Posts Tagged ‘unpublished’

One U.S. District Court’s Lonely Gesture Toward Open Access and Medium-Neutral Citation

Tuesday, January 24th, 2017

I. Introduction

As 2017 opens one U.S. District Court – that for the District of New Hampshire – begins its eighteenth year as an isolated (and incomplete) model of how all federal courts might handle opinion distribution.  (Hat tip to Andrew P. Thornton of Little Rock for bringing its record to my attention.)

II. The Simple Steps this One Court Has Taken

In January 2000, the U.S. District Court in New Hampshire started identifying some of its decisions by year, numbering them sequentially.  It designated Silva v. Nat’l Telewire Corp., No.  99-219-JD, decided on January 3, for example, as “Opinion No. 2000 DNH 001“.  Panza v. Grappone Cos., No. 99-221-M, decided on October 20 of the same year, is “Opinion No. 2000 DNH 224“.  Immediately, upon release, the decision texts, carrying these identifiers, were placed in a court-hosted, searchable database.

The following year the court adopted a local “citation format” rule.  That rule directs those citing decisions released after January 1, 2000 and published at the court site to do so “using the four-digit year in which the opinion is issued, the letters ‘DNH,’ [and] the three-digit opinion number located below the docket number on the right side of the case caption ….”  For decisions published in “the Federal Supplement, the Federal Rules Service, or the Federal Rules Decisions” the rule authorizes volume and page number citations to those print reporters as an alternative.

This took place well before the E-Government Act of 2002 called upon federal courts to provide web-access to “all written opinions.”  While this island of non-print-based citation has escaped the notice of The Bluebook, the 2001 local rule remains in effect and the practice continues.  McFadden v. Walmart, 2017 DNH 002, was decided on January 5 of this year.  The district’s judges themselves do still, on occasion, cite using opinion numbers.  See, e.g., Hersey v. Colvin, 2016 DNH 203, 10 (citing  Corson v. Soc. Sec’y Admin., Comm’r, 2013 DNH 144, 24–25).  So do  attorneys.  The New Hampshire Bar Association publishes a monthly “US District Court Decision Listing” that contains summaries of selected decisions of the prior month.  The decisions covered are cited by their “medium neutral” or “public domain” opinion numbers.

Since the court-attached opinion numbers appear within the texts they identify, researchers need no other citation to retrieve a decision from any electronic source.  They do the job on Bloomberg Law, Casetext, Google Scholar, Lexis Advance, Ravel Law, and WestlawNext.  They also work with the GPO’s FDsys case law repository (about which more below).  For the same reason these sources also provide the opinion number required for a conforming District of New Hampshire citation.

III. Critical Respects in Which the Model Falls Short

A. The Use of Pagination as the Means for Pinpoint Citation

Although nearly all legal research services retain the opinion numbers attached by the U.S. District Court for New Hampshire, only Casetext, Fdsys, and the court’s own database preserve the location of the page breaks in the original version of a decision that the court’s rule directs be used for pinpoint citations. Arkansas and Louisiana, two state systems that, similarly, adopted neutral citation but sought to avoid paragraph numbering by specifying the pagination in a court-released pdf file as the basis for pinpoint references, have suffered the same fate in research services that, like Google Scholar, base their texts for many jurisdictions on the versions published in the Thomson Reuters National Reporter System.  Not only are paragraph numbers more precise and more tightly connected to the logical structure of a cited document than pagination, they travel far more reliably with the portions of text they denote into the full range of data services used by those doing legal research.

B. A Failure to Include All Substantive Opinions (Including Magistrates’ Reports and Recommendations)

Not all decisions rendered by District of New Hampshire judges receive court-applied opinion numbers, only selected ones.  In compliance with the E-Government Act of 2002 all written opinions of the court, including reports and recommendations by magistrate judges, are made available without charge through the PACER system, where they can be gathered by the online services.  A non-trivial number of those opinions – ten percent or more – have not been given opinion numbers nor placed in the court’s searchable database.  That is particularly true with categories of cases such as inmate suits and Social Security appeals that are routinely resolved by a magistrate’s report and recommendation, followed by a short judicial order adopting it.  As a result, a significant body of district case law cannot be found in the court’s searchable database nor cited by means of opinion numbers.  Because of this incompleteness, responsible case law research cannot be carried out using the court’s database.  Thoroughness requires use of one of the comprehensive research services.   And that leads to citations by the court of its own prior decisions that employ Westlaw or Lexis proprietary cites rather than, or in parallel with, the court’s public domain, medium neutral scheme.

C. Inherent Limits on a Single-District Citation System within a Federal Court with 93 Other Districts

The situation in the District of New Hampshire is categorically different from that in the numerous states that have adopted similar plans of electronic publication and court-applied citation.  Matters litigated in state court can often be argued and decided solely on the basis of that state’s own case law.  By contrast, rarely if ever can those representing parties to a matter before the U.S. District Court for the District of New Hampshire or the judge handling the case disregard decisions from the First Circuit and other U.S. Courts of Appeals and decisions by other district courts as well.  For the district judge that calls for use of one of the two commercial systems available to the federal judiciary; for attorneys, use of those same systems or some comparably comprehensive alternative.  The court’s less-than-complete database of decisions may, conceivably, be a useful place to start research but never a place to finish it.  Thorough research and consistent citations of relevant decisions lead almost inexorably to the use of one or more of the proprietary systems.  With this district’s judges the dominant system is Westlaw.  Their pinpoint cites to unpublished decisions, including those in citations to cases that have court-applied opinion numbers, overwhelmingly use Westlaw pagination instead of the page numbers contained in the court’s original version.  The citations to Mudge v. Bank of Am., N.A.Gasparik v. Fed. Nat’l Mortg. Ass’n, and Dionne v. Fed. Nat’l Mortg. Ass’n in LaFratta v. Select Portfolio Servicing, Inc., 2017 DNH 007, as released by the court, are examples.  LaFratta and other recent decisions reveal a declining use of the court’s opinion numbers and a growing practice of linking citations to authority of all kinds into Westlaw.

IV. The Sorry Fate of Other Single-Court Citation Schemes within the Federal Judiciary

A. The Sixth Circuit’s Ancient DOS-Based Naming Scheme

Since 1994 decisions of the U.S. Court of Appeals for the Sixth Circuit, both published and unpublished, have carried a “file name” identifier.  Designed to fit within the name space of the MS-DOS operating system of that era those identifiers consist of eight characters, followed by a period, followed by two more.  The file name of one unpublished decision released in January 2016 is “16a0051n.06”.  Miller v. Comm’r of Soc. Sec., 811 F.3d 825 (6th Cir. 2016) decided the same month is: “16a0020p.06”.  (The “n” and “p” indicate whether the decision is to be published or not.)  While Lexis retains these identifiers, they don’t follow opinions into volumes of F.3d or Westlaw.  As seems gradually to be happening with the District of New Hampshire opinion numbers, the Sixth Circuit file names have become useless data.

B. The Relatively Brief Run of Neutral Citation in the District of South Dakota

Effective January 1, 1996, the Supreme Court of South Dakota began attaching medium neutral citations and paragraph numbers to its opinions.  The practice continues; the court’s rules of appellate procedure still require use of this public domain citation system.  Later in that year, by local rule the U.S. District Court for the District of South Dakota laid down the same steps.  Even at the time not all of the district’s judges bought into the change.  With the appointment of a new chief judge in 1999 who was not an enthusiast, the system continued in the opinions of only one of three active district judges and a magistrate judge.  When the district judge in question took senior status in 2008, all trace of the scheme disappeared.

V. Missed Opportunities to Implement Non-Print-Based, Non-Proprietary Citation across the Federal Courts

A. The Judicial Conference Response to the 1996 ABA Resolution

In 1996 the American Bar Association House of Delegates recommended that all U.S. jurisdictions “adopt a system for official citation to case reports that is equally effective for printed case reports and for case reports electronically published.”  The resolution proceeded to spell out the key elements of such a system: 1) attachment of identifiers to all decisions, consisting of the year, the court, and a sequential decision number, 2) insertion of paragraph numbers, and 3) adoption of court rules requiring that citations employ these elements.  In response the Automation Committee of the Judicial Conference of the United States and the Administrative Office of the Courts simply surveyed federal judges and clerks regarding the ABA citation recommendation.  Without asking the Federal Judicial Center for a study or furnishing rationale or context, it simply asked all these individual actors whether they favored the steps.  Overwhelmingly they expressed satisfaction with the status quo, hostility to paragraph numbering, and puzzlement over the grounds for change.  The recommendation died in committee and has not since been revived.

B. Terms of the E-Government Act’s Mandate

The E-Government Act of 2002, in a section immediately prior to the one addressing the federal courts,  directed the creation of and authorized appropriations for an integrated online information system covering all federal administrative agencies.  That portal was to be designed to allow public access to agency material “integrated according to function or topic rather than separated according to the boundaries of agency jurisdiction.”  In contrast, reflecting the highly decentralized administrative structure of the federal courts, the act’s directive that all federal court opinions be made accessible online was directed at the chief judge or justice of each and every court in the federal system.  A more coordinated approach might have drawn attention to the citation issue.

C. Addition of Rule 32.1 to the Federal Rules of Appellate Procedure

Similarly, the reform movement that led to the addition of Rule 32.1 of the Federal Rules of Appellate Procedure in 2006 might well have focused attention on how the “unpublished” decisions of the U.S. Courts of Appeals, which by the terms of the new rule became citable, could or should be cited.  Its sponsor, the Advisory Committee on the Federal Rules of Appellate Procedure, heard concerns about how those for whom Lexis and Westlaw were beyond reach would access to this large body of case law.  Ignoring the citation challenge the committee pointed to the E-Government Act’s mandate as addressing the problem.

The strategic appearance of the West Federal Appendix in 2001, which furnished the means for proprietary volume and page number citation for these “unpublished” decisions to members of the federal judiciary (all of whom have access to Westlaw) almost certainly encouraged this blindness.

D. Implementation of the Federal CM/ECF System, its PACER overlay, and the Fdsys Decision Archive

Federal court electronic case management systems trace all the way back to applications developed by the Federal Judicial Center in the late 1960s.  Those established the fundamental structural model that persists to this day: central development of a set of electronic tools, with most decisions about whether, when, or how to use them left to the individual courts.  It is probably significant that, having its own administrative and technical support, the U.S. Supreme Court has taken no part in promoting or coordinating technology adoption in the subordinate federal courts.  In 1990 Congress catalyzed the opening of existing court-located case and document management systems for remote electronic access.  At the time that meant dial-up.  The move to electronic filing began in 1995.  At around the same time the Administrative Office of the U.S. Courts began work on a national party and case number index to the electronic records of the federal courts that had implemented its CM/ECF system.  For many federal courts this Public Access to Court Electronic Records service (PACER) subsequently became the mechanism for compliance with the E-Government Act’s mandate.  While access to other documents through PACER carries a fee, all documents tagged by the deciding court as “opinions” can be retrieved without charge.  However, PACER provides no full-text index of those opinions.  They can only be tracked down using docket number, party names, court, and case type.

As filed in a court’s CM/ECF system an opinion is stamped with identifiers that consist solely of case docket number, filing date, and the document’s place in the sequence of filings in the matter – “Case 1:15-cv-00200-LM Document 5 Filed 11/17/15” for example.  A uniform federal court citation system could have been appended to this system, either initially or in the “next generation” version now being rolled out.  It was not.

In recent years the Government Printing Office Federal Digital System (FDsys) has begun drawing opinions from participating federal courts and loading them into a text-searchable database.  Following a pilot phase, the Judicial Conference of the United States authorized national implementation of this inter-branch cooperative venture in September 2012.  Over four years later, it remains seriously incomplete in scope; only 49 out of 94 districts courts are included.  Furthermore, among included courts, the chronological depth and currency of the data vary considerably.  And while GPO authenticates each PDF file it receives from a participating court system and associates a useful array of metadata with it, it has not, as it could, attached an identifier that a lawyer or judge would recognize as a citation.  To date, this is simply another more missed opportunity.

At the beginning of 2017, the prospects of a system-wide citation scheme modeled on that launched in New Hampshire at the turn of the century appear dim.

VI. How Should Decisions of the U.S. District Court for the District of New Hampshire Be Cited?

As noted above, this one district court still attaches medium-neutral citations to many, although not all, of its decisions.  Whether one obtains such a decision from the court’s database or a commercial source, its opinion number is available and, when included in a citation, it furnishes a highly efficient retrieval identifier.  Decisions that have been given a place in F. Supp. or F.R.D. can be retrieved by volume and page number from nearly all research services.  Adding the opinion number as a parallel adds negligible value.  For “unpublished decisions” whether or not, given an opinion number, Westlaw or Lexis citations may suffice for the court, its judges having access to both.  But limiting a citation to one or the other or even both in parallel may leave the opposing party and others who might rely on Google Scholar or Casetext or Ravel without an efficient retrieval hook.  Pincites pose a further problem.  Lexis includes Westlaw cites for unpublished cases but not Westlaw pagination.  Westlaw ignores both Lexis cites and Lexis pagination.

Useful guidance and models come from the court’s own decisions.  In Bersaw v. Northland Group Inc., 2015 DNH 050, Judge Joseph LaPlante offered this advice: “[I] would recommend that, with respect to unpublished cases that appear solely on electronic databases such as Westlaw or Lexis, counsel provide as much alternative identifying information (e.g., case number, issuing court, and opinion date) as possible.”  The judge, himself, practices what he recommends.  A citation appearing in Locke v. Colvin, furnishes a fully fleshed out example of this approach.  It reads:

Brindley v. Colvin, No. 14-cv-548-PB, 2016 U.S. Dist. LEXIS 10757, 2016 WL 355477, at *5 (D.N.H. Jan. 29, 2016) (quoting Ortiz, 890 F.2d at 528) (remanding where ALJ neither called vocational expert nor explained why reliance upon the Grid was appropriate, but “merely stated, without explanation or citation to record evidence, that [the claimant’s] non-exertional limitations have little or no effect on the occupational base of unskilled light work”) (internal quotation marks and citation to the record omitted).

Four aspects of the example warrant notice:

  • While Brindley v. Colvin has an opinion number (2016 DNH 021) it is not included.
  • Westlaw pagination rather than pagination from the version held in the court’s database provides the pinpoint reference.
  • The addition of docket or case number and full date, as counseled by Judge LaPlante, make it possible to retrieve the Brindley decision from sources that hold it but neither its Westlaw or Lexis citation, including the court’s own database.
  • The parenthetical notes provide a clear path to the cited portion of Brindley for any reader who is inspecting that decision on a system in which having the Westlaw star page number is useless.

California (Finally) Ends Automatic Depublication

Monday, June 20th, 2016

California’s intermediate appellate courts, the Courts of Appeal, produce approximately ten thousand written opinions each year. Fewer than one in ten are published. In most cases, the decision to publish or not is made by the deciding court applying criteria set out in Cal. Rules of Court 8.1105(c). Except where res judicata or related doctrines are involved, opinions that are not certified for publication may not be cited or relied upon by “a court or a party in any other [California] action.” Cal. Rules of Court 8.1115(a).  While the deciding court makes the initial call, the California Supreme Court can “depublish” an opinion even as it lets the lower court’s disposition of the case stand. Cal. Rules of Court 8.1105(e). During 2015 the court did so in a dozen cases. (It can also direct that a Court of Appeal decision be published, but that is a rare occurrence.)

In a year’s time the California Supreme Court receives nearly eight thousand petitions for review, agreeing to hear less than ten percent.  Prior to a rule change that takes effect on July 1, 2016, the high court’s decision to take a case automatically placed the opinion being appealed in the “unpublished” category.  Of course, in the modern era, this did not prevented the circulation of the previously “published” decision in print or online.  Indeed, all “unpublished” opinions of the Courts of Appeal are released to the public at a judicial branch website. But automatic depublication blocked citation of it and any subsequent judicial reliance.

This unique rule dates from a time when the California Supreme Court reviewed trial court decisions de novo, so that its agreeing to hear a case effectively nullified the prior opinion of the intermediate appellate court in the matter. A 1984 constitutional amendment altered that framework. Bar groups and judges urged that the depublication rule be revisited, but without success. Three decades later the California Supreme Court released a set of proposed amendments for public comment. With some modification those changes were adopted in June 2016, effective July 1.

After that date a grant of review by the California Supreme Court will no longer automatically remove “published” status from a Court of Appeal opinion. Under the revised rule, the Supreme Court can take that step but only upon an affirmative decision to do so. Even with that change, a grant of review does automatically affect the weight to be given the opinion by other California courts. Pending resolution of the appeal, the Court of Appeal opinion “has no binding or precedential effect.” It may be cited but only for its “potentially persuasive value.”

Chalk this up as a very modest reform. As Professor David Cleveland reports in the most recent issue of The Journal of Appellate Practice and Process, the last decade has seen a significant and steady shift in state rules governing “unpublished” or “non-precedential” decisions. His article counts seven states as having moved to permit citation of unpublished decisions, one as going the further step of granting them precedential weight, and five as having eliminated the “unpublished” category altogether. California’s change comes nowhere near such measures or even the situation in the federal courts under Rule 32.1 of the Federal Rules of Appellate Procedure. Perhaps, in another thirty years?




The Complex Relationship between Citations and Citators

Wednesday, April 29th, 2015

Shepard’s Citations

In 1873, Frank Shepard began compiling and selling lists of citations to Illinois decisions printed on gummed paper (Shepard’s System of Adhesive Citations).  Purchasers pasted them into the margins of their bound case reports.  Shepard’s lists linked each reported case to any subsequent reported decision that referred to it.  When gummed addenda proved too cumbersome a tool (even more troublesome to maintain than looseleaf volumes), Shepard’s Citations moved to separate volumes.  These were books of citations designed to stand beside law reports – volumes that simply pointed from one book to others by means of citation.


For over a century law students, lawyers, and judges conducted forward citation searches on key decisions using the Shepard’s publications.  So tight was the association that the process became known as “Shepardizing”.  One “Shepardized” a case to assure it had not be overruled by a higher court, to determine its status and range of interpretation within the jurisdiction of origin, to see how it had been treated elsewhere.

Cases and Citators Go Digital

Once electronic databases were central to case research, their incorporation of a citator function became essential.  The value of providing the digital equivalent of Shepard’s gummed list proximate to every retrieved opinion was obvious. And in a hypertext environment that list of citing cases could itself offer point and click access to each one of them.  Moreover, once held in a database the entries could be filtered and sorted.  Today, all case law database services of professional quality offer retrieval of subsequent citing cases as an option adjacent to each opinion.  Some not only list the citing cases but analyze and characterize those references as the Shepard’s print publications once did.

As electronic case law collections evolved, however, they posed fresh challenges for these companion citators.  Increasingly the leading online databases added decisions that the Shepard’s lists had ignored, cases without standard print citations.  These included opinions that would never be published in print, either because of court designation or publisher discretion, as well as “slip” versions of those whose publication was anticipated but had not yet occurred.  Generally unexamined is the extent to which the relative performance of today’s online citators is affected by how they deal with citations in and citations to opinions falling in these two categories.  That performance varies considerably.  Researchers who assume complete results are, with some services, likely to miss important cases.  Those who know the limitations of the citator on which they rely can, when necessary, augment its results with their own database search.

The Citator Challenges Posed by Unpublished Decisions

Citations to Not Yet Published Decisions

Because of their high volume Social Security cases provide a particularly clear illustration of the problem posed by the delayed application of citation parameters and the range of responses to it by the citators now embedded in the major online services.  As of April 23 five “precedential” decisions in cases appealing a denial of benefits by the Social Security Administration had been released by the Seventh Circuit U.S. Court of Appeals since the beginning of 2015.  (Decisions the Court does not deem significant to other cases it labels “Nonprecedential” and withholds from publication in the Thomson Reuters Federal Reporter series.)  “Four of the five were written by Judge Richard Posner.  Three of his decisions and one by Judge Daniel Manion reversed trial court decisions that had affirmed the agency’s benefit denial.

From the moment of release, the potential ripple effect of opinions like these is substantial, throughout the district courts falling within the Seventh Circuit and beyond.  Consider the numbers.  During the twelve months ending June 30, 2014, those districts received 1,441 Social Security appeals.  Within weeks, in some cases days, the five 2015 Court of Appeals decisions were being cited.  Curvin v. Colvin, No. 13-3622 (7th Cir. Feb. 11, 2015), the earliest of the set, has now been cited at least 12 times.  (A pro-claimant Social Security decision of the Seventh Circuit handed down a little over a year ago  – Moore v. Colvin, 743 F.3d 1118 (7th Cir. 2014) –  has been cited over 125 times, at least twice outside the circuit.)

Curvin illustrates the difficulty faced by anyone or any system attempting to track these citing references.  The decision was handed down on February 11, 2015 but did not receive its “778 F.3d 645” designation until a month and a half later.  During the intervening weeks it was cited at least eight times by district courts within the Seventh Circuit.  Perforce those citations identified the Seventh Circuit opinion by docket number and exact date or a proprietary database citation (“WL”).  Most, but not all, used both in parallel, yielding citations in the following form: Curvin v. Colvin, No. 13-3622, 2015 WL 542847 (7th Cir. Feb. 11, 2015).  A straight database search on “778 F.3d 645” will not retrieve those cases.  A database search on “2015 WL 542847” will retrieve those using the Westlaw cite (but not those employing the LEXIS equivalent “2015 U.S. App. LEXIS 2170” or the “F.3d” cite).  A search on “13-3622” and “Curvin” will retrieve those including Curvin’s docket number but not those relying solely on a proprietary database cite or the ultimate “F.3d” cite.

Most case law databases purport to do this messy work for the researcher.  With some Curvin’s rank in a set of search results may even be determined by how many citations to it there have been.  What not all manage to do is to include those instances of citation that occurred so soon after Curvin’s release they could not refer to the case as “778 F.3d 645”.  A review of how the major systems actually address this issue (or don’t) follows.


The dominance of Westlaw within the federal judiciary gives that system a clear advantage.  So long as the early decisions cite the not-yet-published version of a case using its “WL” citation, Westlaw can employ that identifier to link them with those citing to the version later published in the company’s National Reporter System (NRS).  But what about decisions written by  federal judges who use LexisNexis and cite using its proprietary system?  Senior Judge Donetta W. Ambrose of the Western District of Pennsylvania falls in this category.  Had she relied on Curvin in late February or early March 2015, her opinion would almost certainly have cited it: Curvin v. Colvin, 2015 U.S. App. LEXIS 2170 (7th Cir. 2015).  (See, for example, her decision in Nickens v. Colvin.)  How would Westlaw have responded?  It would have added a parallel “2015 WL 542847” to her Lexis cite, as it does to all opinion citations to “not yet published” or “never to be published” cases contained in the Westlaw database.  That editorial step simplifies aggregation of all citations to a case prior its print publication.  While Westlaw no longer displays the “WL” cite for decisions that have been given print citations in the National Reporter System, the service’s citation listings rest on its maintaining the association between preliminary “WL” cites and their subsequent NRS equivalents.  This approach enables Westlaw’s listing of cases citing Curvin to include the early ones that did not use its F.3d volume and page number.



Lexis follows a similar strategy.  Since most federal judges use Westlaw most of the early decisions citing Curvin used its Westlaw cite.  See, e.g., Haire v. Colvin, No. 1:14-CV-00322-TAB-JMS (S.D. Ind. Feb. 20, 2015).  On Lexis the cite to Curvin in Haire includes an added “U.S. App. LEXIS” cite.  That enables the inclusion of Haire in the service’s dynamically generated list of decisions citing Curvin.  It also facilitates another Lexis practice, the subsequent addition of parallel “F.3d” cites to decisions that did not, as written, include them.


Bloomberg Law

Bloomberg has a “BL” citing scheme which it now deploys much like the Lexis cites, but with greater clarity.  When a case in its database is cited by a later decision using only docket number and date or a Westlaw or Lexis cite, Bloomberg inserts a parallel “BL” cite.  This editorial addition is, however, placed in square brackets, an acknowledgment that it was not part of the original text.  Bloomberg Law has expanded Haire’s cite to Curvin written by the court as “Curvin v. Colvin, No. 13-3622, 2015 WL 542847, at *4, — F.3d —- (7th Cir. Feb. 11, 2015)” to “Curvin v. Colvin, No. 13-3622, [2015 BL 34654], 2015 U.S. App. LEXIS 2170 , 2015 WL 542847 , at *4, ___ F.3d ___ (7th Cir. Feb. 11, 2015)”.  This practice appears relatively new.  Decisions of an earlier vintage Bloomberg loaded as received without adding “BL” parallel cites.  As a result decisions from that period are missed by Bloomberg’s linked retrieval of citing documents.  (The fact that Bloomberg’s versions of decisions now also include the Lexis cite, without the square brackets, suggests a data sharing arrangement between the two companies.)


Judging at least from this sample of one, Bloomberg appears to add cases more rapidly than either Westlaw or Lexis.  During the week of April 20th two more district court decisions citing Curvin were released.  Both were in the Bloomberg database and listed as citing cases the following day.

The More Limited Approach of Google Scholar, Fastcase, and Casemaker

Google Scholar does not to attempt to track citing references for cases until they have received a permanent citation in the Thomson Reuters books.  To date it does not have the NRS version of Curvin.  When one clicks on the “How cited” link for the “slip” version of the  case, one gets the message: “We could not determine how this case has been cited.”  To find those cases a researcher must know to search on the party names and Curvin’s docket number or, alternatively, on its proprietary cites.  The latter, of course, do not appear on Google Scholar or the public domain version of Curvin released by the Seventh Circuit and now (and forever?) available from the GPO’s Federal Digital System (FDsys).  At some point Scholar will replace the original version of Curvin with that published by Thomson Reuters.  Once it has, the decision’s “How cited” link will work, but it will not retrieve the early cases which did not cite Curvin by volume and page number because they could not.  Researchers who know that can augment Google’s automatically generated list by doing the sort of searches suggested above.

Like Google Scholar both Casemaker and Fastcase limit their retrieval of citing cases to those that cite by means of NRS volume and page number, thereby missing the earliest references.  Leavitt v. Cohen, No. 1:12-cv-1427-DKL-JMS (S.D. Ind. March 4, 2014) cited Moore v. Colvin, 743 F.3d 1118 (7th Cir. 2014), released less than a week before, using the format: Moore v. Colvin, ___ F.3d ___, 2014 WL 763223, *1(7th Cir. 2014).  Since neither Fastcase nor Casemaker later fill in such blank “F.3d” citations or employ an enduring identifier for Moore (like the proprietary citation schemes of Bloomberg, Lexis, and Westlaw) neither includes Leavitt as a case citing Moore as those services do.

What about Newcomers like Ravel Law and Casetext?

Casetext does not yet have a fully developed method of indexing citing cases.  It is designed to allow the ranking of search results by “Cite count” but while its database includes many more it lists only two cases as citing Moore.

Ravel has stronger incentive to solve the citator problem because its visualization of search results derives in significant part from citation links.  However, to date Ravel’s cite count does not include case citations that pre-date the availability of the canonical NRS volume and page cite for a case.  It counts only 70 cases as citing Moore v. Colvin.  Those in its database not using that decision’s full “F.3d” cite do not make the list.

Citators and Never-to-be-Published Decisions

A 2013 “unpublished” Social Security decision of the Ninth Circuit illuminates this closely related citator issue. In Farias v. Colvin, No. 11-57088 (9th Cir. May 20, 2013), the court reversed a district court decision that had affirmed a denial of disability benefits.  Its memorandum opinion faulted the Administrative Law Judge’s uncritical acceptance of testimony from a vocational expert.  Being an unpublished memorandum opinion the Farias decision does not enjoy the status of precedent even within the courts that comprise the Ninth Circuit. Print-based Shepard’s would have ignored it.

On the other hand, unpublished decisions like Farias can be cited by counsel as persuasive authority.  In fact, at least fifteen subsequent (unpublished) district court decisions refer to the Farias case.  Because of the Thomson Reuters Federal Appendix reporter, Farias did in fact receive a print citation before 2013 was over, notwithstanding its “unpublished” designation, but not before being cited in at least two district court decisions.  Thus, in one sense cases like it pose the same problem for citation compilers as those posed by cases eventually published in the Federal Reporter – a need to gather the earliest citations together with later ones expressed in terms of print volume and page numbers.  However, the decision’s “unpublished” status and the dubious value of “Fed. Appx.” cites has led some case law services to stumble over providing useful citator results.  The major three –Bloomberg, Lexis, and Westlaw – use their respective systems of proprietary citation to link Farias to the full spectrum of citing district court decisions.  In contrast users of Google Scholar, Casemaker, and Fastcase are led to believe that Farias has not been cited unless they know enough to undertake a forward citation search on their own.  And because some of the citing cases use the Farias decision’s “Fed. Appx.” cite and others don’t, some include the case docket number but most don’t, some use a proprietary database citation and others not, no single search other than one based simply on the case name (“Farias v. Colvin”) will retrieve them all.

One More Argument for Adoption of Court-Applied Systems of Citation

In jurisdictions that attach official citations to decisions at the time of release there is little difficulty generating a complete list of subsequent citing cases.  Assuming that the court-attached citations are routinely used (whether or not in parallel with the National Reporter System or any other citation) a simple database search will retrieve all citing references.  In 1999 the Oklahoma Supreme Court decided an influential case dealing with attorney malpractice liability.  When released it carried the designation “1999 OK 79”.  A search on that string, whether carried out directly by a researcher or automatically by software generating a citator list, should gather a comprehensive list of references to Manley v. Brown.  That fact has enabled the Oklahoma State Courts Network database to append a list of citing cases to the decision in Manley.  Although the case appears in the National Reporter System as “989 P.2d 448” a researcher or automated citator searching cases for references to Manley will not be thrown off by use of that print reference so long as it appears in parallel with the court-attached cite, as it does in all Oklahoma decisions and in a 2013 decision of the Illinois Appellate Court.  Any citation search that relies solely on NRS citations for Oklahoma cases runs the risk of missing some.

What does the start of a new year mean in legal citation?

Tuesday, January 7th, 2014

A year change prompts reflection on the roles dates play in legal citation. I use the plural “roles” because of the diversity of functions dates serve in citations.  With some sources they are largely superfluous; with others, they are critical to retrieval.


As noted in a prior post, the full date of release is a crucial part of the identifying data for any unpublished opinion. Among the decisions released this week by the Second Circuit is one that should now and into the future be cited as: Wager v. Littell, No. 13-1683-cv (2d Cir. Jan. 6, 2014).

In all U.S. jurisdictions that have adopted systems of medium neutral case citation, the year of a decision is an integral part of its cite or retrieval tag. The decision, People v. Radojcic, 2013 IL 114197, would have been designated 2014 IL 114197, had it been released by the Illinois Supreme Court during the early part of 2014 rather than on Nov. 21, 2013. On Dec. 23, 2013, the Colorado Supreme Court released its opinion in People v. Cunningham, 2013 CO 71. Had it instead been the court’s first decision of this month it would have been 2014 CO 1. (Illinois and Colorado employ different systems of designating decisions rendered within the year.)

In a majority of U.S. jurisdictions, however, the year of a precedential decision is not critical for identification or retrieval. Nonetheless, it is routinely included as one element of a complete case citation. Thus, when a 2013 decision of the U.S. Supreme Court cites an earlier decision, that decision is identified as Trainor v. Hernandez, 431 U. S. 434, 443 (1977). Why include the year? “431 U.S. 434” provides all the information one needs to retrieve that earlier decision from any database or library shelf? The reason presumably is that knowing the year of a decision may help a reader to decide whether to look at it. It provides useful but not critical information.

Finally, it may be worth noting that the year incorporated within a decision citation is not the year that the case was compiled into a print volume or that the volume was finally published but rather the year the decision was issued.


Citations to session laws generally include the year of enactment. Indeed, the year is often part of an act’s name. If the legislation has not been named, its full date will be employed for that purpose, as, for example, “Act of Dec. 9, 2013”.

What to do when citing to a section of a jurisdiction’s codified laws is bit of a puzzle. Should a year be furnished and, if so, what year? Now that the year is 2014 has section 110 of title 17 of the United States Code become 17 U.S.C. § 110 (2014)? Is that a function of the cutoff date of the writer’s source?

The most recent print version of 17 U.S.C. § 110 published by the Government Printing Office appears in a set denominated the 2012 edition. However, since that edition extends through the term of the 112th Congress it, in fact, includes laws passed and signed into law in the early days of 2013. The volume in which 17 U.S.C. § 110 appears was printed in 2013. Other volumes of the 2012 edition have yet to appear.  Westlaw doesn’t provide an “as of” date for this or other sections of the U.S. Code but it does report that the most recent amendment of this particular section took effect on April 27, 2005. Lexis represents that its version of the U.S. Code is “Current through PL 113-57, approved 12/9/13.” The LII notes of its version “Current through Pub. L. 113-52” without providing a date. However one interprets of The Bluebook’s prescription on this point, it definitely calls for some date to be appended to 17 U.S.C. § 110, in parentheses.

The more sensible approach, at least in legal writing produced by or for courts, is that followed by the U.S. Supreme Court. So long as an opinion of the Court is referring to sections of the code currently in effect, its citations include no date element. The lower federal courts follow the same practice as do most lawyers submitting briefs to federal courts. One also finds dateless statutory citations in the appellate decisions and briefs from such prominent states as California, Massachusetts, New York, Pennsylvania, and Texas.

Only when the provision being cited has, by the time of writing, been repealed or amended or has only recently been enacted does it become important to specify the date of a compilation that contains the language being cited. The precise form this takes will necessarily be governed by the form in which that compilation presents its cutoff date, and it ought to report the compilation date not the year that compilation appeared in print or online.


The considerations bearing on citations to regulations appear very much the same. However, professional practice is less consistent. Justices of the U.S. Supreme Court and judges of lower federal courts will often include the year of compilation for a Code of Federal Regulations provision in an opinion’s first citation to it, leaving the year off all subsequent references. Arguably, for as long as annual print compilations constituted the principal source for codified regulations that approach furnished useful information. Today, with constantly updated compilations maintained not only by commercial online services but the Government Printing Office, it makes better sense for both writer and reader to understand that a citation in a brief or opinion to 37 C.F.R. § 205.22 refers to the provision in effect at the time of writing. With a section that has not changed since 2008, the addition of 2013 or, as of last week, 2014 in parentheses serves little purpose. For that reason many judges and lawyers would cite to 37 C.F.R. § 205.22 (or a state equivalent) without indicating a year, again, unless the litigation concerned an earlier version or the regulation in question has undergone recent change.


Dates are far less precise and therefore less useful in citations to journal articles. Because publication delays are common with student-edited journals, numerous articles that failed to appear in 2013 will nonetheless carry that date.  Many destined to appear in 2014 last received attention from their author or an editor in 2013. Judicial opinions, legislative enactments, and regulations all carry specific release or effective dates.  Individual journal articles do not. Notwithstanding the imprecision and limited utility, attachment of the nominal year of publication to article citations is accepted practice.

The same holds for treatise  citations with greater reason and despite a further difficulty. Most major treatises have been acquired by an online research service and are bundled with the service’s primary law materials.  In both print and online form they are updated at least annually. In print, the updates may be integrated, the case with treatises published in a looseleaf format, or they may be issued in a separate supplement.  Online, they are integrated without any indication of what was changed or when. Under these circumstances, how should one date a section of A. Wright, A. Miller & M. K. Kane, Federal Practice and Procedure or M. Nimmer & D. Nimmer, Nimmer on Copyright in a brief or opinion prepared during 2014? Should that depend on whether one accessed the material in print or online? Assuming that one is citing to the current work rather than a prior edition or version, the best practice is to cite to the year of the most recent update or revision of the source relied on. Following that practice one would in January 2014 still use the year 2013 for both those works since they were last updated during that year, a fact noted in their print and electronic versions.

The updating phenomenon bestows greater importance on the date associated with a treatise citation. Unlike journal articles these are not static works. A reference to a particular section as it existed some years ago, 2004 say, or 1994, may well, if followed into the current version of the treatise, take the reader to significantly different text . The year accompanying the citation provides a reminder of that reality even though it may be difficult verging on impossible for those working in a contemporary research environment to determine exactly how the cited section read in 1994 or 2014. Neither Lexis (Nimmer & Nimmer) nor Westlaw (Wright & Miller) archive past versions of their treatises online as they do past statutory codifications.

Cite thoughtfully in 2014!



Citing unpublished decisions

Thursday, December 5th, 2013

On December 16, 2010, a panel of the Eleventh Circuit, U.S. Court of Appeals, issued a per curiam opinion interpreting the Armed Career Criminal Act (ACCA), 18 U.S.C. § 924(e)(1) as it related to specific Florida crimes. The panel designated that opinion not for publication (“DO NOT PUBLISH”). This December opinion vacated an earlier one, dated September 8, also unpublished, that had misstated one of the defendant’s prior convictions. The new decision corrected the error. In all other respects it was identical. Although unpublished, under the Federal Rules of Appellate Procedure (Rule 32.1) that December 16 decision can be cited. A rule of the Eleventh Circuit (p. 147, Rule 36.2) explicitly provides that unpublished opinions are not binding precedent but “may be cited as persuasive authority.”

The issue to be considered here is how to cite such unpublished, non-precedential decisions.

Both the September and December opinions are available on the Eleventh Circuit Web site. They and other Eleventh Circuit opinions applying the same sentence enhancement provision of the ACCA can be found with a Google web search ( “Armed Career Criminal Act” “residual clause”) or through a search on Google Scholar limited to the Eleventh Circuit. Anyone finding the court’s decision in United States v. Hayes on the open Web would, however, be unaware that, notwithstanding, the “DO NOT PUBLISH” label the editors of Thomson Reuters selected the decision for publication in a set of books that no law library I’ve ever used has seen fit to buy or shelve, the Federal Appendix of the National Reporter System. (The Federal Appendix is for sale. The full set, currently 523 volumes, covering a mere dozen years, can be yours for only $7,093.80, just marked down from $10,134, perhaps for the holidays. However, the print market was never that publication’s aim.) Within that series the Hayes decision is reportedly located in volume 409, at page 277. That information is not available on the open Web. Furthermore, unless a person finding and wanting to cite Hayes is a subscriber to Bloomberg Law, Lexis, or Westlaw, she would not be aware that those services have designated it, 2010 BL 299236, 2010 U.S. App. LEXIS 25741, and 2010 WL 5122587, respectively. Those high end services also provide the case’s Federal Appendix cite, 409 Fed. Appx. 277 (or as converted by The Bluebook, 409 F. App’x 277). Persons with access to Casemaker or Fastcase could discover and retrieve the Hayes decision using a suitable query, but neither of those services adds their own proprietary citation or reports the citations added by their competitors.

One further point about the Federal Rules of Appellate Procedure and Eleventh Circuit additions – they provide no explicit guidance on how to cite “unpublished” but widely available decisions like Hayes. One can, however, find indirect policy guidance in the same Eleventh Circuit rule that allows their citation. It provides that “If the text of an unpublished opinion is not available on the internet, a copy of the unpublished opinion must be attached to or incorporated within the brief, petition, motion or response in which such citation is made.” Patently, this requirement is not focused on judicial access to such decisions. The judges of the Eleventh Circuit, like other federal judges, have access to both Lexis and Westlaw.  Rather the rule addresses the problem of access faced by parties without access to Westlaw, Lexis, Bloomberg Law and the rest, and citation format bears directly on access.  A citation to Hayes in a brief, memo, or court opinion reading: “United States v. Hayes, 409 F. App’x 277 (11th Cir. 2010)“ is utterly useless on the open Web. It will also fail to retrieve the decision on Casemaker and Fastcase. Yet that is precisely how The Bluebook would have the case cited once it has been selected for and received volume and page numbers in the Federal Appendix. (See Rule 10.5(a).) No doubt that is because The Bluebook is written by and for law journals, whose editors have access to at least one, if not all, of the Bloomberg Law, Lexis, and Westlaw trio. The ALWD Citation Manual similarly assumes the universal utility of a Federal Appendix citation. (See its Rule 12.14(b).)  In fact the ALWD manual goes farther down this false path than The Bluebook, for it authorizes citations to unpublished decisions that rely totally on Lexis or Westlaw cites, which are even less effective across systems, e.g., “United States v. Hayes, 2010 U.S. App. LEXIS 25741  (11th Cir. 2010)” or “United States v. Hayes, 2010 U.S. App. LEXIS 25741  (11th Cir. 2010).”

Until the federal courts begin attaching neutral citations to their own decisions, the only effective way to cite Hayes or any other “unpublished” but widely distributed decision is to include both the docket number and the full date of the decision, as in “United States v. Hayes, No. 09-12024 (11th Cir. Dec. 16, 2010).” The docket number, coupled with deciding court, enables retrieval of the opinion from all competing commercial research services, from Google Scholar and the open Web. The full date, particularly important with this example, allows anyone following the citation to realize that the vacated September 8 opinion, which the docket number will also retrieve, is not the target of the reference.

In sum, both The Bluebook and the ALWD Citation Manual have been led astray. An unpublished decision should be cited as an unpublished decision. Docket number, court, and full date work effectively to identify and retrieve a cited case across sources, including importantly the open Web. A citation to the Thomson Reuters Federal Appendix is no substitute. Nor is a citation using the proprietary numbering system of one of the commercial online services. Of course, there is no harm, beyond the space consumed, in adding a Federal Appendix, Bloomberg Law, Lexis, or Westlaw cite to that essential core. On the other hand, unless one is confident that all important readers of a document will have access to a system on which such a proprietary cite will work, the added value is not likely to be worth the increase in citation length.

Unfortunately, the judges of the Eleventh Circuit and the district courts over which it sits do not model this approach. Just as they impose no particular citation format on those appearing before them, they practice none. Hayes has been cited in numerous subsequent decisions, both published and unpublished. In United States v. Nix, 628 F. 3d 1341, 1342 (11th Cir. 2010) the earlier Hayes opinion is cited as “United States v. Hayes, 2010 WL 3489973 (September 8, 2010).” The dissent in Rozier v. United States, 701 F.3d 681, 688 n.5  (11th Cir. 2012)  cites to the Federal Appendix reporter, “United States v. Hayes, 409 Fed.Appx. 277 (11th Cir. Dec.16.2010).” United States v. Morris, No. 11-13064 (11th Cir. Aug. 15, 2012) (which appears in volume 486 of the Federal Appendix at page 853, if that is useful to you) cites the case, without either docket number or exact date, as “United States v. Hayes, 409 Fed. App’x 277, 278-79 (11th Cir. 2010).” Citations to Hayes, in recent decisions of the U.S. District Court for the Middle District of Florida, appear in the form: “United States v. Hayes, 409 F. App’x 277 (11th Cir. 2010), cert. denied, ___ U.S. ___, 132 S. Ct. 125, 181 L. Ed. 2d 47 (2011).”

Under the influence of those appearing before them and the guidance of their clerks, federal judges need to bring their citation practice into accord with the concern over access expressed in the Eleventh Circuit rule.