should exist in Federal archives: what a Senator or Representative introduced in Congress; then what Congress proposed to the states.

        Ideally, both would be posted for all amendments. While the introduced version would have bound no one if Congress thought better of it and no convention picked it up, if Congress had simply voted to propose the introduced version directly to the states and if what each state considered differed that could be significant. Likewise, if Congress proposed a text that was not what the states considered, because the originally introduced version was publicized among state legislators or conventioneers who then approved it, that, too, would be error. Under the circumstances, these events are not so unlikely. Availability of both texts would facilitate research.

        Illustrative of the possibility of extensive differences is that differences were found between ratified and proposed texts as published in United States Statutes at Large, for which see the additional listings in Appendix 20.

        The posted images of all such amendment versions should be at least life-size and have high resolution, wide color range, and low contrast, in order to permit identifying, for example, a mark as a character stroke or an ink spill.

        4. I sampled at least one popular almanac and found a significant level of error, and it wasn’t alone. Among the popular versions are the fake-parchment, crinkly, brownish, single-sheet-shoehorned souvenir Constitutions often sold to tourists. One at a major public library, at a librarian’s desk, wasn’t clear enough to use for precise comparisons and lacked provenance and a printing date.

        5. I did no detailed study of electronic versions. I know no reason why they’d necessarily be more accurate or less.

        6. For this research, I defined the Bill of Rights as the sheet inscribed with the first ten amendments and what ultimately became Amendment XXVII plus one proposed amendment that was never ratified, together with all other writings on the same sheet. Those other writings presumably were ratified with, and relate to, all amendments on that sheet and state, inter alia, why they were proposed (and thus should appear where full amendment texts are professionally published, just as the preamble to the original articles is published).

        Other definitions are useful elsewhere. The most common is that it consists only of Amendments I–X; e.g., see notes in pamphlets of the Senate and the House of Representatives (pp. 20–21, both). The National Archives and the Library of Congress agree. Another definition eliminates Amendment X. And Judge Learned Hand posited it as Amendments I–VIII and the Fourteenth (see L. Hand, The Bill of Rights (1964), as quoted in Essays Commemorating The Bicentennial of The United States Constitution and the Bill of Rights: Looking Toward the Third Century, by Eric A. Klein (Lanham, Md.: Univ. Press of America, 1991), p. [1] & n. 1).

        7. See The Rise of Literacy and the Common School In the United States: A Socioeconomic Analysis to 1870, by Lee Soltow and Edward Stevens (Univ. of Chicago Press (Chicago Originals), pbk. 1981), at 50–53 (military enlistment sign-ups, assuming little disincentivized the literate into X-marking).

        8. “Perhaps 5 percent of the slaves knew how to read and write by 1860.” We Are Your Sisters: Black Women in the Nineteenth Century, ed. Dorothy Sterling (N.Y.: W. W. Norton, 1984, reissued, Norton pbk. [2d printing?] 1997), p. 44 fn.

        The percentage was likely even smaller in 1789.

        9. See Appendix 21.

        10. See Appendices 21 and 22.

        11. The Constitution of the United States of America: Analysis and Interpretation: Annotations of Cases Decided By the Supreme Court of the United States to June 29, 1992, prep. Congressional Research Service, Library of Congress (Washington, D.C.: USGPO (Sen. Doc. 103-6): 1992 ed. 1996), passim (short title: Constitution Annotated) (book challenged in this research; to be succeeded by 2002 ed., not yet published as of mid-2004).

        12. See, e.g., The Congressional Record: Fact or Fiction of the Legislative Process, by Howard N. Mantel, in 12 The Western Political Quarterly 981–995 (issue no. 4) (Dec., 1959) (JSTOR archive (stable URL http://links.jstor.org/sici?sici=0043-4078%28195912%29123A4%3C981%3ATCRFOF%3E2.0.CO%3B2-T (downloaded as 018dd5534000501279a60&backcontext=page&config=jstor&dowhat=Acrobat&0-150.pdf (PDF Economy format))), accessed Jul. 17, 2004) (author apparently affiliated with Columbia Univ., per id., at 981).

        13. Revisiting the content of the Constitution itself every twenty years was proposed around the time of its ratification.

        14. Are modern amendments inherently less prone to being rendered with errors? I haven’t drawn conclusions on that angle. The answer could have an implication for predictions on the frequency of errors in statutory renditions, since statutes now in effect are likelier to have more recent promulgations than Constitutional provisions.

        15. Everyone’s capacity for teary-eyed boredom is finite. Rotate early.

        16. One attorney, with a large metropolitan firm, did proofreading. She never mentioned any other work there. She left to run a political campaign. Her candidate won. Could her firm have been underutilizing her abilities?

        17. Const., Art. I, § 10, cl. 1, in USCS Complete Text. See Appendix 6.

        18. For example, one listener could work by telephone, so that listeners won’t interact with each other during proofreading. There may be a tendency for listeners to reinforce each other rather than for one to catch a discrepancy another misses, especially with errors of spelling, word choice, punctuation, and other relatively simple matters, which are not errors of purpose, lawfulness, and such that are relatively complex, which might more readily be found through group discussion and analysis. Apparent discrepancies reported by separate listeners should be pooled and checked regardless of how many listeners detected each one. See Applying Code Inspection to Spreadsheet Testing, by Raymond R. Panko, in 16 J. of Mgmt. Info. Systems 159 or 159 ff. (issue 2) (Fall 1999) (EBSCOhost (Business Source Premier database) (HTML Full Text), accessed Dec. 29, 2003) (code inspection of computer software comparable in part to proofreading).

        19. Normal office skills training may not reach this level of skill. See, e.g., Programmed Proofreading, by H. Frances Daniels, Thadys Johnson Dewar, & Carol W. Henson (Cincinnati, Oh.: South-Western Publishing, 3d ed. [1st printing?] 1992), esp. at 4–5 (team proofreading method, per id., at 5, described without reference to, e.g., second listener or character-by-character analysis) (authors are prof., prof. emer., & asst. prof., respectively, from two schools; book is “[exclusively endorsed]” by Professional Secretaries International (PSI) for the business communications fundamentals course in a postsecondary model curriculum).

        20. See Applying Code Inspection to Spreadsheet Testing, supra n. 18.

        21. Antiquity in a source rendition otherwise qualified can excuse an editor, especially when combined with repetition. Some renditions have long reprinting histories, possibly using the same typeset masters. Unchecked, these gain a patina of unimpeachability. Such antiquity could excuse not only low-level editors but arguably attorneys of all levels, clients, courts, legislators, and other authorities needing to refer to such renditions, but the excuse must be limited, or the effect would be to amend the Constitution by a new means, that of noncorrection of error.

        22. See Appendix 19.

        23. Making the large collection of discrepancies available was advanced by the nature of the Internet. Print publication of this research might have had substantial visibility and impact, but might have required printing only a space-restrained sampling of discrepancies, instead of all of them. That would have made it more convenient for criticized publishers to attend only to the discrepancies in the sample, and to ignore the rest, knowing that most lawyers and most courts wouldn’t see the rest, either, nor repeat the research I had done.

        24. Might a lower-level editor act in the public interest and avoid the risk to employment by sneaking in the corrections without telling a supervising editor? At the rate of one or two corrections a year, and given that some will get pulled by someone else or will be themselves wrong, this could take anywhere from twenty years to a century and a half to complete.

        Do we want to count on an institution counting on its staff to correct law texts by sneaking fixes past supervisors? I don’t think so.

        If proofreaders at law publishers are attorneys (see supra n. 16) and are interested in public service of a higher calibre than spell-checking, they may well leave for greener courtyards.

        Management will have to take the initiative to confirm its own errors, fess up to their responsibility, and apply remedies.

        25. Consider a method common in scientific research. A scientist reports a discovery, submitting an article describing it to a peer-reviewed journal. The editor, once deciding that it’s probably worth publishing, forwards it to a peer of the author, with both author and peer being somewhat anonymous to each other.

        The peer spends about an hour on the article and replies to the editor, listing specific critiques, if any. If the author thinks the peer is a moron, it’s usually permissible to request of the editor that another peer look at it. The author is usually well-advised to accept as many of the critiques as possible, but can refuse, leaving to the editor whether to publish the article. Some articles are subjected to review by several peers, as many as five.

        After publication, other scientists may attempt to replicate the discovery. For that reason, such articles are written much like recipes. If replication fails, that may be publicly reported, creating an often-justifiable demerit against the first scientist.

        Employment relationships between scientist and peer are generally out of the picture, and vendettas are harder to carry out. If a peer were to claim that the author must have a brain of curdled milk, and not list specific criticisms, the editor will have little reason to use that peer’s services again.

        This system strongly encourages scientists to be more reliable in their work and in their performance of peer review, and the result, generally, is more solid science. There are lessons here applicable to legal work.

        26. Generally, a lawyer has to proofread where concepts and legal dialect must be understood, with competence in character accuracy sufficiently available among nonlawyers.

        27. See Oxford English Dictionary Online (accessed, e.g., Jan. 2, 2003).

        Praeess also is not in the Dictionary of American Regional English (final vol. not yet seen), now the premier dialectal dictionary of the U.S. language, as a lemma.

        28. Citations supporting Constitutional quotations rarely, if ever, specify source renditions.

        29. Since the ratifications of only 9 states brought the Constitution into effect, their 9 originals are the most important to examine, followed by the remaining four of the original 13, followed by the next 37 in order of ratification.

        30. Quite a few later states had been territories for spells of decades. They probably received their official copies circa when they first established territorial governments, and may have ratified those when they became states. If several states were carved from one territory, did each carvee receive its own copy? When? Which version?

        The older the transmission of the document, roughly speaking, the less accurate it’s likely to be. Not only does technology get more capable over centuries, but human management systems, including organizational methods, scientific research standards, and historiography, get more sophisticated, too, partly because there are more people and more capital, allowing more harmonized solutions to once-baffling problems.

        More problematic might be a transmission that occurred long ago but well after the original inscribing of the Constitution, since the evolving of calligraphic customs might have meant a temporary loss of recognition of meaning when readers confronted unfamiliar papers before modern methods were available. One misunderstood stroke could swap senses.

        31. There are potentially serious issues. I, for one, on balance, am glad for the antiquity argument, where apropos.

        If antiquity doesn’t matter, then, arguably, the Constitution is invalid, because the last two of the original 13 states to ratify it may have done so under duress caused by excessive pressure from the states that had already ratified it, assuming duress is relevant to states. While by the Constitution’s terms agreement by fewer than 13 states sufficed to bring it into effect among those ratifying it, amending the Articles of Confederation required unanimity, and they remained in effect until all 13 states decided otherwise.

        Alternatively, one could argue that legislation, executive actions, and adjudications within the Federal government under the Constitution were both null and void until all 13 states had ratified the Constitution. Since legislation creating the judiciary inferior to the Supreme Court was enacted during that window, if that legislation was invalid, then the precedents emanating from it were also invalid, as were the decisions of the Supreme Court based on cases from the inferior Article III courts.

        The attorney who wants to pursue any such issue will have their work cut out and waiting. On the other hand, as my view of modern times, I don’t believe that changing the law through such a pursuit is desirable.

        32. High-register writing formerly capitalized most nouns; I don’t address whether that could mean that a noncapitalized word there was semantically weakened or was not a noun.

        33. For that word and discrepancies, see Const., Art. I, § 10, cl. 2, and appropriate appendices hereto.

        34. The Oxford English Dictionary Online, for example, is based on written records ranging beyond nineteen centuries, and includes all known spelling variants, even those long obsolete. No legal dictionary comes close. And possibly no dictionary of English dialects covers much law.

        35. See The Congressional Record: Fact or Fiction of the Legislative Process, supra n. 12, at 992 & esp. n. 46, in which the author (in 1959) suggested a difficulty if the judiciary were to order the executive or the legislature to prove what legal text it had approved or enacted, respectively; but I question whether such a difficulty, grounded in the separation of the three Federal branches, would apply with respect to a private party engaged in publishing the same, especially the Constitution, which is an enactment jointly of states and Congress (absent any convention), and at least part of one copy of which is already publicly displayed.

        Even were that subpoena power limited, voluntary disclosure by a branch, a state, or the custodian of convention or other records might suffice to shift or reinforce a burden of proof of accuracy, and there would seem little reason beyond inconvenience to resist voluntary disclosure (the real risk to physical integrity of original documents, especially older ones, being accommodatable by allowing such disclosure to be occasional and by allowing especially high-quality photographs to be official records as prima facie evidence of their subjects), not to mention Constitutional issues in concealing law that purports to bind an uninformed party, a party being uninformed if previous publication has failed to be accurate.

        Publication may create prima facie evidence of its content, but it cannot be unrebuttable, as that would constitute a usurpation of the lawmaking power vested, e.g., in the legislature, the executive, and the judiciary by the Constitution in the Federal government and in the republican government Constitutionally guaranteed for each state. Doubtless many persons would like to assume a power to write laws binding others if all that were required to put the same into effect is paper, ink, and publicity.

        Publishers public and private who publish texts that do not purport to be only interpretive of binding texts and with invitations that they be relied upon in determining precise civil and criminal obligations and rights, invitations that have been accepted often with considerations paid over, would seem to have a duty of faithful accuracy in the rendering of those legal texts, and thus a duty proactively to find their errors and to correct them promptly and with specific publicity. Specific publicity is necessitated by the likely continuing circulation of older renditions that have long enjoyed trust, along with older opinions, treatises, and other derivative documents also long trusted, in order that the trust be appropriately modified among wide audiences.

        36. Non-Federal statute is probably rendered worse on average than Federal statute, and the latter is demonstrably not always fault-free. E.g., a modern omission from a Federal statute due to a clerical error was reported by the President as late as 1986. See his statement, 100 Stat. 3341-388 n., on his approval of Pub. L. 99-591; cf. 100 Stat. 1783 n. (re Pub. L. 99-500, also approved). The earlier-approved enactment appears in 100 Stat. as “a subsequently typeset print”; subsequent to what, given that its mention was presumptively warranted, is unclear. The page counts in 100 Stat. appear to differ by three and the President said that “a small number of paragraphs of text were omitted due to clerical error” (100 Stat., at 3341-389), but that is not enough to determine exactly which provisions were promulgated only when the error was substantively corrected. No further clarification was found on the corresponding pages of 1986 U.S. Code Cong. & Admin. News. Unless someone has already performed that research and it has been published, given that the mentioned error may not be exclusive and, e.g., a punctuation mark may also have been erroneous, the research’s scale could be daunting.

        37. One distinction among subpopulations could be according to the probability of being composed of entities likely already to have full notice, such as major institutions, assuming the law doesn’t create collateral effects for others lacking that notice.

        If that explains the noncodification of one Federal statute applicable by its terms to a certain Indian tribe, a question of quality of notice arises from its collateral applicability to some non-Indians and perhaps other tribes.

        38. Rather than print provisions that apply only to small subpopulations, publishers can correlate by relevance code sections and uncodified provisions available in enactment series. Indexing in code section notes and in a subject index would ease research, and ultimately might aid legislative transparency.

        39. Among attorneys who should undertake this are those who are idle because their firms underwork them or who are developing new niches, especially a niche where the legal framework has long been in place and law has been much interpreted. That a body of law may have long settled down may mean little more than institutional lethargy.

        40. Those who practice largely within a regulations-based framework should examine its purportedly authorizing statutes, while those who practice in a specialty largely within agency directives should examine their purportedly authorizing regulations for inconsistency with superior authority, lack of authority, and other defects. Challenges on point have been successfully adjudicated despite years of administrative habit.

        41. Odd is the exclusion of the Constitution from U.S.C.A.’s Organic Laws, contrary to the same publisher’s Black’s Law Dictionary (7th ed.), at its entry for organic law, and Corpus Juris Secundum, 16 C.J.S. Constitutional Law § 2 subsec. a, and contrary to the classification in the U.S. Code. By its own terms, the Constitution qualifies as organic. The exclusion in U.S.C.A. should be reversed by the publisher.

URL: http://www.GeoCities.com/Nick_Levinson/law/const_proofrdg/article.htm
Article first publicly posted to the World Wide Web July 24, 2004;
        last revised (to add the link at the top) and posted August 25, 2004.
Article and all appendices:
        Revisions (in the case of one or more appendices, minor) and, in article, all HTML coding:
                Copyright © 2004 Nick Levinson
        Balance: Copyright © 2003 Nick Levinson
        All rights reserved.