ChatGPT on Picking Evidence in Statutory (and Constitutional) Interpretation (9/13/26)

Today, I was reworking the part of my 2027 Working Draft for my Federal Tax Procedure Editions on the problem of cherry-picking or selecting evidence used in constitutional and statutory interpretation. The issue is presented anecdotally by conservative judges trotting out Judge Leventhal’s famous remark that using legislative history is like looking over the crowd and picking your friends. There is a similar phenomenon with so-called “law office history” deployed to support arguments made by advocates and judges in their opinions. Even broader, the same phenomenon is presented for the entire body of law that advocates and judges deploy—such as cases, scholarly treatises, law review articles, etc. The risk is always present that we as lawyers (or judges) pick our friends rather than make a neutral assessment of the evidence.

Continuing my experimentation with AI, I asked ChatGPT Plus about the issue and got what I think are pretty good responses. I presented the issue in two prompts (the second asking for more than the first) noted below in red. The bold face was from the original responses. I checked the citations to the two Supreme Court cases and they are correct (i.e., no hallucinations). I will use the responses in re-working my discussion in the Federal Tax Procedure Editions (not verbatim but to inspire how I state the concepts in my own words which I hope are at least as good as ChatGPT's).

As an introduction, I use the term "law office history." That term commonly describes the phenomenon of lawyers and judges evaluating history as a process of cherry-picking to support arguments. Of course, selecting the evidence is what we all do when making arguments or taking positions; lawyers do that in briefs and oral arguments; judges do that in opinions; we all do that in our daily discourse. Even when a person--say a judge--sees his or her duty to make an objective or fair assessment of the evidence, personal prejudices or preferences may influence the evidence used. I wanted to explore what that may mean.

One caveat: The same phenomenon occurs in factfinding where a judge selects from a body of factual evidence the evidence that he or she finds persuasive to determine facts relevant to the outcome of the case. I don't think factfinding is materially different from lawfinding--selecting among the relevant evidence to reach the outcome.

Prompt #1

Please write a short discussion of the features of "law office history" and "legislative history" that permit a judge or lawyer to cherry pick (or as one judge called it to look over a crowd and pick your friends).

Response to Prompt #1

Here is a short version in a law-review style that emphasizes the structural similarity between the two problems:

Law-Office History and Legislative History

Both “law-office history” and legislative history present a similar danger: they often supply the interpreter with a large and heterogeneous body of material from which supporting evidence can be selected without any obvious neutral rule for determining what evidence should receive the greatest weight. Alfred H. Kelly coined “law-office history” to describe the lawyerly practice of selecting historical data favorable to a position while failing adequately to confront contrary evidence or evaluate its significance in historical context. The problem is partly inherent in advocacy. The historical record is usually extensive, incomplete, and capable of supporting competing narratives; choices must be made about the relevant period, the appropriate level of generality, which practices are representative, and which sources deserve weight. Those choices can permit the advocate—or judge—to begin with a desired conclusion and assemble a plausible historical narrative supporting it.

Legislative history presents much the same problem, perhaps in an even more concentrated form. A statute may generate committee reports, sponsor statements, floor debates, hearing testimony, rejected amendments, conference materials, and statements by legislators who supported or opposed the final bill. These materials may point in different directions, and statements by particular legislators do not necessarily represent the understanding of the legislative body that enacted the statute. Moreover, participants aware that courts consult legislative history may have incentives to place favorable statements in the record. The resulting abundance of potentially conflicting material makes selective quotation unusually easy. The Supreme Court has accordingly repeated Judge Harold Leventhal's famous description of using legislative history as akin to “looking over a crowd and picking out your friends.” Exxon Mobil Corp. v. Allapattah Services, Inc., 545 U.S. 546, 568 (2005).

Judge Holmes Decides on Burden of Proof (Persuasion) (9/3/26)

In Risan v. Commissioner, T.C. Memo. 2026-78, TC here at #93, TN here, and GS here [to come], Judge Holmes decided significant factual issues on burden of proof. The key reason I thought this case interesting is that Tax Court Judges often, after noising about burden of proof, will just hold that, in effect, burden of proof is irrelevant to the outcome because the Judge in the particular case has resolved those factual issues by a preponderance of the evidence. That is simply to say that, in the normal case where the preponderance of the evidence standard applies, burden of proof is relevant only if the Judge is in factual equipoise unable to determine whether or not a critical fact is more likely than not. In that state of equipoise, the party with the burden of persuasion loses.

Judge Holmes sets up the discussion at the beginning of the Opinion section (starting at *17, footnote omitted):

          We begin with a number of issues for which Mr. Risan presented no evidence or argument, or about which he included only an underdeveloped argument in his brief. Taxpayers usually bear the burden of proof in cases before us. Rule 142(a). The Code does sometimes shift this burden to the Commissioner. See § 7491(a). But when neither party presented evidence in these cases, we simply find that Mr. Risan did not comply with the Code’s recordkeeping requirements and did not cooperate with the Commissioner’s reasonable requests for information. This means that, for a great many issues, we find against Mr. Risan and for the Commissioner.

 Other JAT comments:

1. Judge Holmes also addressed the statute of limitations. Judge Holmes found (i) there was a 25% omission in the years involved invoking the six-year statute of limitations and (ii) for some of the years during the six-year period, the taxpayer timely signed consents to extend. Judge Holmes said that, although the taxpayer testified that he did not sign the forms, visual review of the forms permitted him to “find it more likely than not that he did sign them.” (*19.) Although Judge Holmes did not say it directly, that finding means that it is more like than not that the taxpayer testified untruthfully (whether or not the taxpayer intended to deceive is a different issue).

Townsend ABA Tax Lawyer Article Titled Loper Bright Flip Flops on Chevron Deference: A Tax Lawyer's Perspective (8/28/26)

I have posted to SSRN my recent article in the ABA Tax Lawyer: John A. Townsend, Loper Bright Flip Flops on Chevron Deference: A Tax Lawyer's Perspective, 79 Tax Law. 323 (2026). The SSRN citation and link is: Townsend, John A, Loper Bright Flip Flops on Chevron Deference: A Tax Lawyer's Perspective (January 01, 2026), SSRN at: https://ssrn.com/abstract=7360440 [Note: I have slightly changed the SSRN "suggested citation" to eliminate duplicating my name (don't know why that happens) and shortened the reference to SSRN.]

One of the themes I develop in the article is that Chevron was not about deferring to an agency interpretation that was not the best interpretation. Rather, Chevron was about “deferring” to an agency interpretation when the court found the statute ambiguous, meaning that the court was in equipoise as to the best interpretation. Loper Bright flogged the false notion that Chevron permitted/required courts to defer to the agency "not best" interpretation over the competing "best" interpretation. Courts, particularly judges of Federalist Society bent, also flog that false notion ad nauseum. E.g., yesterday, I read Piedra v. Blanche, ___ F.4^th ___ (4^th Cir. 8/24/26), CA4 here and GS here, which starts off (Slip Op. 3):

For decades, Chevron deference restricted courts from exercising their duty "to say what the law is." Marbury v. Madison, 5 U.S. (1 Cranch) 137, 177 (1803). Newly freed from that restraint, courts now exercise their independent judgment about the best meaning of statutes.

To repeat, Chevron required courts to apply the best interpretation at Step One; if, after rigorous interpretation (Chevron footnote 9) the statute was still ambiguous, only then did courts “default” in equipoise to the agency interpretation. But that default was not applying the "not best" interpretation. Of course, I flog my criticism of Loper Bright in the article, pp. 339-341.

So, after Loper Bright, the problem for which Chevron applied a default rule still exists when a court, being honest with itself, cannot determine whether the agency interpretation or the (or any) competing interpretation is the best—a state of ambiguity (or equipoise). In the article, I develop this issue in a footnote (p. 346 n. 87 emphasis supplied):

CFC Adopts U.S. Interpretation of U.S.-Canada Tax Treaty Over Canadian Interpretation (8/27/26)

In The South Saskatchewan Comm. Fdn., Inc. v. United States, ___ Fed. Cl. ___  (8/25/26) (“South Saskatchewan”), CFC here, TN here, and GS here [to come], the Court rejected a Canadian charity’s attempt to apply an exemption in the United States-Canada Convention with Respect to Taxes on Income and on Capital (including a key Protocol). I found South Saskatchewan interesting because of the CFC’s application of interpretive techniques to resolve the treaty dispute between the parties in a situation where the Canadian tax authority and the U.S. did not agree on the interpretation as applied to the facts.

South Saskatchewan starts with the goal of treaty interpretation being to effectuate the shared understanding of the parties to the treaty. This is the contract model of interpretation. But, in this case, Canada’s understanding of the treaty is not the U.S.’s understanding. I infer that, assuming Canada’s current interpretation was taken in good faith, Canada did not have a shared understanding with the U.S. What does a court do when it cannot discern a shared understanding?

South Saskatchewan says (Slip Op. 6) after citing Loper Bright:

The meaning of both U.S. and Canadian laws, as well as that of the Tax Treaty, is susceptible to being construed, interpreted, and applied using traditional methods of statutory interpretation.

I am not sure how traditional methods of statutory interpretation flange with the goal of interpreting treaties to apply the shared understanding of the parties. Since, for U.S. purposes, the treaty is the law of the land and the treaty is text, I guess it is close enough to written law (statutes) that the interpretive model for statutes can apply. Maybe.

The Court adopts the U.S. interpretation based on a textualist reading of the Treaty Protocol. The Court finds helpful “extrinsic evidence” in the “legislative history” [caveat, legislative history is my term not the Court’s] of the Senate’s ratification of the Protocol. This legislative history is (i) the Treasury Technical Explanation accompanying the submission of the Protocol to the Senate for approval and (ii) the Joint Committee on Taxation (“JCT”) Report which “reflects the contemporaneous understanding of the Senate that ratified the Fifth Protocol.” (Slip Op. 9.) Both documents informed the Senate of the meaning of the Protocol.

Recommended Article on Constitutional Interpretation Applicable to Statutory Interpretation and Loper Bright (8/22/26)

Yesterday and today, I have been thinking about an article that implicates some of the interpretive issues I have been writing on post-Loper Bright. The article is: John O. McGinnis & Mike Rappaport, Originalism’s Better Meaning (Law & Liberty 8/13/26), here. Although developed in the context of constitutional interpretation and originalism, the article makes some points that seem to me applicable to statutory interpretation more broadly than the originalism spin.

First, the authors define the difference between interpretation and construction:

Interpretation is the process of determining the meaning of a constitutional provision. Construction begins only when that meaning is deemed indeterminate—that is, when it is thought to run out—and the decision must be made on grounds other than the provision’s original meaning.

I would restate that for statutory interpretation as follows (with changes in bold):

Statutory interpretation is the process of determining the meaning of a statutory provision. Statutory construction begins only when that meaning is deemed indeterminate—that is, when it is thought to run out—and the decision must be made on grounds other than the provision’s meaning as determined by statutory interpretation.

In these constructs, Chevron deference was construction rather than interpretation because it applied only when the meaning could not be determined and applied as a default rule to resolve cases before the courts without regard to the indeterminate meaning of the statute.

One interesting construct the authors deploy in the article is the “51-49 rule” which permits the “best” interpretation to be determined by the preponderance of the evidence standard (applying fact-finding concepts to a state of statutory interpretive ambiguity). In this construct, interpretation of a constitution (or statute) requires a confidence level for an interpretation to be at least 51%. The construction zone is between 50.9999% and 49.0001% where some default rule is required to resolve cases.

(Caveat: I don’t think that is a realistic range in statutory interpretation (or fact-finding for that matter); can anyone meaningfully distinguish between an interpretation that is 49.9% and one that is 50.1%?; but let’s just accept that construct for now; for my discussion of such ranges in valuation context, see John A. Townsend, Burden of Proof in Tax Cases: Valuation and Ranges—An Update, 73 Tax Lawyer 389 (2020).)

First Circuit Holds that the Time Period to File Petition in § 6213(a) is NonJurisdictional But Mandatory, Hence No Equitable Tolling (8/19/26; 8/20/26)

In Kyick Holdings, LLC v. Commissioner, ___ F.4th ___ (1st Cir. 2026), CA1 here and GS here [to come], the Court held that

1. The IRS properly sent the notice to the taxpayer’s last known address. This is the less exciting holding, but I will address it below in my Comment #2.

2. Section 6213(a)’s timely filing requirements are not jurisdictional. (Caution on this label “jurisdictional” though, as I note below.)

3. As a matter of statutory interpretation § 6213(a)’s timely filing requirements are “mandatory,” meaning that the periods to file the petition is not subject to equitable tolling. This latter holding has the same practical effect as would have applied if the Court had agreed with the consensus holdings in the Courts of Appeals that § 6213(a)’s time periods were jurisdictional.

The key holdings for purposes of this blog are ## 2 & 3, although they are not outcome-determinative in terms of the holdings of other Circuit Courts of Appeals. That is because regardless of whether a Court bases its outcome on jurisdiction or a mandatory analysis, the result is the same in denying the availability of equitable tolling.

For that reason, I worry whether courts futzing around with “jurisdiction” in this context is meaningful. Courts could just go to the heart of the disposition that the time periods are mandatory. In many cases, courts are just writing at length about jurisdiction which has some esoterica when mandatory should do.

Added 8/20/26 3:25pm: For a different outcome see Maniktala v. Commissioner, ___ F.4th ___ (8th Cir. 8/11/26), CA8 here and GS here, allowing equitable estoppel for time deadlines in § 6213(a) based on holdings that (i) § 6213(a) is nonjurisdictional (consistent with Kyick) but (ii) the § 6213(a) deadlines are not mandatory (inconsistent with Kyick). It appears that there is a clear conflict in the Circuits on the statutory interpretation issue which is often a substantial basis for granting a petition for writ of certorari. I should note that, in effect, Maniktala seems to be based on the Eighth Circuit's application of the default rule allowing equitable estoppel in a case where the court cannot determine the statutory meaning (in interpretive equipoise), whereas Kyick is based on an interpretation of the statute where it could it could determine the statutory meaning. Although an agency interpretation that might have qualified for Chevron deference (i.e., a Treasury interpretive regulation) is not in issue, this split does highlight one of the central fanciful notions of Loper Bright that courts can always interpret statutes to reach the best interpretation without the need for default rules for decision.

Other JAT Comments:

Eleventh Circuit Affirms Judge Halpern's Rejection of Bullshit Tax Shelter--the Conservation Easement Variety (8/15/26)

In Evans v. Commissioner (11^th Cir. Nos. 24-11882 & No. 24-11884 Unpublished Opinion dated 8/13/26), CA11 here, GS here [to come], and Tax Notes here, the Court affirmed the Tax Court’s decisions based on the opinions Carter v. Commissioner, T.C. Memo. 2023-133 (T.C. Dkt. 23647-15, here, at #112 and GS here), a fine Judge Halpern opinion.

The Eleventh Circuit majority gave the appellants’ bullshit arguments the treatment they deserve (even more than they deserve). (There ought to be an appellate disposition simply by rejecting an appellant’s arguments in two lines:

Appellant’s arguments are bullshit. [No exclamation mark needed; a simple statement of fact]

Affirmed.

For that reason, I hesitated to write anything on the Eleventh Circuit’s unpublished opinions in this case that have already involved grossly disproportionate attention from all involved (including perhaps not-so-innocent readers such as me).

Fifth Circuit on Rehearing Makes the Same Mistake on the Limited Partner Gambit to Avoid Tax (8/13/26; 8/20/26)

On 8/20/26 at 12:25 am, I substantially revised this blog entry. My additions are indicated in red; but I do not show strikeouts.

In Alain, L.L.L.P. v. Commissioner, ___ F.4th ___ (5th Cir. 8/12/26), CA5 here; GS here, the Court on panel rehearing (also denying en banc rehearing), again rejected the Tax Court and IRS interpretation of the limited partner exception to the “self-employment income” in § 1402(1)(13). Readers paying attention in the tax procedure universe are surely familiar with that issue. Basically, active service participants in a business enterprise who would have had self-employment income in straight-forward characterization of their earnings have tried to avoid that tax on self-employment income by the magic of labeling their income as a distributive share for share for limited partners.

Alain was originally decided in Sirius Solutions, L.L.L.P. v. Commissioner, 165 F.4th 374 (5th Cir. 2026), GS here, which the Court withdrew on the rehearing. The judges on the panel rendered opinions to the same effect as before. The Sirius majority opinion was nominally authored by Judge Oldham. The panel rehearing opinion is per curiam with neither of the majority judges stepping up as the author.

I critiqued the majority in Sirius Solutions in Fifth Circuit Knows a Limited Partner When Reads It (Federal Tax Procedure Blog 1/24/26; 1/20/26), here. In my original blog on Alain, I did not pay attention to the difference between the original opinion in Sirius and the revised opinion in Alain. The panel majority states its bottom-line interpretation in the opening (Slip Op. 1): 

Today, we hold its original public meaning is a partner who plays no significant role in managing or running a business.

As stated, the panel majority’s revised formulation narrows the limited partner exception it approved in its earlier opinion. The panel majority deploys the tools of interpretation that textualists so love (such as contemporary dictionaries) to determine (divine) the original public meaning to possibly include only a smaller subclass of nominal “limited partners”—specifically those limited partners who have no significant role in managing or running the business. That narrowing is at least an improvement on the earlier opinion. 

Judge Graves engages the majority on its determination of original public meaning. (Slip Op. 21-25.)

Questions for Tax Crimes and Tax Procedure Students on Trump's "Settlement" in Trump v. IRS (8/6/26)

I have posted my Federal Tax Procedure Book to SSRN, see here. Since it is timely, although a moving target that we sure have not heard the last of yet, I thought I would post the concluding paragraphs of my discussion of Trump v. IRS that Trump and his buddies at DOJ (most notably Blanch, but not only Blanche) tried to use as a pretext to raid the Treasury.

In my opinion, Trump and his buddies new they would draw flack on that package, so they must have realized that the Anti-Weaponization Fund was just a ploy, a stalking horse. The real goal was tax audit immunity which for now has not been taken off the table. I discuss that in the Federal Tax Procedure Book (Student Edition pp.  802-805; Practitioner Edition pp. 1118-1192). I conclude that discussion with some further discussion questions for students of tax crimes and tax procedure:

Exercise for Students: Readers of Chapter Six discussing tax and tax-related crimes should be able easily to spot several tax and tax-related crimes that this conduct might implicate, particularly the ubiquitous defraud conspiracy. One possibility is that anyone participating materially in the audit immunity could be an affirmative act of evasion with respect to the taxes covered or an overt act of conspiracy (offense or defraud). In theory, if that were viable, all of the key players in this drama stand exposed to criminal prosecution. Of course, the prosecutions, if any, will have to be brought by DOJ which Trump can prevent while he is President but can then be brought in the next Administration unless Trump gives sweeping pardons to all who were involved. (I suspect and have read that those actors are counting on such pardons.) One final question worth asking is whether Trump’s control of DOJ and IRS could implicate some conduct for which the Supreme Court has given full or qualified immunity in Trump v. United States, 603 U.S. 593 (2024).

Another Exercise for Students: Tax procedure students should think about the settlement authorities discussed earlier in this text. Thus, generally, only the IRS has settlement authority for cases that it has not yet referred to DOJ. DOJ has settlement authority only for cases that the IRS referred. Facially, it appeared in Trump v. IRS that the IRS had only transferred authority to DOJ over the § 7431 wrongful disclosure suit. There is no indication that the IRS referred all of the matters sweepingly released in the Release Order to DOJ, and there is no indication that the IRS “settled” those claims.

Concluding thoughts: If nothing else, Trump v. IRS and its resulting commotions will occupy tax procedure and tax crimes enthusiasts for a long time.

 This blog entry is cross-posted on the Federal Tax Crimes Blog here.

2026 Federal Tax Procedure Book Student and Practitioner Editions (8/4/26)

I have published to SSRN my Federal Tax Procedure Book Student and Practitioner Editions. For information on the SSRN links to download see the page to the right titled “Federal Tax Procedure Book (2026 Editions) (8/4/26)” here.