Tuesday, June 08, 2021

The Absurd Formalism of the Mississippi Supreme Court

by Michael C. Dorf

The Mississippi Constitution provides for two means of amendment. The legislature can propose amendments or the People can do so via ballot initiative. State ballot initiatives--for good or ill--frequently change state law in ways that the ordinary legislative process does not. Drug policy is one area where we tend to see a divergence between popular opinion and legislator opinions. Thus, last November, Mississippi voters used the ballot initiative process to amend their state constitution to allow medical marijuana. Or so they thought. It turns out, the Mississippi Supreme Court recently decided, the Secretary of State acted unlawfully by placing the issue on the ballot.

Why? Because the state constitution says that for a petition to successfully place a measure on the ballot "signatures of the qualified electors from any congressional district shall not exceed one-fifth (1/5) of the total number of signatures required to qualify an initiative petition for placement upon the ballot." When that requirement was adopted, Mississippi had five congressional districts, but following the 2000 census, it was reduced to four. If you're any good at math, you'll realize that the state Supreme Court ruling means that not only was the medical marijuana initiative wrongly included on the ballot, but that the ruling effectively wiped out the possibility of amending the state constitution by ballot initiative. With only four districts, there will be at least one-quarter (1/4) of signatures coming from at least one district. And, take my word on this one, a quarter exceeds a fifth.

The Mississippi Supreme Court Justices in the majority are not bad at math. They're just bad at law. They realized that their ruling eliminated the ballot initiative, but they thought that was the inevitable result of the census and the text of the state constitution. Indeed, they had the gall to say that they had to rule as they did, because otherwise they would be changing the constitution. This is the sort of ruling that gives formalism a bad name.

Consider an everyday sort of analogy. Sheila is a shopper for Instacart. She wants to fill an order for  Edgar. The order includes various quantities of tomatoes, onions, garlic, beer, and more. It also says "I'm going to make my famous five-bean chili. Please purchase a total of five pounds of the following kinds of dry beans: kidney; black; pinto; garbanzo; and chickpea. But make sure that no bean amounts to more than a fifth of the total by weight." Sheila puzzles over the order. She surmises that Edgar doesn't realize that garbanzo beans are the same thing as chickpeas. Indeed, the grocery store where Edgar has asked her to shop doesn't sell anything called "chickpeas." All of its dry chickpeas are labeled "garbanzo." She tries to contact Edgar to clarify, but he's unavailable. What should Sheila do?

(1) One option would be to buy a pound each of kidney, black, and pinto beans, and two pounds of garbanzo beans. That would fulfill Edgar's desire for five pounds of beans, even though it would violate his requirement that "no bean amounts to more than a fifth of the total by weight." After all, garbanzo beans now amount to two-fifths. Still, Sheila might fairly say, that's what Edgar really wanted; he just didn't realize that's what he wanted because of his mistaken belief that garbanzo beans and chickpeas are different.

(2) Another option would be for Sheila to get a pound of each of the four kinds of beans. This would violate the no-more-than-a-fifth rule also, because now each kind of bean accounts for a quarter of the total, but it seems to respect the general spirit of Edgar's wish to balance the varieties of beans.

(3) Yet option (2) does not satisfy Edgar's requirement that there be five pounds of beans in total, which is probably important in ensuring the proper proportion of other ingredients. A seemingly better option would be to buy 1.25 pounds of each of the four kinds of beans, which would total five pounds and, like option (2), would also seem to respect the general spirit of "equal beans."

I think either (1) or (3) could plausibly be defended. Maybe even option (2). The one thing that it would make no sense whatsoever for Sheila to do would be to say that she can't comply with Edgar's wishes perfectly--which she cannot--so she won't buy any beans. That is a much greater departure from anything Edgar might want than any of the other approaches.

The Mississippi Supreme Court bought no beans.

Now, hold on, you might say, there are at least two differences between Sheila and the Mississippi Supreme Court. First, Edgar made a mistake. And second, law is not shopping. I find neither distinction persuasive.

It's true that Edgar made a mistake, but so did the drafters of the no-more-than-a-fifth provision of the Mississippi constitution. When they wrote that provision in 1992, they seem to have mistakenly assumed that Mississippi would always have at least five congressional districts. Is it possible that they intended for the initiative process to disappear in a puff of mathematics if the state's representation declined? Well theoretically, sure, but there's nothing else in the text or history of the provision that suggests as much. It seems that the drafters simply overlooked the possibility of a decline in the number of districts. It was at least as much of a mistake as thinking that garbanzo beans and chickpeas are different things.

What about the second objection? One might think that intentionalism is appropriate for a hired shopper but textualism is the preferred mode of construing authoritative legal texts like constitutional provisions. Does that notion distinguish beans from legislative districts?

Nope. For one thing, as the dissenters in the Mississippi case point out, it is possible to read "congressional district" in Section 273(3) of Article 15 of the Mississippi Constitution to refer to the state statutory division into five, not four, congressional districts. To be sure, federal court rulings enjoined the use of that statutory definition for congressional elections, but as the dissenters further argue, that doesn't preclude its use for some other purpose--like making sense of an otherwise unusable ballot initiative procedure in the state constitution. So the textualist case for the majority's ruling is not nearly so strong as the majority thinks.

Moreover, as the dissenters also point out, even if Mississippi had amended the statute with respect to congressional elections, it would still be possible to read the constitutional provision to allow the ballot initiative process to proceed by reading "congressional district" to refer to the 1992 districts, not the current districts. Would that be the most natural reading? Of course not. But it would avoid absurdity.

And that takes us to what might be the heart of the matter. Neither the majority nor the dissents cite the 1892 federal Holy Trinity case, but (if you will excuse the pun) its spirit hangs over the Mississippi case. Textualists routinely denounce Holy Trinity for its pronouncement of what is sometimes called an absurdity doctrine. The U.S. Supreme Court concluded its opinion in that case by saying that where the "whole history and life of the country affirm" that some category of case or cases "could not have been intentionally legislated against . . . [i]t is the duty of the courts . . . to say that, however broad the language of the statute may be, the act, although within the letter, is not within the intention of the legislature, and therefore cannot be within the statute."

Wait! If I'm advocating the strong intentionalism of Holy Trinity, does that mean I'm rejecting the proposition on which Justice Gorsuch relied last Term in the Bostock case in finding that Title VII's prohibition on sex discrimination in employment extends to sexual orientation and gender identity discrimination? There he quotes Justice Scalia's unanimous opinion in the Oncale case, which said that "it is ultimately the provisions of our laws rather than the principal concerns of our legislators by which we are governed." If the Mississippi Supreme Court was wrong to nullify the procedure by which citizens can place proposed constitutional amendments on the ballot, was SCOTUS also wrong in Bostock?

In a word, no. The Oncale/Bostock proposition does not say "do what the text requires even if it's absurd." It says "follow the words of a statute to their logical and natural extension, even if doing so goes beyond the concrete intentions and expectations of the drafters." The Oncale/Bostock proposition is perfectly compatible with a Holy Trinity-rooted absurdity canon that generally gives the text its full even if unexpected meaning but rejects a construction of a statutory or constitutional provision that would render that provision essentially pointless. To be sure, that's not how the absurdity doctrine was used in Holy Trinity, but the textualist critique of the absurdity doctrine goes well beyond critiquing the particular result in Holy Trinity.

Accordingly, I think it fair to read the Mississippi Supreme Court and strong textualists more generally as promoting absurdity.


Frank Willa said...

Admitting that I have not researched the history of events herein, but how in 2000 did someone not pick up on this '1/5' provision and realize there was a problem that needed fixing? So, I suppose my question is rhetorical commentary.

Joe said...

Since there is a dissent by people with some expert knowledge of state law, I will suppose if there is a logical way out, the ability of the people to express their will in a core way was not removed here.

But, the whole thing does not really seem totally absurd.

First, we are dealing with a provision from 1890 Mississippi. As the majority notes, it's possible that the intention (to the degree that matters) was that if there were too few districts, that the ballot measure route would be blocked. We are talking about an anti-democratic period. What if they feared a heavily populated area with the "wrong" people dominating? (This is my supposition; the majority leaves it open.)

Second, constitutions have provisions that over time become problematic. The U.S. Constitution has a few. The provision here worked okay for a hundred years. That is the value of open-ended language. When you provide specific numbers and percentages, you can run into problems. In fact, that might be why a proposed amendment involving House seats was never ratified unlike the eventual 27th Amendment.

Again, if there is a workaround here, go at it. The chance of one seems to be, per Frank's question, a basic reason why it was not deemed necessary to amend the constitution. But, it's possible, I think, the majority is correct. I'll assume not, but such "glitches" are going to arise. You have to fix them.

hardreaders said...

This is another great post. Between "equal beans" and "The Mississippi Supreme Court bought no beans." I think you might have won the internet for today. Were you also channeling Justice Kagan's "blame the bean" quip from Bowman v. Monsanto? It's certainly nice to have a change of scenery from the usual fare of SCOTUS and other federal decisions. When's the last time DoL covered a state case before this one? It seems like it's been a while.

I got sort of strong King v. Burwell vibes from this case. While there are obviously distinctions between the former being federal/statutory and the latter being state/constitutional, I think the bad faith/disingenuous formalism arguments made by the challengers in each case are quite similar. Likewise the preposterous idea that the schemes (constitutional or statutory) in question contain hidden "self-destruct mechanisms" that can suddenly be triggered by a certain set of facts. I'm maybe a little surprised none of the dissents cited King at all, as it would still be relevant persuasive authority even in a state constitutional case.

While as I said, it's a great post, I have to disagree a little that the "heart" of the case is absurd results flowing from the majority opinion, although certainly that's the outcome. The "absurdity" doctrine was discussed only in passing in one of the dissenting opinions. Rather, I see the core issue as a different 'a' word: ambiguity. There's unquestionably a glaring ambiguity created by the interplay of the constitutional provision and the changed circumstances of the 2000 reapportionment. If we were discussing estate law, that would be a considered a "latent" ambiguity. The bean hypo in contrast is a "patent" ambiguity because it appears on the face of the document (i.e., Edgar's order). Anyway, the type of ambiguity isn't really important. What matters is that there is one, and so you must choose between competing interpretations (e.g., Sheila's options (1)-(3)), potentially considering extrinsic evidence in the process. Absurd results is certainly an important heuristic when evaluating the different options, but again, I'd say the critical point is recognizing the existence of ambiguity. Framed that way, it's not so much that the majority ignored the potential for absurd results as it just denied that the provision had any ambiguity in the first place. But that denial was totally unconvincing, because the majority immediately started consulting dictionaries, which are just a kind of extrinsic evidence. I also laughed at the majority's citation to a then-professor Barrett article as some kind of support for its position. Were they aware that her very recent opinion in Van Buren v. U.S. went out of its way to acknowledge (and resolve) an ambiguity in the statutory text under consideration?

Well, at least according to the article linked below, it seems like people in MS are not taking this decision sitting down. Hopefully a solution can be obtained.


[Speaking of numerical requirements, I will follow up separately with replies to Frank and Joe so as to avoid running afoul of the character limits.]

hardreaders said...

@Frank Willa
There's no particular reason at the time of the 2000 reapportionment that anyone would have made the connection between that and an otherwise completely unrelated constitutional provision that just happens to include a reference to Congressional districts. And, from what I can make of the data on Ballotpedia and the 2009 AG opinion discussed in the case, no ballot initiative was proposed for years later until 2008. When that happened, the "problem" was properly recognized and the then-AG conducted an initial analysis, but there was no reason to do so any earlier than that.

First, and this isn't my area of expertise either, but I'm not sure you're right on the basic facts. While I understand the MS constitution is still referred to as the constitution "of 1890" it appears the relevant provision was introduced or at least updated over a century later—in 1992. Or so says the AG opinion letter as linked below. I also note that MS didn't even start having 5 districts until 1960. (It also had 5 at times prior to 1890, but that's obviously not relevant here.) So the anti-democratic theory just lacks any support.


Second, I'm not sure what U.S. constitutional provisions you allude to since you don't mention them, but if you're comparing a hard numerical constraint such as the minimum age for the presidency, I don't think that works. Nobody would say there's any ambiguity in what it means to be 35 years old—standing alone—such that we have to interpret it in a way that nudges the number up or down a few years to achieve the requirement's purpose. Likewise, nobody is claiming that "1/5" by itself is ambiguous. (I'm really trying not to make a joke about the 3/5 compromise!) Rather, as discussed at length in the dissents, the OP, and by me, the ambiguity is in the term "congressional district" in light of the rest of the provision, its purpose, and the changed circumstances since 1992—and consequently how the 1/5 requirement should apply, given that ambiguity. Honestly, it's just not that complicated; except that you have lawyers involved. Even a moderately intelligent high schooler could recognize that "1/5 of the total number of signatures" is just a convenient shorthand for something like "the total number of signatures divided by the number of the congressional districts". Unlike the 35 year requirement, there's nothing intrinsically significant about the 1/5 number. It's just a point-in-time snapshot of the underlying ratio, which carries the actual significance.

Joe said...

it appears the relevant provision was introduced or at least updated over a century later—in 1992.

The majority opinion specifically references the 1890 date, how long it was in place etc. The debate seemed to me over the "air" in the provision. Prof. Dorf also didn't seem to suggest it was that late.

The AG letter refers to the 1890 Constitution. It says "The current initiative procedure was proposed by Senate Concurrent Resolution 516 in 1992, when Mississippi had five congressional districts" but still going under the 1890 Constitution. It puts forth a way to apply the constitutional rule, from what I can tell, to avoid making it impossible. Again, I'm fine with that if there is any reasonable way to do it.

The letter also in part notes "initiative has broad support throughout the state and to help assure that the initiative process is not used by citizens of one part of the state to the detriment of those in another." An argument can be made that the 1890 anti-democratic ethos (I'm just offering it as a possibility) set FIVE as the minimum there. Let's say Mississippi had three reps. "Throughout" the state with five covers more diversity than if there is only three.

claiming that "1/5" by itself is ambiguous

I'm not sure. There is a dispute on what it allows.

I'm not sure what U.S. constitutional provisions you allude

I think you can cite various ones, depending on one's sentiments.

The U.S. Senate, e.g., made more sense under the current rule when population ranges were much lower and democratic concerns were less present. Some argue the 2A is sort of obsolete given the development of the "militia." The practices of agencies in the 21st Century is arguably stretching the set-up of the Constitution. (Legislative vetoes, e.g., might make sense, but then we had Chadha). I think other examples can be cited, including before later amendments changed the situation. Slavery in the 1850s etc.

Unknown said...

Isn't there a persuasive argument that this decision is right but that it violated the state constitution for the legislature to create a situation that effectively voided the constitutional right of the people to amend the constitution?

Michael A Livingston said...

Even with five districts, wouldn’t the number of signatures in each district have had to be exactly equal to avoid exceeding one-fifth? That can’t have been the intent.

Asher Steinberg said...

So the court says:

"The effect of the twenty percent cap, tied to congressional districts,
was that no more—and no less—than twenty percent of the total number of required
signatures must come from each congressional district."

But that's not all. The provision also prohibits turning in signatures in excess of the required number. The fifth cap isn't based on the total number of signatures turned in, but the minimum required. Even if each district submits an equal number of signatures, each would violate the provision if the total exceeds the required minimum.

Asher Steinberg said...

Sorry, to be clear that's a reply to Mr. Livingston.

hardreaders said...

The below is most of my response, but I have to separate out the last part into a different comment.

I don’t think it’s worth dwelling on dates a lot, but I certainly recognize the starting point for the current MS constitution was indeed the version adopted in 1890. But that version has since been amended over 100 times. And as I said, MS had 7 districts in 1890, so there's just no way the 1/5 requirement can date back that far. Your theory of democracy suppression is interesting and would of course be concerning if true, but you still haven’t put forward any actual evidence either to support it or to contradict how the AG letter characterizes the purpose of the requirement. Common sense alone dictates that if such a nefarious purpose were truly lurking in the background, it would have been mentioned at least once in a combined 43+ pages of majority and dissenting opinions. But it wasn’t.

Maybe I was a little confusing on 1/5 the first time, so let me try and clarify a bit. Nobody disputes that, in a vacuum, it’s the ratio of 1 to 5. But in context of the provision and after reapportionment, it is ambiguous, as I said. It certainly presupposes that at the time (which as I’ve argued, was at least after the 1960 redistricting), there were 5 districts. There’s no other reason I’m aware of for why 1/5 is special, other than it corresponds to the inverse of the number of districts. And as I noted, it’s a convenient placeholder, given that spelling out the ratio in prose would be quite cumbersome. Finally, there’s no suggestion anyone wanted the provision to suddenly expire after a redistricting that happened to reduce the number of seats. So yes, when that number unexpectedly decreases, thereby bucking the historical trend that had persisted for decades (since 1960), the reason for using the particular 1/5 fraction and its application as part of the entire provision do become ambiguous.

Again, I think reference to estate law can be somewhat instructive here. Courts resolve these sorts of latent ambiguities in wills all the time. For example, someone who has 5 children might draft a will with a provision that gives 1/5 of that person’s IBM shares to each child; however, the actual number of children is not expressly stated anywhere in the four corners (no pun!) of the will. Then, suppose one of the children predeceases the testator, but the will isn’t updated to reflect that. Maybe that’s a bit unrealistic, but let’s say the impact on the will simply didn’t occur to anyone somehow. So, when the estate assets are ultimately distributed, will the 4 surviving children still get 20% of the shares each, totaling only 80% of the entire amount, while the remaining 20% sits there idly? Or will a court look at the external facts of there being 5 children initially, followed by the passing of 1 that was not accounted for, and realize that correctly resolving the ensuing ambiguity means giving the surviving children equal portions of the total share amount, i.e., 25% each? Admittedly, I haven’t searched for a case like that, but I strongly suspect a court would do the latter. If courts can do that in low-stakes situations like estates, I’d hope they can do the same for constitutional interpretation when so much more is riding on getting the outcome right.

hardreaders said...

Below is the final portion.

If you have anything to say about my contrast with the presidential age floor, let me know. Likewise if you want to cite a different numerical requirement you think is a good comparison with the MS provision. I will say I don’t see 2A as a particularly good example. Maybe some will disagree, but I’d argue that becoming obsolete or irrelevant due to changed circumstances isn’t at all the same as the change rendering a provision ambiguous. Yes, traditional state militias have gone the way of the dinosaur, but that particular circumstance doesn’t somehow make the 2A text ambiguous all of a sudden. (Of course, there’s plenty of debate about whether it’s ambiguous for other reasons, but this isn’t one of them.) It just means it has—or should have, anyway—a lot less relevance in modern life. And nobody ever promised that every constitutional provision would remain relevant for all time. With regard to changed facts actually making a provision ambiguous, there might be some examples, but I don’t know of any offhand.

hardreaders said...

Another thing occurred to me. I don't think this was discussed by the majority or dissents either. You can also get ambiguity, or in Prof. D's view, absurdity, under the majority's approach if the number of districts ever increased past 5. Then, instead of making initiatives effectively impossible, it makes them easier than they should be. For example, with 6 districts, you get the ability to completely ignore a district of your choosing when collecting signatures. With 7, you can ignore 2 entire districts, and so on. Not only does it give signature collecting an unwarranted advantage, but it also in fact gives rise to Joe's suggested "anti-democratic" scenario. The folks in the bypassed districts would miss out on the ability to participate in the initiative process (at least the proposal aspect) and lose some corresponding amount of representation.

@Prof. L.
As noted by Asher, they do have to be exactly the same. The AG opinion letter I cited above explains that the intent behind this "is to help ensure that an initiative has broad support throughout the state and to help assure that the initiative process is not used by citizens of one part of the state to the detriment of those in another."

I read the prohibition on excess signatures a little differently. The provision states that "the excess number of signatures from [a] congressional district shall not be considered ... in determining whether the petition qualifies for placement on the ballot." So it doesn't seem like a petition will be rejected altogether if the signatures from one or more districts exceed the cap. Instead, those excess signatures will just be disregarded. Assuming that's right, there would be no harm in running up the total in a particular district, other than unnecessary time and expense, as long as the totals from the remaining districts also satisfied the floor amount. That seems to make sense, because then you don't have to worry about balancing out the signatures among districts perfectly on the front end. You can collect a little bit more than you need as a buffer, with any leftovers just being ignored in the final tally. I'm sure people also want a certain amount of buffer to withstand inevitable challenges to some of the signatures.

Joe said...

MS had 7 districts in 1890, so there's just no way the 1/5 requirement

I don't understand this. It very well can. Prof. Dorf referenced the assumption that there would always be at least five districts. But, there could be more.

The majority does not reference to some later amendment. I don't see it in the dissent either. The majority at one point suggests possibly the provision was intentionally set up to burden the initiative process to some degree. Making it 1890 is an easier case there to some degree, but it is far from necessary. Could be much later & the same possibility might hold.

The bean example to me is not really a good one. There, you have an item that is known by different names. "Liz" and "Beth" (Elizabeth) can be the same person. Here, it is more complicated. And, the assumed rule here need not be bad. There is a reasonable concern about having diverse support throughout various communities. Five districts as a minimum might be a rough way to do that. Again, imagine if there is only three.


I also noted that there can be specific details in a constitution that over time become out of date and/or lead to problems. Your response appears to want a more narrowly on point example. I'm not going to worry too much there. Focusing merely on state constitutions, they repeatedly are much more detailed than the U.S. Constitution, and sometimes that will result in lack of wiggle room as facts change.

I think in theory this can be such a case though as noted there seems to be a reasonable enough workaround for it not to be so.

Asher Steinberg said...

Yes, hardreaders, you only throw out the signatures from that district. But if you throw them out, then one of two things is possible: the remaining four collectively satisfy the minimum, in which case at least one of those four must account for more than one-fifth of the minimum, or they don't. The only way to make things work is to submit exactly the minimum number of valid signatures and submit exactly 1/5 of the minimum from each district.

Accordingly, I'm inclined to think this provision was an absurdity from the beginning and that the court here hasn't made matters much worse than they always were.

hardreaders said...

At the risk of belaboring the dates issue yet again, I'm just going by the underlying facts and documents, not what was said in the opinions about the history and progression of the provision.

Wikipedia has the historical number of districts, and there's no reason I can think of to doubt the data's accuracy. In 1890, there were 7, not 5, districts.


And I managed to dig up the unamended 1890 version. You can see for yourself (in Section 273) it doesn't even have an initiative process. So that had to have been added later.


Further, Ballotpedia shows that Section 273 was amended repeatedly during 1958-1992. Referencing the Wikipedia data again, that corresponds almost perfectly to the 1956-1990 timespan when the number of districts remained constant at 5 for multiple decades. So, I take this evidence as strongly indicating that the 1/5 in Section 273 simply reflected an assumption that the number of districts would always be the same. Given how long it had been stable at the time—in 1992, when the most recent amendment happened—I don't think it was a terrible assumption either.


With all respect to Prof. D., I do agree with you on the bean example. For what it's meant to show, i.e., absurdity, I will say it works on its own terms. But as far as a good match for what's happening in the MS case, I agree it doesn't reflect the same kind of ambiguity. As I said above, the bean example is a patent ambiguity. My estate law example I think lines up better with the latent ambiguity in the 1/5 provision.

Yes, of course I agree some provisions can become obsolete or unworkable over time. That's sort of a general statement. I don't really see how it shines much light on the specific issues here.

hardreaders said...

I'm just not sure that's correct. You seem to suggest any district that comes in over the floor will have all signatures for that district knocked out. Although I agree in that case it would be fairly unworkable, I don't understand the provision that way. To me, it indicates that only the excess signatures will be disregarded, which is different. So, while you have to at least reach the floor in each district—which, as pointed out by the AG letter, fulfills the purpose of showing broad-based support—there's no danger if some of the districts also surpass the floor. Any excess just gets ignored. It also makes the process reasonably workable.

Joe said...

Wikipedia has the historical number of districts, and there's no reason I can think of to doubt the data's accuracy. In 1890, there were 7, not 5, districts.

I looked up this in the beginning myself.

My latest comment, to be clear, was that a 1 in 5 rule can make sense even if there was seven districts.

the unamended 1890 version

I appreciate the clarification, since as I noted, my reading of the opinion and so forth was that it wasn't clear. But, as I said, I'm not sure how much this changes the bottom line of what I'm saying. I'd just have to reword it a bit.

simply reflected an assumption that the number of districts would always be the same

Quite possibly. To quote you, "some provisions can become obsolete or unworkable over time" ... such as a provision that assumes there would be always the same number of district. If the number changed, the result might be anti-democratic as here. It is not an absurd result for a court to say "sorry, we are stuck with it."

The only issue here is if the specific provision is so problematic. That depends on the specific text involved. It seems the court should have not said "sorry." But, in some other case, they would rightly do so.

Asher Steinberg said...

Sorry, hardreaders, you're correct. Only excess signatures are discarded.

hardreaders said...

Well glad to know I’m not losing my marbles quite yet!