Richard J. Malouf, D.D.S. v. the State of Texas Ex Rels. Christine Ellis, D.D.S. and Madelayne Castillo ( 2024 )

Supreme Court of Texas
══════════
No. 22-1046
══════════
Richard J. Malouf, D.D.S.,
Petitioner,
v.
The State of Texas ex rels. Christine Ellis, D.D.S. and Madelayne
Castillo,
Respondents
═══════════════════════════════════════
On Petition for Review from the
Court of Appeals for the Eighth District of Texas
═══════════════════════════════════════
JUSTICE YOUNG, joined by Justice Lehrmann, dissenting.
What do robots and lawyers have in common? Maybe more than
we would care to admit, but at least one answer is that only robots and
lawyers could read the statute at issue the way the Court does today. The
Court purports to do so in the name of textualism. But a textualist’s
obligation is to construe a statute in its context, giving it the meaning an
ordinary English speaker would have given it at the time it was enacted.
The Court instead imposes an implausible reading that no one would have
given it when it was written. Today’s reading at best adopts a post hoc
construction of the sort that good lawyers scrambling for an escape hatch
might concoct for their clients.
Specifically, the Court holds that Dr. Malouf did not violate Texas
Human Resources Code § 36.002(8). Chapter 36 is titled “Health Care
Program Fraud Prevention.” Section 36.002 is titled “Unlawful Acts” and
provides that
[a] person commits an unlawful act if the person: . . .
(8) makes a claim under a health care program and
knowingly fails to indicate the type of license and the
identification number of the licensed health care provider
who actually provided the service[.]
Tex. Hum. Res. Code § 36.002(8). Dr. Malouf submitted nearly 2,000
claims to the State that falsely listed his own “identification number”
rather than “the identification number of the licensed health care
provider who actually provided the service.” Did he commit what the
statute calls an “unlawful act”?
The Court says that he did not. True, he failed to provide the
identification numbers of the dentists who actually provided the services
at his dental chains, and instead listed someone (himself) who certainly
did not provide those services. That sounds bad. But, the Court says, not
to worry: Dr. Malouf implicitly provided those dentists’ license type (i.e.,
“dentist”) because, as it turns out, Malouf is also a “dentist.” According
to the Court, a lie about who “actually provided the service” is just as good
as the truth, so long as the unnamed person who did it was a dentist.
How could the Court misread the statute so badly? How could it
read the text that I quoted above and think that the legislature was
indifferent to knowing “who actually provided the service” for which the
State is being billed? It is because the Court does not engage in a
“context-sensitive interpretation of [the statute] as a whole.” United
States v. Palomares, 52 F.4th 640, 649 (5th Cir. 2022) (Oldham, J.,
2
concurring). It instead zeroes in on the word “and” to justify giving the
statute a meaning that its context will not remotely bear. I hope the
Court is not serious about what it says, because if it is, every statute,
contract, deed, will, ordinance, or other document will now be subject to
a hyper-literal insistence on how to read the word “and”—and will often
still get it wrong. It is the Court, not I, who makes “and” mean “or”— a
checklist using “and” to link requirements now is just stating two mere
options. Fun days ahead—at least for lawyers who bill by the hour.
The canons of construction are the essential tools for interpreting
legal documents. But like the tools of every trade, the canons must be
wielded with reference to their object. Our goal is to understand the
meaning of a text in its context. Common English usages—like the fact
that sometimes there is no semantic difference between choosing “and”
rather than “or,” as we all know—need not be sacrificed at the altar of
textualism. Such a sacrifice does no honor to its purported deity. I fear
that the Court’s approach reflects a turn to a false god—one who tempts
with the lure of easy answers and happy outcomes—while disavowing the
hard work of faithfully and accurately discerning a text’s true meaning.
The result today is that Dr. Malouf walks away scot-free. Among
those who will be surprised by this result is Dr. Malouf. The argument
that the Court adopts is his “alternative argument,” as the Court
acknowledges. See ante at 13–14. His far more modest lead argument
merely asserts that there are fact questions about whether he
“knowingly” listed the wrong provider, so the Court should send the case
back for trial. Id. at 13. Unlike the lead argument, the “alternative” is a
kill-shot—one on which he spends scant pages in both his opening and
3
reply brief, compared to the many pages devoted to his lead argument.
The Court should not have fallen for it, and I doubt that Dr. Malouf or his
counsel thought that we would. But as is often said at legal conferences
and when lawyers tell their war stories, one never really knows what a
court will do. Today’s decision should be added to the lore.
I respectfully dissent.
I
We all agree that this case turns on the text of the Medicaid Fraud
Prevention Act. Specifically, the statute says that it is unlawful to
“make[] a claim under a health care program and knowingly fail[] to
indicate the type of license and the identification number of the licensed
health care provider who actually provided the service.” Tex. Hum. Res.
Code § 36.002(8) (emphasis added). The case turns on this question: what
does the italicized “and” mean?
In my view, “and” plays its normal role of joining both prongs. The
statute creates a short checklist of two things (not just one or the other)
that a provider must list: the license type + the identification number.
Failing to “indicate” either of them violates the statute, just like failing to
put both ham and cheese on a sandwich would violate the Court’s
hypothetical deli order: “Don’t forget to put ham and cheese on my
sandwich.” Ante at 22. In our new age of artificial intelligence, I suppose
that a robot waiter or a robot cook (or should I say “and a robot cook”?)
would feel justified in serving a sandwich as soon as either ham or cheese
is placed between slices of bread. But a human would regard the order
as incomplete and would feel no guilt in sending it back.
The conjunctive meaning of the stated statutory requirements is
4
clear from the statute’s text, context, and textually expressed function.
But the Court insists on an arid reading that would make even a 1950s
high-school English teacher blush. Treating the “and” as really meaning
“or”—that the provider can list one or the other to escape any
consequence—is grammatically permissible (barely), as with the
concomitant sandwich in the deli. But it is not a remotely reasonable
reading and gives no hint of pursuing an accurate rather than a
tendentious interpretation of the statute.
A
As I read it, “and” conveys its normal grammatical meaning and
plays its normal role—it imposes a joint requirement by listing two things
(A and B) that a provider must indicate. But suppose for the moment
that the Court would accept the dual-requirement reading if the
legislature instead had used “or” in the statute. Even under the dubious
assumption that “or” would satisfy the Court,1 that does not mean that
“and” means anything different in this context. We have properly held
that “and” and “or” generally are not interchangeable. In re Brookshire
Grocery Co., 250 S.W.3d 66, 69–70 (Tex. 2008). They are, of course,
usually distinct in meaning: “in a legal instrument, and joins a
conjunctive list to combine items, while or joins a disjunctive list to create
alternatives.” Conjunctive/Disjunctive Canon, Black’s Law Dictionary
(11th ed. 2019).
“Not A, not B, and not C,” however, is logically equivalent to “not
1 [Narrator: The Court would not. If it will not accept “and,” it would
not accept “or,” but would make the very same argument in the photographic
negative, saying if the legislature wanted both, it would have said “and!” ]. See
also infra note 3.
5
(A, B, or C).” Linguistic context thus enables us to “move back and forth
between disjunctive and conjunctive propositions as long as we are
mindful about negations, as well as ‘our p’s and q’s.’ ” R.E. Houser, Logic
as a Liberal Art: An Introduction to Rhetoric and Reasoning 343 (2020).
Therefore, as Chief Justice Greenhill put it for the Court, we have also
recognized that “there may be circumstances which call for such a
construction” in which “and” is construed to mean “or,” even if we try hard
to avoid those constructions. Bayou Pipeline Corp. v. R.R. Comm’n, 568
S.W.2d 122, 125 (Tex. 1978); see also De Sylva v. Ballentine, 351 U.S. 570,
573 (1956) (“the word ‘or’ is often used as a careless substitute for the
word ‘and’; that is, it is often used in phrases where ‘and’ would express
the thought with greater clarity”); Bryan A. Garner, Garner’s Modern
English Usage 49 (4th ed. 2016) (“and is frequently misused for or where
a singular noun, or one of two nouns, is called for”).2
Depending on its statutory context, therefore, “and” can have
several meanings. Some may be the opposite of how that word normally
functions; some may simply be interchangeable with the word “or.” When
that happens, a court does not rewrite “and” to mean “or”—that is what it
meant all along. It is basic to our language that the same word can mean
2 This linguistic phenomenon is not merely an academic or logical trifle,
but in fact occurs in parlance ranging from the mundane to the literally divine.
Did Jesus mean that someone who “left” his “children, for the kingdom of God’s
sake,” but who refused to leave his “house, or parents, or brethren, or wife” for
that cause will still be richly rewarded? See Luke 18:29–30 (“Verily I say unto
you, There is no man that hath left house, or parents, or brethren, or wife, or
children, for the kingdom of God’s sake, Who shall not receive manifold more
in this present time, and in the world to come life everlasting.”); cf. Luke 14:26
(“If any man come to me, and hate not his father, and mother, and wife, and
children, and brethren, and sisters, yea, and his own life also, he cannot be my
disciple.” (emphasis added)).
6
different things or function in different ways, so we rely on context to
discern the applicable meaning. Sometimes that meaning is so evidently
clear that we barely notice. “He tapped the mouse” and “he caught the
mouse” give the word “mouse” two very different meanings—although
both sentences could deploy either meaning. This point is so
commonplace as to be truly banal.
Particularly relevant here, “and” can be used in “a distributive (or
several) sense as well as a joint sense.” Bryan A. Garner, Garner’s
Dictionary of Legal Usage 639 (3d ed. 2011) (emphasis added). The Court
chooses the distributive sense: no problem unless a claimant knowingly
fails to provide both his license type and his identification number. See
ante at 32. Sometimes this use is the only reasonable one. Judge Willett
offered this example: “ ‘Do not mix heat, fuel, and oxygen’ instructs the
reader to prevent the unity of all three ingredients unless she wants a
fire.” Palomares, 52 F.4th at 653 (Willett, J., dissenting). Mixing any two
is fine—no fire unless heat, fuel, and oxygen are present. So I readily
agree that the Court’s reading—that there is no statutory violation unless
the claimant (1) fails to provide her license type and (2) fails to provide
her identification number—is grammatically possible. But if three
colleagues will get into trouble in any combination, one could say “Do not
mix Adam, Ben, and Cole” without meaning that it is fine to have two but
not all three of them. I take the Court to agree with this basic principle.
Ante at 21–22.
The examples and counterexamples are tiresome because they are
inexhaustible. That very point, however, proves that context is what
matters. The ham-sandwich example hardly stands alone. To take
7
another, suppose a dentist (maybe even Dr. Malouf—or at least someone
using his identification number) tells a patient to “brush and floss your
teeth.” Or, to inject the negative, “you’ll be paying dental bills if you forget
to brush and floss your teeth.” Even simply “don’t forget to brush and
floss.” One who brushes without flossing will pay a price—gingivitis or
worse—for giving the dentist’s advice a bizarre but grammatically
tolerable meaning.
I suppose that the legislature could commit itself to drafting only
asyndetic statutes—avoiding conjunctions at all costs and using
structure, subparts, and other language to eliminate any contrary
grammatical reading. And sure, the legislature could have written this
statute to more directly say what it meant.3 Greater clarity is always
3 The Court thinks it is attacking me by saying that “[i]f the Legislature
mistakenly used the term ‘and’ instead of ‘or’ in Section 36.002(8), it is up to
the Legislature—not the courts—to amend the statute to fix that mistake.”
Ante at 29 n.19.
The Court mistakes my point. I agree that courts are not in the business
of correcting drafting errors (at least those that are not scrivener’s errors, see
Antonin Scalia & Bryan A. Garner, Reading Law: The Interpretation of Legal
Texts 134–35 (2012)). But this statute contains no drafting (or scrivener’s) error.
As drafted, it means that a provider must indicate both the license type and the
actual provider’s identification number. As I explain below, the word “and” (like
many words) can mean different things in different contexts, and in this context,
the Court’s reading is unreasonable. In other words, I hardly propose to
(improperly) “fix” some substantive legislative “mistake,” but instead to
(properly) apply the statute as the legislature wrote it, in light of its statutory
context and the statute’s textually expressed function.
Regardless of all that, I doubt that the alternate phrasing would “fix” the
legislature’s supposed “mistake.” If the statute replaced “and” with “or,” the
parties would still be before us. The State would argue that a knowing failure
to provide either the license type or the identification number violates the
statute. Dr. Malouf would argue that “or” gives claimants a choice: so long as
8
desirable and nearly always possible. (For judicial opinions, too—not
just statutes.) But courts cannot “demand (or in truth expect) that [the
legislature] draft in the most translucent way possible.” Pulsifer v.
United States, 144 S. Ct. 718, 729 (2024).
Statutes end up as they do for a myriad of reasons—speed as the
session comes to an end, compromise in merging House and Senate
versions, desire to minimize revisions to existing laws, the fusion of
competing versions of a text, human frailty, and so many others. True,
such circumstances sometimes lead to an enactment that unambiguously
requires something the legislature likely did not want—and when that
happens, we must follow the actual law. But when we interpret the law
in the first place, we are not required to give the legislature the least
charitable reading we can. Indeed, we are not authorized to do so. We
must instead “ ‘ascertain and give effect to the Legislature’s intent,’ ” and
we do so by “enforc[ing] the plain meaning of statutory text, informed by
its context.” Hegar v. Health Care Serv. Corp., 652 S.W.3d 39, 43 (Tex.
2022) (emphasis added) (quoting In re D.S., 602 S.W.3d 504, 514 (Tex.
2020)). After all, “[w]ords in a vacuum mean nothing. Only in the context
of the remainder of the statute can the true meaning of a single provision
be made clear.” McLane Champions, LLC v. Hous. Baseball Partners,
LLC, 671 S.W.3d 907, 920 (Tex. 2023) (quoting Bridgestone/Firestone,
Inc. v. Glyn-Jones, 878 S.W.2d 132, 133 (Tex. 1994)). Specifically in
the claimant provides the license type or the identification number, there is no
violation. We would still have to resolve the question if the legislature made the
single wording switch. See Pulsifer v. United States, 144 S. Ct. 718, 729 (2024)
(“[W]e doubt that substituting ‘or’ for ‘and’ would have delivered us from
interpretive controversy.”).
9
statutes like this one—where the use of “and” within a statement is at
issue—“context may drive such a statement in either direction.” Pulsifer,
144 S. Ct. at 729 n.5 (offering examples using an identical linguistic
structure but where “and” will bear opposite meanings).
B
So if context is indispensable to determining meaning—and the
Court at least says that it admits this, ante at 6, 23–27—the case should
be easy. The Court’s reading is unreasonable, and the answer to why that
is so “lie[s] in considering the [statute’s] text in its legal context.” Pulsifer,
144 S. Ct. at 731.
Context comes in many forms. Some are irrelevant, but those
drawn from the text itself are certainly proper for courts to use. More
to the point, we cannot smuggle in impermissible grounds just by calling
something “context.” But the full text is always legitimate. When
interpreting a statute to resolve another “and-or” dispute, for example,
we referred to a “succeeding clause” as one way to contextualize the
meaning of “and.” Bd. of Ins. Comm’rs v. Guardian Life Ins. Co. of Tex.,
180 S.W.2d 906, 909 (Tex. 1944). Here, the succeeding clause is this
phrase: “of the licensed health care provider who actually provided the
service.” Tex. Hum. Res. Code § 36.002(8). This phrase modifies both the
(1) license-type and (2) identification-number prongs. See ConocoPhillips
Co. v. EPA, 612 F.3d 822, 839 (5th Cir. 2010) (“Nouns joined by
coordinating conjunctions are usually treated as a single, compounded
unit, and a postmodifying prepositional phrase is most naturally read to
modify that single unit.”).
The successive phrase contextualizes how we should construe the
10
word “and.” The statute tells a claimant to provide the license type and
identification number to enable the State to know the true identity “of the
licensed health care provider who actually provided the service.” Tex.
Hum. Res. Code § 36.002(8) (emphasis added). Providing only one of those
two requirements contravenes this clear textual mandate.
The two required data points are not substitutes but are quite
notably at opposite ends of the spectrum. The license type is the most
general (“dentist”); the identification number is the most granular (one
specific dentist and no other). The statute is openly and expressly an anti-
fraud statute, too—not just one to punish fraud, or even just to detect
fraud, but to prevent it altogether. Again, Chapter 36 of the Human
Resources Code is titled “Health Care Program Fraud Prevention.” So is
it really plausible that the legislature is indifferent about which of two
fundamentally dissimilar data points are provided? Is it plausible that
the legislature is indifferent to receiving false information about one data
point, so long as it receives accurate information about the other—yet
does not even care to know which one is true and which is false?
The answer, of course, is no. Both data points, after all, could be
useful in various anti-fraud efforts. They could work together in
individual cases as a kind of check—if a license type and identification
number did not match, that would signal that something has gone awry,
flagging the claim for more attention. Even taken separately, both data
points could be useful in different ways. For example, given how many
Medicaid claims are made, it would surely be useful to sometimes
generate reports to see if there are outliers or patterns about the type of
provider who performs specific kinds of treatments. If certain services
11
are normally provided by “dental hygienists,” running reports to see when
they are provided by “dentists” could help identify anomalies, which
might lead to further analysis, audits, improvements, or the like. Other
circumstances—especially if there is an audit—might make the license
type irrelevant, but make it very important to know the individual
provider. Many of these purposes would not be possible if the form
includes one data point but not the other—or, perhaps worse, if the form
accurately includes one but falsely includes the other.
So can anyone seriously think that the legislature just does not
care if the data it demands is corrupted in this way? That the legislature
only cares about getting one piece of accurate data for any individual
claim, without any interest whatsoever in being able to reliably detect
trends in specific practices, regions, or across the State? Reading the
statute in such a way is puzzling at best.
To be clear, all we are doing is reading the statute to see what the
statute contemplates. But the Court reads the statute anachronistically—
it relies on a current form that is not part of the statute, and reasons
backwards to conclude that this later-adopted form reached back from the
future to determine the meaning of the statute. See ante at 24–25 (relying
on the “Medicaid-approved claim form”).
Let me explain what is going on. The statute (which again is what
matters) requires the license type and individual identification number.
The statute applies to all Medicaid providers, not just dentists. As it
turns out, at least for dental providers and during the time at issue, the
Health and Human Services Commission could ascertain both required
data points if given the “Texas Provider Identifier” number, so the then-
12
current claim form requested only that.4 It is easy to see how that could
happen. Imagine that a statute demanded a license number and a date
of birth, and that when the statute was enacted, nothing in the license
revealed a birth date. But imagine that over time, the licensing authority
began adding the birth date as the final digits of a license. At that point,
asking for only the license number would satisfy both statutory
4 As the Court explains, the Medicaid claim form at use here included
only one box—the identification-number box—and did not have a place for
claimants to separately indicate a provider’s license type. Ante at 15. Because
the license type is linked to a provider’s identification number, I agree with the
Court’s conclusion that the submission of an accurate identification number
provided both pieces of information. Id. at 14–15.
The record indicates, however, that the claim form in use during the
relevant timeframe for this litigation (2007–2010) was “dental specific.” And in
the current Texas Medicaid Provider Procedures Manual, dentists participating
in the THSteps program are required to submit an ADA Dental Claim Form.
Tex. Health & Hum. Servs., Texas Medicaid Provider Procedures Manual: Vol.
1, Section 6: Claims Filing (2024). By submitting a “dental specific” form, one
could argue that the claimant impliedly indicated the license type as “dental.”
Would that be enough for the Court?
The point, though, is that the Court’s holding would apply with equal
force if the form included two boxes: one for the identification number and one
for the license type. As it happens, the current claim form appears to ask for
both. The Texas Medicaid Provider Procedures Manual requires a dental
claimant to submit a unique provider identifier number and a taxonomy code for
all paper claims. Id. “A taxonomy code is a unique 10-character code that
designates [a medical provider’s] classification and specialization.” Find Your
Taxonomy Code, Centers for Medicare & Medicaid Services, https://
www.cms.gov/medicare/enrollment-renewal/providers-suppliers/health-care-
taxonomy (last visited June 18, 2024). In other words, the taxonomy code is a
unique number that identifies a provider’s license type. The manual then states
why the claim form requires a taxonomy code in addition to the provider’s
identifier number: the former is “used to crosswalk the [provider’s unique
identifier number] to the billing provider.” Tex. Health & Hum. Servs., Texas
Medicaid Provider Procedures Manual, supra.
But good news, dentists: the Court has now dispensed with this
requirement, despite how sensible and useful it likely is.
13
requirements. But if the licenses stopped including birth dates, then the
form would need to ask for both.
Both in the real circumstances of today and in my hypothetical,
however, it remains important to accurately reflect both data points. In
today’s case, for example, the form requested only the identification
number because, if listed accurately, it would accurately generate both
required data points. Falsely listing the identification number
guarantees that both required data points cannot be accurate. At most,
as here, a false identification number might match the correct license
type. (And in the dental world, it often will—“dentist” is hardly rare.) But
at no point could the Commission ascertain both data points from the
license type alone, no matter how a claimant provided that information.
A false identification number certainly cannot generate an accurate
identification number.
Yet under the Court’s construction, a claimant may just make up
the identification number, or use someone else’s, or leave it blank—as
long as he somewhere scribbles what his license type is or otherwise so
indicates. Any of those actions ensure that the State will not know the
identity of the person “who actually provided” the medical services—the
one thing that the statute makes abundantly and expressly clear it wants.
Yet the Court excludes such falsehoods from the statute’s prohibitions,
and says that all the State needs to know is that some anonymous
“dentist” did the work—maybe the one whose identification number is
used, maybe not. Who knows? That really narrows it down. How helpful.
The Court, in other words, says that “indicat[ing]” the “license
type” is all the statute requires—form or no form. This case shows the
14
consequences of that view. When a particular identification number
embeds both data points, as with the “TPI” that the then-current form
requested, see ante at 15–16, the truth about both is revealed with one
accurate number. Telling a lie will often accurately generate the license
type, but never the identification number. The Court thinks that the
legislature is fine with the lie—that it unambiguously approved the lie.
The Court invokes statutory context to defend this remarkable
conclusion in three ways: (1) that the construction that I advance is
allegedly surplusage, id. at 24–25; (2) that the statute’s use of the word
“or” in nearby provisions proves that it uses “and” the way the Court
claims, id. at 25–27; and (3) that the statute is not stated as an
“affirmative command,” id. at 27 n.19. That these are the best arguments
just proves that there are no good arguments.
1
First, the Court argues that both the license type and the
identification number of the person who actually provided the medical
service are not necessary, so reading the statute to require both would
“result[] in superfluity,” id. at 24, because the claim form only requires
the identification number, which (if accurately supplied) necessarily
provides both pieces of information, id. at 24–25.
As I have just described, I accept the point as far as it goes: an
accurate identification number would supply both pieces of information.
But the converse is never true. Providing just an accurate license type
is to fail to provide an accurate identification number (using any
mechanism other than actually providing an accurate identification
number, of course—the very thing that both the form and the statute
15
expressly request, and what the Court says can be ignored). Giving the
Commission only the license type would, as here, leave the Commission
in the dark about who actually performed the service—either forever or
only after substantial effort and expenditure of time and resources to
get at the truth.
But again, to understand what the statute means, who cares what
the form requires? The form—which an agency adopted after the
legislature enacted the statute—tells us nothing about the statute’s
meaning. I can imagine one scenario where the fact that the form only
asks for the identification number might matter: to a dentist who
accurately listed that number. If the State were to later seek to impose
penalties on such a dentist for not separately listing a license type—which
the form does not request and which is embedded in the identification
number the provider supplies—it might well violate due-process
principles. See, e.g., Mosley v. Tex. Health & Hum. Servs. Comm’n, 593
S.W.3d 250, 262–69 (Tex. 2019) (holding that due process forbids the
government from insisting even on correct legal requirements when the
government has misled the regulated party about what they are). But no
one is holding Dr. Malouf accountable for not separately listing the license
types of the providers at his dental chains—the State seeks to hold him
accountable for not listing the providers’ actual identification numbers,
which both the form and the statute demand.
The Court is wrong to collapse the identification-number and
license-type requirements for another reason: the legislature is perfectly
free to impose overlapping requirements. Creating an account on most
any website requires someone to enter a new password twice, just to make
16
sure there was no typo; overlapping informational requirements can
likewise minimize error or promote accuracy in other contexts. 5 There
are likely many other reasons why the legislature might choose to require
more than the bare minimum to achieve a valid regulatory objective, such
as detecting fraud or other problems (or, more benignly, to ensure that
payment gets to the right place as rapidly as possible). It is beyond this
Court’s authority to invalidate or minimize such a legislative choice
simply because a hypothetical set of facts would render that choice
meaningless. See BankDirect Cap. Fin., LLC v. Plasma Fab, LLC, 519
S.W.3d 76, 86 (Tex. 2017) (noting that “we read unambiguous statutes as
written, ‘not as they make the most policy sense’ ” (quoting Combs v.
Health Care Servs. Corp., 401 S.W.3d 623, 629 (Tex. 2013))); MCI
Telecomms. Corp. v. Am. Tel. & Tel. Co., 512 U.S. 218, 231 n.4 (1994)
(noting that courts “are bound, not only by the ultimate purposes [the
legislature] has selected, but by the means it has deemed appropriate,
and prescribed, for the pursuit of those purposes”).
The real facts here prove the point. Providing an accurate
identification number may turn out to satisfy both requirements—which
5 As further examples of why more information may be required in a form
than what appears necessary at first glance, consider that when a provider
enrolls in Texas Medicaid, he or she enrolls in one of four enrollment types:
Individual, Group, Performing Provider, or Facility. The Medicaid scheme
requires some licensees to enroll in certain categories—for example, “dieticians,
licensed vocational nurses, occupational therapists, registered nurses, and
speech therapists” must enroll in the Individual category. The Medicaid scheme
also requires that some providers enroll in the Performing Provider category
whenever that individual “is seeking enrollment under a group.” Whatever the
underlying policy rationales were, however, it is entirely improper for courts to
vitiate the legislature’s choices simply because they make little policy sense.
Separation of powers demands more; it is the courts’ job to say what the law is,
not what it should be.
17
is presumably why the form was structured as it was—but this case
involves a dentist who provided an inaccurate number on thousands of
claims. There is, in short, no “superfluity” here.
2
Second, the Court invokes the presumption of consistent usage—
that the statute’s disjunctive use of the word “or” in nearby provisions
supports reading “and” to have a conjunctive meaning (or “distributive
sense”) here. Ante at 25–27. Even assuming that the Court correctly
categorizes the various uses of various conjunctions, the canon has no
role to play in this case.
What is the canon? In Colorado County v. Staff, the Court described
it by quoting a statute: “Stated another way, ‘words and phrases that
have acquired a technical or particular meaning, whether by legislative
definition or otherwise, shall be construed accordingly.’ ” 510 S.W.3d 435,
452 (Tex. 2017) (brackets omitted) (quoting Tex. Gov’t Code § 311.011(b)).
Defined terms, for example, usefully allow the legislature to use a single
term to convey a concept that would otherwise require tedious repetition
of lengthy, specialized meanings.
This basis does not remotely support the Court’s theory that
provisions like § 36.002(10), which use “or” disjunctively, make its
construction of the word “and” in § 36.002(8) reasonable. As far as I know,
neither this Court nor the U.S. Supreme Court has ever applied the
presumption of consistent usage to the word “and.” To the contrary, the
Supreme Court recently noted the novelty of such an argument in this
very context, rejecting the applicability of such a presumption “to words
as ubiquitous and . . . sometimes context-dependent as ‘and’ and ‘or.’ ”
18
Pulsifer, 144 S. Ct. at 735.6
Understandably so: “and” and “or” are the exact opposite of words
that take on a technical or particular meaning, whether by statutory
definition, context, or otherwise. As the Supreme Court put it, “[t]he
principle is mostly applied to terms with some heft and distinctiveness.”
Id. (using “principal activity” and “money remuneration” as examples).
Frankly, it is hard to imagine a less technical, substantive, particularized,
hefty, or distinctive word than “and.”
Trying to force a “consistent usage” discipline on the legislature’s
uses of conjunctions, particularly in sentences with highly complex
structures, is also troubling because we know that the presumption of
consistent usage “is so often disregarded,” which is why it “is particularly
defeasible by context.” Antonin Scalia & Bryan A. Garner, Reading Law:
The Interpretation of Legal Texts 171 (2012); see also S.C. v. M.B., 650
S.W.3d 428, 445 (Tex. 2022) (“the consistent-use canon depends heavily
on context”).
Without the “defeasib[ility]” of this canon, we probably could not
have the canon in the first place. The canon, after all, is an aid in
reaching the meaning the legislature actually adopted, not a way for
courts to punish the legislature for being less punctilious than members
of this Court might (sometimes) demand. In other words, we use the
6 To quote the Court more fully, it observed that the petitioner in
Pulsifer “breaks new ground in applying the [presumption of consistent usage]
to words as ubiquitous and (as shown above) sometimes context-dependent as
‘and’ and ‘or.’ ” 144 S. Ct. at 735. Amazingly, the Supreme Court of Texas now
chooses to till the same “new ground,” and even quotes Pulsifer for the general
principle, see ante at 26, without mentioning that Pulsifer pulverized the
attempt to do so here.
19
canon because it makes sense. When the legislature repeatedly uses a
particularized term and gives no reason to think that any of those uses
bear a different meaning, it would disrespect the legislature to interpret
the same word differently. The same is true in ordinary speech. Someone
who says “the president” to refer to the president of the school board six
times in a conversation probably is not referring to President Biden in her
seventh usage—unless context shows a changed meaning (such as the
seventh usage coming an hour later and after someone else starts talking
about national politics).
We thus presume that a use of the same term is purposeful—that
drafters use the same term to mean the same thing, just as anyone
typically does in ordinary speech. But as with ordinary speech, it is just
a rebuttable presumption for statutory interpretation—drafters often
“use different words to denote the same concept,” so statutory context
(with a little common sense sprinkled in) is essential. Scalia & Garner,
supra, at 170 (emphasis added). Forcing the same term to bear the same
meaning when context refutes consistent usage is just as bad as giving
terms different meanings when nothing in the context suggests any such
variant. This defeasibility principle applies even to specialized terms,
much less to a conjunction, for goodness’ sake.7
7 The Court claims that I would “only” apply the canon to terms that have
a technical or specialized meaning, not conjunctions. Ante at 26 n.18. I am
perfectly happy to assume that, in some contexts, the canon of consistent usage
might apply to conjunctions. My point transcends all that: courts should wield
the doctrine carefully. It should never be used to justify adopting an otherwise
acontextual reading of a term. And because conjunctions are so far away from
the kinds of words whose repetition implies shared meaning, they are among the
least likely to warrant application of the canon (or, in Justice Scalia’s term, the
20
Examples of statutes that illustrate the danger of whipping out
the consistent-usage canon on conjunctions are almost endless. Here is
one from the Transportation Code. Vehicles’ headlights must be turned
on “(1) at nighttime; and (2) when light is insufficient or atmospheric
conditions are unfavorable so that a person or vehicle on the highway is
not clearly discernible at a distance of 1,000 feet ahead.” Tex. Transp.
Code § 547.302(a) (emphasis added). Under the Court’s reasoning,
because that section uses the word “or” disjunctively, we can reasonably
construe “and” conjunctively. So drivers must use lights only when both
(1) and (2) are present, even though the law obviously commands that
lights be used in either situation, independent of the other. No rational,
law-abiding citizen would read the statute as not requiring lights in heavy
fog or pouring rain during the day. See State v. Gammill, 442 S.W.3d 538,
541 (Tex. App.—Dallas 2014, pet. ref’d) (holding that, when viewing “the
statute as a whole, the plain meaning of the text imposes a duty to display
lights during two alternative time periods”). But someone convicted
under this provision for a daytime violation now has the prestige of the
Supreme Court of Texas itself to argue that the statute unambiguously
precludes any duty to have the lights on.
Here is another example. Section 48.02 of the Penal Code is titled
“Prohibition of the Purchase and Sale of Human Organs.” Tex. Penal
Code § 48.02 (emphasis added). Under the Court’s reasoning, that title
suggests that the statute would prohibit only the purchase and sale of
organs; a seller of organs who did not first purchase those organs is in the
most likely where the canon is “defeasible”). So I am where the U.S. Supreme
Court is—applying the canon to conjunctions like “and” is unprecedented and
weird, but I need not disclaim the possibility. See Pulsifer, 144 S. Ct. at 735–36.
21
clear. Such a construction makes no sense, as the statutory text (enacted
at the same time as the title) reflects—it punishes one who “knowingly or
intentionally offers to buy, offers to sell, acquires, receives, sells, or
otherwise transfers any human organ for valuable consideration.” Id.
§ 48.02(b) (emphasis added).8 The Court ties itself in knots about this
example, see ante at 28 n.19 (the fourth paragraph of the footnote), but
again misses the point. The same legislature enacted both the title and
the text in the same bill, which shows that the legislature does not feel
quite as bound to “consistent usage” of conjunctions as the Court now
thinks it ought to.
The larger point, of course, is that there are many contexts in
which there is no real difference between “or” and “and.” Because of that
linguistic truism, the legislature was not really inconsistent in any
meaningful way in any of these examples, and certainly not here.
3
The Court also argues that I cannot be right because the statute
does not affirmatively require two pieces of information, but instead
deems “wrongful” a “fail[ure] to indicate the provider’s identification
number and license type.” Ante at 27 n.19. Semantics. The following
two circumstances are not different: (1) affirmatively requiring two
things to avoid punishment and (2) punishing a failure to provide only
one of two things. Either way—because it is in fact just one way, written
differently—the statute here requires both.
8 We have ways of dealing with statutory titles that conflict with
statutory text, of course. But because I do not read conjunctions the way the
Court insists on doing, I find no conflict here. In any event, the point I make is
not substantive—it is about how the legislature uses conjunctions.
22
C
Part of statutory context is what the statute tells us about its role
in a larger statutory scheme. “Purpose” has a bad name because of
“purposivism”—the school of thought that, without any particular textual
anchor, determines at a high level of generality what the “purpose” of a
statute is and then forces the text to conform to that discovered purpose.
See, e.g., Scalia & Garner, supra, at 19; accord Pulsifer, 144 S. Ct. at 737
(“No law pursues its . . . purposes at all costs.”) (brackets omitted)
(quoting Luna Perez v. Sturgis Pub. Schs., 598 U.S. 142, 150 (2023)). This
historic abuse of purpose is unfortunate for many reasons, not least this:
“The term purposivism suggests, wrongly, that its supposed antonym—
namely textualism—precludes consideration of a text’s purpose. This is
not so. It is untrue that a textualist judge must ‘put on blinders that
shield the legislative purpose from view.’ ” Scalia & Garner, supra, at
20 (quoting William D. Popkin, An Internal Critique of Justice Scalia’s
Theory of Statutory Interpretation, 76 Minn. L. Rev. 1133, 1142 (1992)).
To the contrary, purpose, “in its concrete manifestations as deduced from
close reading of the text,” is utterly essential to true textualism. Id.Understood in this light, “[t]he evident purpose of what a text seeks to
achieve is an essential element of context that gives meaning to words.”
Id. (emphasis added); see also Tex. Dep’t of Transp. v. City of Sunset
Valley, 146 S.W.3d 637, 642 (Tex. 2004) (noting that “[w]e also consider
the objective the law seeks to obtain” when construing statutes).
Put another way, a statute’s purpose, as revealed by the text’s
plain meaning, is just another part of the statute’s legal context.
Statutorily expressed purpose can helpfully eliminate alternatives that,
23
while grammatically possible, are deemed unreasonable because they are
inconsistent with the statutory text.
Here, the statutory text reveals both a general and specific
purpose, which provides essential context that requires rejecting Dr.
Malouf’s reading. Start at the top, with the title, which I have mentioned
before. I wholly agree that the “title of a statutory provision cannot
override the plain meaning of the underlying text,” but “a title can at least
‘inform the inquiry into the Legislature’s intent.’ ” Brown v. City of
Houston, 660 S.W.3d 749, 754 (Tex. 2023) (quoting TIC Energy & Chem.,
Inc. v. Martin, 498 S.W.3d 68, 75 (Tex. 2016)). The heading to
Chapter36 of the Human Resources Code was recently amended to read:
“Health Care Program Fraud Prevention.” Act of May 16, 2023, 88th
Leg., R.S., S.B. 745, § 2. It was similar as originally enacted: “AN ACT
relating to the prevention of Medicaid fraud.” Act of May 27, 1995, 74th
Leg., R.S., ch. 824, H.B. 2523. The legislature’s overarching motive is
pretty obvious: to avoid squandering limited resources, it wants to pay
only valid claims, including for treatments performed only by those
licensed under the Medicaid scheme. The State does not want to pay for
unauthorized medical services, and to that end has chosen to minimize
the number of fraudulent claims by making it easier to detect fraud and
by penalizing claimants who make fraudulent claims or claims that
violate the anti-fraud requirements.
The claim-reporting requirements—and the heavy consequences
of violating them—are part of the effort to “prevent” fraud. Knowing the
identity “of the licensed health care provider who actually provided the
service” is one of the legislature’s specifically enumerated means of
24
accomplishing that end. Tex. Hum. Res. Code § 36.002(8). Knowing the
healthcare provider’s identity greatly facilitates determining the claim’s
legitimacy in a cost-effective and efficient manner. According to the
Court, though, the State has really asked for nothing more than a
provider’s license type. The Court’s holding today tells the State (and
providers) that a complete lie—writing in some other person’s number
who did not provide the treatment—is treated as compliance, just as long
as that other person has the same license type.
This reading fundamentally—and, I must say, rather obviously—
botches the whole thing. Giving the State nothing but “license type”—
how many dentists are there?—reads the text in a way that essentially
renders the most important part (the identity of the actual provider) a
dead letter. The Court’s interpretation impedes the statutory goals and
function. Telling the State that it cannot require the provision of an
accurate identification number (so long as the license types match up—
but how would the State know that if it does not know who the actual
provider is?) makes it harder for the State to verify that the claim is in
all respects proper, and makes it harder to conduct the sort of anti-fraud
monitoring that I described above.
The Court’s reading of this fraud-prevention statute is one that
makes fraud easier. The statute’s textually derived purpose confirms
yet again that the only reasonable way to construe the statute is to
require claimants to provide both pieces of information (which, in this
context, is achieved by giving an accurate identification number, but is
not achieved by giving only an accurate license type).
The Court recasts the statutory purpose as “preventing fraudulent
25
harm to the Medicaid program.” Ante at 32. As long as the fraud does
not disrupt the State’s ability “to properly process claims and distribute
the appropriate amount of funds to the provider,” the statutory purpose
is to not punish those who make false claims. Id.
This argument fails for at least two reasons. First, even accepting
the Court’s formulation, the “harm” to the State extends beyond paying
unauthorized claims. The State is harmed because it lacks the
information that it needs to investigate whether a claim is authorized in
the first place. If all the State knows is that some anonymous dentist
did the work, it will either pay up despite being in the dark or have to
determine (1) who actually performed the services and (2) whether the
claim is otherwise authorized. When that inefficient and costly process
is multiplied by the millions of Texans enrolled in Medicaid, it is easy to
see why the legislature demanded information to enable cost-effective
and efficient ways to verify claims or detect patterns of noncompliance.9
So even if the State ultimately determines that a particular claim is
authorized, incomplete or false information necessarily harms its ability
“to properly process claims.”
Second, and more fundamentally, it is quite dangerous to suggest
that some lies are no big deal. Even if the procedures were authorized,
providing false information about them is a big deal, and not only because
of the indirect costs that such false information imposes. Yes, the parties
9 There were over five million people enrolled in Medicaid in Texas in
2022. Tex. Health & Hum. Servs., Texas Medicaid and CHIP Full Benefit
Caseload By Risk Group By County—State Fiscal Year 2022 (2023),
https://www.hhs.texas.gov/sites/default/files/documents/enrollment-by-county-
final-sfy-2022.xlsm (last visited June 18, 2024)
26
have stipulated that properly qualified dentists performed the procedures
in this case. If one conceives of the harm as limited to the delta between
what the State actually paid and what it would have paid had the forms
told the truth, then the lies may have been “harmless.” But we are not
talking about tort or contract damages. The lie here is a violation of an
important regulatory-enforcement law that requires accuracy for different
reasons.
Securities laws enforced by the government provide a useful
analogy. When Congress passed the Investment Advisers Act of 1940, it
“empower[ed] the courts to enjoin any practice which operates ‘as a fraud
or deceit,’ ” but did not “intend[] to require proof of . . . actual injury to
clients.” SEC v. Cap. Gains Rsch. Bureau, Inc., 375 U.S. 180, 192 (1963).
Similarly, the Securities Act of 1933 “makes unlawful the making of
untrue statements of material fact or the omissions of such a fact.” United
States v. Tallant, 547 F.2d 1291, 1296 (5th Cir. 1977). It is the making of
the untrue statement that is unlawful, “not the occurrence of a dollar loss
as a result of the actions, statements, or omissions.” Id. Put another way,
“a violation of the Securities Act is [not] a result-oriented crime.” Cook v.
State, 824 S.W.2d 634, 638 (Tex. App.—Dallas 1991), pet. ref’d, 828
S.W.2d 11 (Tex. Crim. App. 1992) (per curiam). At least for government
enforcement, securities laws protect the integrity of the markets—they
do not just target actual financial losses caused by a lack of integrity.
Likewise, the State legitimately demands that the Medicaid-
reimbursement system be one of integrity. The statute does not just
target general government “fraud”—it requires compliance with specific
27
anti-fraud and fraud-prevention requirements.10 Remarkably, however,
intentional lies to the government that are material to its management
of a massive and expensive program turn out to be of no great importance
to the Court, which holds that § 36.002(8) was designed only to target
those who ultimately receive more money than they would have received
if they had told the truth. (And even this standard seems insufficient to
support the Court’s judgment, because at least some of the false
information supplied on the forms was apparently written to obtain
money more quickly than would have been possible if the correct
information had been supplied—for example, when the actual providers
did not yet even have an identification number.)
* * *
Remember that all of this comes just because of the Court’s root
canal of the word “and.” The foregoing reasons explain why, as a matter
10 But suppose “fraud” as generally defined is the only thing that matters.
As the Court notes, the Government Code defines fraud as “an intentional
deception or misrepresentation made by a person with the knowledge that the
deception could result in some unauthorized benefit to that person or some other
person.” Tex. Gov’t Code § 531.1011(4) (emphasis added). So even if it were just
standard government “fraud” at issue, the general definition still clearly
contemplates that some frauds will not ultimately “result in some unauthorized
benefit.” Id. The punishable act is not the realization of an unauthorized
benefit, but the “intentional deception or misrepresentation.” Id.
After all, unlike private plaintiffs, the government does not need to
establish an actionable injury to bring an enforcement action; the violation of
law is itself sufficient. See, e.g., Universal Brands, Inc. v. Philip Morris Inc., 546
F.2d 30, 34 (5th Cir. 1977) (“A mere showing by the private plaintiff of a
violation of the anti-trust laws has no actionable significance because, while in
a government action there need be established only an antitrust violation, a
private litigant ‘must not only show the violation of the antitrust laws, but
show also the impact of the violations upon him.’ ” (internal quotation mark
omitted) (quoting Credit Bureau Reps., Inc. v. Retail Credit Co., 476 F.2d 989,
992 (5th Cir. 1973))).
28
of statutory construction, it is wrong for the Court to have interpreted the
word as it did.
That my reading of the statute aligns with the State’s does not
mean that I think the statute cannot be abused. Maybe it was abused
here. Dr. Malouf’s strongest point is not his argument that the statute
does not even really care about knowing the “actual” provider. Nor is it
his contention that the lies he told on 1,842 false claim forms did not
really harm the State. His strongest point is instead that because all the
procedures would have been authorized and paid for, the State is using
the law as a sledgehammer against a gnat. Really, the State seeks $16
million for a bunch of forms that were all wrongly filled out in exactly the
same way, none of which was used to actually rob the treasury of even $1
that should not have been paid? I hasten to add that I know nothing of
the underlying investigation or negotiations, but it seems excessive.
The Court’s solution is every bit as excessive. The meaning of
“and” cannot turn on a collateral matter like whether the total penalty for
these false claims was $1,000 or $16 million. The words of a statute mean
what they mean. Giving this statute an excessively parsimonious reading
does not fix enforcement proceedings that are excessively grasping. The
law has other and better ways to solve that sort of problem—assuming it
is one here—without rewriting a statute. Being over-penalized raises
potential claims under our due-course clause or, perhaps most clearly,
under the provision of our Constitution promising that “[e]xcessive bail
shall not be required, nor excessive fines imposed, nor cruel or unusual
punishment inflicted.” Tex. Const. art. I, § 13 (emphasis added).
Dr. Malouf, however, has not invoked these constitutional
29
protections. The Court should not find a backdoor way to give him aid or
comfort when he refuses to knock at the front door.11
One might also argue that the State’s construction harshly
penalizes claimants who inadvertently provide the wrong license type or
identification number via “unintentional technical, clerical, or
administrative errors.” Tex. Gov’t Code § 531.1011(4) (excluding such
“errors” from the definition of fraud). But the statute covers that
possibility, too—it only punishes claimants who “knowingly” make claims
without the required information. Tex. Hum. Res. Code § 36.002(8). In
other words, the legislature included a mens rea element to prevent
penalizing mere negligence—and remember, this is Dr. Malouf’s lead
argument, for which he requests a remand.12 But any desire to hold the
11 I take no position on the merits of any such contention, although I am
quite open to them if over-penalizing is shown. My only point is if anything is
wrong with the State’s case it is not its interpretation of the word “and.”
12 The Court declines to resolve Dr. Malouf’s lead argument because it
rules for him on his alternative argument. See ante at 14 n.13. But because I
reject that argument, I must resolve whether summary judgment for the State
was improper with respect to whether Dr. Malouf knowingly failed to provide
his identification number and license type.
I would mostly affirm on this ground, too. Dr. Malouf testified that he
generally did not know that his clinic was providing false information. This
testimony is probably enough to create a fact issue as to Dr. Malouf’s mental
state—i.e., whether he genuinely did not know that his clinic was filing false
claim forms. As to others, he testified that he knew his clinic was providing
the wrong identification number, but thought that it was doing so based on prior
approval from Texas Medicaid. In other words, Dr. Malouf admits to knowingly
making false claims—he just thought that his conduct was excused. The
statute, however, does not require “specific intent to commit an unlawful act
under Section 36.002 . . . to show that a person acted ‘knowingly.’ ” Tex. Hum.
Res. Code § 36.0011(b). Put another way, the statute does not excuse telling lies
in good faith, so even if we credit Dr. Malouf’s testimony that he thought it was
30
State to its burden of proof has nothing to do with reading the statute to
not require the specified information in the first place.
The legislature struck a careful balance between two competing
objectives: enacting strong protections against fraud while simultaneously
excluding honest mistakes from the statute’s reach. The Court’s
approach disrupts that balance, departing from the fundamental
principle that courts “enforce a statute as written, and avoid construing
individual provisions of a statute in isolation from the statute as a whole.”
Hlavinka v. HSC Pipeline P’ship, LLC, 650 S.W.3d 483, 491 (Tex. 2022)
(internal quotation marks omitted).
D
The Court also indulges in substantial dicta concerning the rule
of lenity—dicta because, amazingly, the Court holds that the only
reasonable interpretation of the statute is the one that I have shown to
be utterly unreasonable, ante at 33, and with only one reasonable
interpretation, there is no work for the rule of lenity to perform. I would
not apply the rule of lenity either—because the only reasonable reading
reaches the result contrary to the Court’s. But if the Court insists on
addressing the rule of lenity, it should at least do so accurately. It is
fine to falsely report that he was the provider, he still knowingly did it.
With respect to the first batch of claim forms (those where Dr. Malouf
allegedly did not know that his clinic was filing false claim forms), therefore, I
would reverse and remand for the trial court to determine which (if any) forms
qualify for the first batch and, as to those forms, let a factfinder evaluate the
credibility of Dr. Malouf’s general denial. As to the second batch of claim forms
(those where Dr. Malouf knowingly lied but believed he had the right to do so),
I would affirm the judgment below that renders judgment for the State. I
disagree with the Court as to both batches, and thus must dissent in the entirety
of the judgment.
31
fortunate that all that discussion, see ante at 7–13, 33, is dicta.13
The Court is mostly right in stating that, in cases of ambiguity, the
rule of lenity requires courts to construe penal statutes narrowly and in
favor of the accused. I also agree that the rule of lenity is not relevant
absent ambiguity. But not just any ambiguity will do; after all, “most
statutes are ambiguous to some degree.” Muscarello v. United States, 524
U.S. 125, 138 (1998).
To the contrary, courts across the country, including this Court,
apply the rule of lenity only in extreme cases. The U.S. Supreme “Court
has repeatedly emphasized that a court must find not just ambiguity but
‘grievous ambiguity’ before resorting to the rule of lenity.” Shular v.
United States, 589 U.S. 154, 167 (2020) (Kavanaugh, J., concurring). In
other words, “the rule of lenity only applies if, after considering text,
structure, history, and purpose, there remains a grievous ambiguity or
uncertainty in the statute . . . such that the Court must simply guess as
to what [the legislature] intended.” Barber v. Thomas, 560 U.S. 474, 488
(2010) (emphasis added) (internal quotation marks omitted). We have
likewise said that the rule applies only when there is “grave doubt as to
the intention of the legislature.” Gulf, Colo. & Santa Fe Ry. Co. v. Dwyer,
13 The Court starts by opining at great length on the rule of lenity, then
agrees that the rule of lenity does not apply unless a statute is ambiguous, then
deems this statute unambiguous, and thus concludes that the rule of lenity
does not apply. Pulsifer reflects the sounder approach—to address lenity only
at the end, and—if the statute is unambiguous—to decline further
consideration of the rule of lenity for that reason, without any extraneous
discursions into that doctrine. See 144 S. Ct. at 737 (“The two possible readings
thus reduce to one—leaving no role for lenity to play.”). The Court’s contrary
approach is akin to a lengthy discourse on the proper headwear for the rain,
and only then looking out the window and seeing the sun.
32
19 S.W. 470, 471 (Tex. 1892) (emphasis added).14 The Court of Criminal
Appeals applies the rule of lenity “when the proper construction of a
statute is in insoluble doubt.” Diruzzo v. State, 581 S.W.3d 788, 802 n.22
(Tex. Crim. App. 2019) (emphasis added); see also Cuellar v. State, 70
S.W.3d 815, 819 n.6 (Tex. Crim. App. 2002) (noting that it is
constitutionally permissible to consider the rule of lenity only when
absolutely necessary to resolve statutory ambiguity). The Fifth Circuit
takes the same approach: “The rule of lenity . . . should be reserved for
those situations in which a reasonable doubt persists about a statute’s
intended scope even after resort to the language and structure, legislative
history, and motivating policies of the statute.” United States v. Orellana,
14 Contrary to the Court’s accusation that this quote reflects “[c]herry-
picking one of our oldest decisions on the topic,” ante at 10 n.11, this Court has
never suggested that we apply the rule of lenity before first applying traditional
methods of interpretation to resolve superficial doubts about the meaning of
statutory terms. The Court’s own citations (which seem to have no great concern
with “old[] decisions”) prove the point. In Estes v. State, for example, the Court
observed that “if doubt existed as to the intention of the legislature in the two
preceding sections” of the statute in question, “that doubt [was] removed by
considering them with [another section] which, in prescribing what shall be a
sufficient indictment for the offense,” revealed the legislature’s intention
“beyond a doubt.” 10 Tex. 300, 309 (1853). We then reiterated that “the
intention of the legislature is manifest by considering the several provisions of
the law in their natural connection as reflecting their meaning one upon the
other,” thus placing “the proper construction of the statute . . . beyond a question
by the application of [the rule of lenity].” Id.“Cherry-pick[ed]” or otherwise, it is hardly an incorrect statement of
Texas law. Unsurprisingly, this Court runs a bit low on cases specifically
addressing what level of ambiguity is required to justify applying the rule of
lenity—we construe penal statutes quite infrequently compared to other courts.
Those that do regularly construe such statutes, however, apply the rule of lenity
only in cases of grave ambiguity. I would not purport to chart a different
course—and especially not in a case where the Court itself admits that the rule
of lenity plays no role regardless of the level of ambiguity at issue.
33
405 F.3d 360, 371 (5th Cir. 2005) (internal quotation marks omitted).
Words like “grave,” “grievous,” and “insoluble”—repeated in case
after case, court after court, century after century—do some real work.
The rule of lenity is not a special gift to one side of the case—a sort of
strict-scrutiny of penal-law construction. It is a “break glass in case of
ambiguity,” an absolute last resort when nothing else helps—not context,
not the canons, not history and tradition, not statutory history. The rule
of lenity means that if it would truly be a coin flip, we do not flip the coin
but instead give the call to the defendant.
But even if there were some ambiguity here, it is not remotely a
“grave,” “grievous,” or “insoluble” ambiguity that escapes meaning. No
coin-flipping needed. To the contrary, the text, structure, and purpose of
this statute resolve the question. I hope that the lower courts do not draw
from the Court’s dicta today that the rule of lenity is suddenly a vibrant
force ready to slay statutory requirements that are otherwise clear.
* * *
“[T]extualists believe that meaning is a function of the way
speakers use language in particular circumstances.” John F. Manning,
The Absurdity Doctrine, 116 Harv. L. Rev. 2387, 2457 (2003). Courts must
construe statutes with the knowledge that “words mean what they
conveyed to reasonable people at the time they were written.” Scalia &
Garner, supra, at 16. So “it is a ‘fundamental principle of statutory
construction (and, indeed, of language itself) that the meaning of a word
cannot be determined in isolation, but must be drawn from the context in
which it is used.’ ” Yates v. United States, 574 U.S. 528, 537 (2015) (quoting
Deal v. United States, 508 U.S. 129, 132 (1993)). This includes statutory
context and purpose, the latter of which “must be derived from the text,
34
not from extrinsic sources such as legislative history or an assumption
about the legal drafter’s desires.” Scalia & Garner, supra, at 56.
The Court does not follow these principles today. Its “textualism”
ignores the statutory context and expressed purposes underlying the
Medicaid Fraud Prevention Act’s language. The rule of lenity has no role
to play here, both because the statute is not ambiguous and because, even
if it were, it would not be triggered as a first impulse.
But if the Court is right and I am wrong, then buckle up. As I show
next, if this statute means what the Court says, then a huge host of
statutes do not mean what anyone has thought they meant.
II
The foregoing reasons are sufficient in and of themselves to merit
my dissent. But the Court’s approach to statutory construction is also
problematic because of multiple Texas and federal laws that use the
terms “and” and “or” interchangeably.15 I mention those statutes for two
15 As noted above, see supra at 5–6 & n.2, it is not just statutes that use
“and” and “or” interchangeably. English speakers in non-legal and legal
contexts have used one word when the other would not meaningfully change the
phrase’s meaning or when the other word would seem better.
To thee and thine hereditary ever
Remain this ample third of our fair kingdom;
No less in space, validity, and pleasure,
Than that conferr’d on Goneril.
William Shakespeare, King Lear act 1, sc. 1, l. 88–91 (emphasis added).
The Texas Pattern Jury Charges also use “and” when “or” could do. For
example, in deciding whether a nuisance is temporary, a jury must determine
if the injury is “such that any anticipated recurrence would be only occasional,
irregular, intermittent, and not reasonably predictable.” Comm. on Pattern
Jury Charges, State Bar of Tex., Texas Pattern Jury Charges: General
Negligence, Intentional Personal Torts & Workers’ Compensation PJC 12.4
35
reasons. First, their mere existence proves the point that I have made
above—that these conjunctions are often used in such a transposable
way. Statutes must therefore be read within their context, which always
“includes common sense.” Morath v. Lampasas Indep. Sch. Dist., 686
S.W.3d 725, 738 (Tex. 2024) (quoting Biden v. Nebraska, 143 S. Ct. 2355,
2379 (2023) (Barrett, J., concurring)). Second, these statutes’ existence
amplifies the threat of today’s decision—if this statute can be treated as
the Court does, what is next? I hope, instead, that today is a one-off, and
that the Court returns to the older tradition of relying on text within
context, “includ[ing] common sense.”
I start with just a few Texas statutes (and mention quite a few
others—but still just the tip of the iceberg—in footnote 16, infra).
Chapter 2 of the Texas Family Code generally forbids county clerks from
issuing marriage licenses “if either applicant . . . fails to submit proof of
age and identity.” § 2.009(a). Applying today’s logic, a county clerk
(2022) (emphasis added). Under the Court’s reasoning today, a temporary
nuisance must be all of those adjectives, even though all need not be found. See
Schneider Nat. Carriers, Inc. v. Bates, 147 S.W.3d 264, 272 (Tex. 2004) (holding
that a nuisance is “temporary if it is ‘occasional, intermittent or recurrent’ ”
(emphasis added) (quoting Bayouth v. Lion Oil Co., 671 S.W.2d 867, 868 (Tex.
1984))); see also Huynh v. Blanchard, ___ S.W.3d ___, 2024 WL 2869423, *18
(Tex. June 7, 2024) (noting a jury’s finding that a “nuisance [was] ‘occasional,
irregular, [and] intermittent’ ”).
Our very own rules of procedure use “and” when context (and this Court’s
routine practice) shows that it means “or.” See Tex. R. App. P. 52.8(a) (requiring
an appellate court to deny relief “[i]f the court determines from the petition and
any response and reply that the relator is not entitled to the relief sought”
(emphasis added)). Under the Court’s reasoning, Rule 52.8(a) would not permit
this Court to deny relief until it considered the petition, the response, and the
reply. But in practice, this Court routinely denies relief without seeing a response
or a reply to the petition for review. Today, the Court holds the legislature to a
higher standard than the one to which we hold ourselves.
36
could issue a marriage license if an applicant submitted proof of identity
but not age (or vice versa), even though the statute requires applicants
to submit both pieces of information (sound familiar?). Id. § 2.002(2)
(requiring applicants to submit “proof of identity and age” in their
application for a marriage license (emphasis added)).
Chapter 312 of the Texas Health and Safety Code applies to
agreements with a “medical and dental unit.” § 312.003. Under the
Court’s reasoning, Chapter 312 would only apply to agreements with a
unit that provided both medical and dental services, even though the
statute is clearly intended to apply to agreements with medical or dental
units. See Tex. Educ. Code § 61.003(5) (specifying that the term
“[m]edical and dental unit” includes both medical and dental schools).
Consider also Chapter 65 of the Texas Parks and Wildlife Code,
which “governs the taking, possession, and sale of alligators.” § 65.002
(emphasis added). Under the Court’s approach, that chapter would
punish someone who took, possessed, and sold an alligator—but one who
took, possessed, and gifted an alligator would be beyond the statute’s
reach. Yet viewed in context, the Code contemplates regulating all three
activities, independent of the others—it permits the Parks and Wildlife
Commission to regulate the “limits, size, means, methods, and places in
which it is lawful to take or possess alligators,” among other things. Id.
§ 65.003(4) (emphasis added).16
16 Besides those discussed in this part of my opinion, many other Texas
statutes contain similar grammatical ambiguities. Here is but a sample: Tex.
Fin. Code § 152.107(f) (providing that a money services licensee loses its
license if it “fails to submit the completed annual report and pay the annual
license fee and any late fee due within the time prescribed by [the statute]”
37
Federal law is no different. For example, 49 U.S.C. § 28301(b)
provides that the preceding section “does not apply to . . . (A) an
independently owned and operated railroad not exceeding one hundred
miles in length; (B) an electric street railroad; and (C) an electric
interurban railroad.” Read in context, the Code does not require a
railroad to satisfy (A), (B), and (C) to be exempted; rather, all three kinds
of railroads are exempt. The Court’s reasoning would rationalize the
(emphasis added)); id. § 392.304(a)(6) (providing that a debt collector may not
use “a written communication that fails to indicate clearly the name of the debt
collector and the debt collector’s street address or post office box and telephone
number if the written notice refers to a delinquent consumer debt” (emphasis
added)); Tex. Lab. Code § 205.016 (providing certain penalties for a
reimbursing employer “who fails to pay a reimbursement on the date on which
the reimbursement is due, or who fails to submit records and reports, as
prescribed by the commission” (emphasis added)); Tex. Loc. Gov’t Code
§ 83.004(a) (permitting a county judge to declare the office of county treasurer
vacant “[i]f a person elected to the office . . . fails to provide an adequate bond
as required by [statute] and to take the official oath on or before assuming the
office (emphasis added)); Tex. Occ. Code § 651.460(a) (providing that “[a]
person violates this chapter if the person . . . (2) fails to retain and make
available to the commission, on request, copies of all price lists, written notices,
embalming documents, and memoranda of agreement required by this
chapter” (emphasis added)); Tex. Parks & Wild. Code § 12.603 (providing that
the Parks and Wildlife Department can refuse to issue permits to applicants
who “fail[] to submit . . . (1) a completed application . . . (2) the required permit
fee; (3) accurate reports as applicable; and (4) any additional information that
the department determines is necessary to process the application” (emphasis
added)); Tex. Penal Code § 21.09(2) (providing that a person commits an
offense if he or she knowingly “fondles or touches the anus or genitals of an
animal in a manner that is not a generally accepted and otherwise lawful
animal husbandry or veterinary practice, including touching through clothing”
(emphasis added)); Tex. Tax Code § 181.001(a) (imposing a tax on one who
“(1) manufactures or produces cement in, or imports cement into, the state;
and (2) distributes or sells the cement in intrastate commerce or uses the
cement in the state” (emphasis added)).
38
opposite result.17
The Court also departs from the reasoning used by courts around
the country. Just this term, the U.S. Supreme Court noted that courts
should resolve ambiguity “by reviewing text in context.” Pulsifer, 144
S. Ct. at 726. The Court rejected Pulsifer’s interpretation not because it
was grammatically impermissible, but because it would “make[] a hash
of the scheme Congress devised.” Id. at 735. Or, as this Court recently
described Pulsifer, “requiring A, B, and C, each, provides a safety valve
for otherwise stricter sentences, while requiring only one or two of the
three would allow more violent criminals to be released sooner and
seriously rupture the sentencing system.” Bd. of Regents of Univ. of Tex.
Sys. v. IDEXX Labs., Inc, ___ S.W.3d ___, 2024 WL 2983170, *5 (Tex. June
14, 2024). Yet here, the Court embraces a construction that ruptures the
statutory scheme by insisting on a dubious definition of “and” even
though only the State’s interpretation is consistent with the statutory
protection against Medicaid fraud. It is hard to read these two opinions
from our Court issued one week apart and make sense of them both.
Likewise, state supreme courts across the country refer to
statutory context and common sense when giving meaning to
conjunctions. See People v. Allen, 968 N.W.2d 532, 538 n.16 (Mich. 2021)
(holding that “the context mandates a disjunctive reading of ‘and’ ”
because a conjunctive construction would have created a result
inconsistent with the rest of the statutory scheme); State v. Irby, 967
17 For further federal examples of statutes that disprove (and would be
threatened by) the Court’s reasoning, see the Solicitor General’s Pulsifer
briefing. She includes dozens. Pulsifer itself quoted 34 U.S.C. § 20101(f) as
an example. See 144 S. Ct. at 728.
39
N.W.2d 389, 395 (Minn. 2021) (“in limited circumstances, ‘and’ can be
read in a several, that is, disjunctive, sense based on context and the
specific way the word is used”). We should too.
* * *
The legislature can fix the specific error the Court makes today.
It could amend the statute to use different language to affirmatively and
expressly require both the license type and the identification number (or
whatever else it may want). But how can the legislature solve the larger
problem that the Court has created—a problem, as the foregoing
discussion reflects, that permeates every part of our statutory law? If
the judiciary is now going to read laws that are as clear in their scope as
this one to mean something quite different than what they say, the
legislature will not only have to scrub all new legislation, but will also
have to race to check all the existing laws. Perhaps a review of legislation
is no terrible thing—but to force the legislature to go on Conjunction
Patrol at the cost of other legislative priorities seems to be a questionable
way to treat a coordinate branch.
Because I disagree with the Court’s articulation of the law and its
result—a conjunction I use with meaning—I must respectfully dissent.
Evan A. Young
Justice
OPINION FILED: June 21, 2024
40