Oral argument — Melanie Glasser v. Hilton Grand Vacations Company

Glasser v. Hilton Mr. Keogh. Thank you. May it please the Court. Keith Keogh on behalf of Ms. Glasser. When the $15.9 million telemarketing calls were made in under six months in this case, all that predictive dollars and list-based dollars were ATDSs and violated the TCPA. As such, Hilton Vacations contrived a system to mask their predictive dollar and list-based dollar. They did so by a computer creating a list. That computer sent the list to an office in Florida and all the office in Florida did was pass it through by clicking a button. The people in Florida had no ability to make a call. They had no phones. They couldn't even control the pacing. They couldn't control how many clicks they made. That was all done by computer. All they did was pass through a computer-generated list to Kentucky and in Kentucky, a computer automatically dialed those calls. And then if an agent was available, they connected to a different place. We had a call center for an agent or they hung up on them. They're unattended calls. No human involvement in the calling. When this case was filed in April 2016, there was no doubt that that system was an ATDS. Since then, we recognize the landscape has shifted a bit. The D.C. Circuit and ACA vacated a portion of the 2015 SEC order regarding ATDS because that portion was inconsistent with prior orders. Now importantly, ACA held that either may be permissible for certain portions but not both. And afterwards, some courts have held the prior SEC orders were vacated, at least implicitly. Other courts, including the court in this case, held they were still valid and followed them. Are you defending that position? I am defending that they're still valid. I'm not defending the way the court applied it because we believe the court misapplied the SEC orders in the plain language. To be clear, Ms. Glasser prevails under either. I mean, I thought the consensus was growing that the D.C. Circuit decision kind of wiped the slate clean. But you're not so sure about that? I am sure about that, Your Honor. I think the consensus is not growing. I think there's more cases that held they're valid than there are that held invalid, including the Ninth Circuit in March. Yeah, I'm referring to the reasoning of the FCC in 2003 and 2008, that we really don't have to worry about that. But you think we should worry about that? I think so, Your Honor. I think it's consistent. And I think ACA, when it vacated a portion of the 2015 order, it got rid of the SEC order. Has the government taken a position since the ACA D.C. Circuit decision? So are there some government briefs out there that explain how this should work today? Not that I'm aware of, Your Honor. So we have courts doing it. And once again, you know, turning to, we think it's error because... And I take it the same is true of any regulations or guidance. There's nothing the government has done since the D.C. Circuit decision, telling courts, litigants, how to construe this language. No, Your Honor. Actually, the FCC has been thinking it over for some time. And whether there's an opinion on the horizon, you know, in six months or five years, no one knows. So there's been no other guidance besides that. But as we pointed out in our briefs, either under the FCC or plain language. Now, turning to the plain language, in 1991, when Congress enacted the TCPA, it heard testimony that 30 to 40 percent of telemarketers were using list-based predictive dollars. So you know, one thing I've seen in some of the briefs, just trying to figure out, is it doesn't sound like there was a lot of this litigation between 91 and 2003. Big increase after 03, quite understandably. But if the plain language is so good, why were there so few of these cases? I mean, you know, lawyers are known for being pretty creative and clients often go along if there's money. Was this just a negligence? They didn't realize all the opportunities? Well, Your Honor, I think when it's heard... My wrong is the predicate of what I'm saying wrong. There really was a lot of litigation between 91 and 03? No, I think you're correct as far as there's a lot of litigation revolving the junk fax portion. And quite frankly... No, but how about the store, you know, stored list, called list? I mean, I would have thought there were lots of... How else did debt collectors do this? They have a stored list and they call the list. I mean, Congress wrote the statute for a reason, right? That was an existing problem. And that's our point, Your Honor, is it... And even when these cases were, you know, I guess in 2002 is when they became more prevalent, Your Honor, but... But I guess I heard your answer then to be debt collection. The debt collection statute is the one people were using as opposed to the TCPA. That's part of your answer? I mean... Not for telemarketing. Remember, this case is telemarketing. I know. I know. I'm still trying... I thought I'm trying to help you. I thought you were giving me an answer from 91 to 03. The reason the TCPA wasn't being used for cases where debt collectors had lists and then robo-called the list is because it was all being done through a different statute. That's not your answer. No, I think in part, Your Honor, that's correct. But also you need to remember that the use of cell phones has also exponentially grown. So they existed in 91 when this statute was created. They have since 2000... I think I get received my first cell phone in 2099 maybe. So I don't know how many people in this courtroom had... The statute, I mean, it's so quaint with paging service, specialized mobile radio service. I mean, all that stuff was being used. In other words, I'm making the point, you're right, the technology has changed, but the original statute captured what the technology was in the 90s. Which explicitly said cellular phones as well as it had, you know, her testimony regarding list-based dialing. And that's why it captured, you know, list-based dialing as well as... It had perhaps something to do with public service commissions and do-not-call lists, which were more effective in an earlier era than apparently they are today. I don't... Remember the... At least based on my experience with my cell phone. Well, Your Honor, I think it was the first public do-not-call list was overturned and then Congress immediately enacted that. It was overturned in Oklahoma, I believe. But... So I don't think they were effective at the time. But in 91, they became effective and that's, you know, the court said people aren't following them. So the fact that people are following the law more, there's more litigation, does not mean that in 91, when it was created, this didn't cover that. There's just an explosion of, you know, violations. And the FCC, this is the number one complaint in the FCC are these robocalls. And every year, it actually increases. How do you address the problem and maybe just accept it that if we adopt your approach that, you know, it's all you have to have is a stored list and then it being called. And given the FCC's interpretation of capacity, or at least the DC Circuit's interpretation of capacity, if we're going to stick with that, that this really puts a lot of iPhone cell phones in play. So is your answer to that tough luck, that that happens every now and then, a statute needs to be rewritten? Not at all, Your Honor. Or is it just a false threat? I think the ACA, as well as Marks, recognize that your cell phone doesn't automatically dial lists. And that's the issue. Because even your smartphone can, there's apps, if you search on your smartphone, random number generator, 15 things pop up and you can modify it. So you can modify a cell phone to automatically dial stored numbers, and you can modify a cell phone to randomly generate numbers. So you can't ignore one because there's a potential liability of people modifying it. And we even cited a case where someone was mass requesting 30,000, you know, texts at a time from a cell phone. But the fact that they can be modified does not take out the protections that the TCP applies to predictive dialers and dialing from lists. They dialed almost 16 million numbers. How does it apply to somebody, I forget how this feature works, obviously I don't know how to use it, but the one where you're driving, you don't want to be bothered, so every call generates, how do you respond to that? Is that just covered? It's not covered. That is a one call, and that is not dialing from a list. That is a one call going out. That is used for a piece of... It doesn't, the statute doesn't say list. It says store. Plurals. It says numbers. It's plural. We made that a point in our brief that the statute explicitly says dialed numbers, plural, not singular. So a singular auto response would not be covered. The plain language of the statute would not cover that. So, and I think the problem is, is kind of like throwing a baby out of the bathwater type thing. You know, is there, looking at these extreme examples, you know, that don't take place, you're going to let off a company that makes 16 million robocalls in under six months. And I think that defeats the congressional purpose. You know, the Dish Network case on the Fourth Circuit had a very poignant quote when it said, it's, you know, disheartening or dispiriting to interpret it in favor of telemarketers who are meant to be regulated at the expense of consumers. We're not looking at the context. And that's what the U.S. Supreme Court says is when looking at modifiers, whether random or sequential applies to store, you have to look at the context. And the context here is Congress' testimony that list-based dialings were a problem in 1991. They had the word store. If you read the word store as they're arguing, it's superfluous, which the Supreme Court has been very clear you cannot do. Store has no other function besides produce under their argument. And the Bailey case in the U.S. Supreme Court said you have, when Congress uses two different words, you have to give meaning to both words. Why is that true? I mean, can't you have something that helps you produce numbers randomly and then you have to store them before they're called? Why does that take out the meaning of the word store? Because if you apply the modifier to store, it's not produce randomly and then store. It's store or produce, or. So if you took out the other half, store using a random or sequential modifier, it doesn't exist. Even the Dominguez case, the Dominguez-Yahoo case in the Third Circuit pointed out that they have a problem. They can't reconcile store using a random or sequential number generator. But what about the possibility of using a random or sequential number generator as modifying the created and stored numbers to be called? No, Your Honor, because that's as Marx dealt with that and said that. I actually thought Marx did not deal with this issue, actually. I thought that was the potential defect in the decision. I didn't see it coming to grips with this possibility. In fact, I don't even know if this possibility was argued to Judge Okuda. Well, I think it dealt with it and said the holding is the automatic dialing is what is key, Your Honor. And the issue here is given the context, which the Supreme Court gave it the context, don't let grammar rules, whether it's a comma or or, trump the purpose and the context in the plain language. And that's what, you know, Hilton cited Justice Scalia and that dealt with the same thing as the context controls. And I realize that I'm over my time, so. Until right, you saved your five minutes for rebuttal. You were answering questions. We'll hear from your opposing counsel. Mr. Dvoretsky. Good morning, Your Honors, and may it please the Court. Shai Dvoretsky with Jones Day on behalf of the appellee. The appellant's interpretation can't be squared with the statute's text, its context and history, or its proper constitutional scope. As the Court was asking, the appellant's interpretation renders every modern smartphone an ATDS because every modern smartphone can store and then automatically dial numbers. The Court should instead read the TCPA, as Congress provided, to require the to require the capacity to use a random or sequential number generator. As a preliminary matter, the Court's free to interpret the statute for itself. On that point, we do agree with Mark's, as the D.C. Circuit explained in ACA Intervention. Can you use one of those random and sequential generators to store numbers? I think you can conceive of storing numbers using a random or sequential number generator. That could be distinct from producing a lot of... There's a lot of petitions that have been filed recently in the Supreme Court, and certainly a lot of folks are telling the Supreme Court that you can do that. That's right, and I think that is a way to understand store. When you use the random or sequential number generator, it generates the number and then stores it and then dials it later. That's what Congress had in mind. And if you look at the context and the background of this statute, hundreds of pages of context and history showing that what Congress was concerned with was the particular problem of random and sequential number generation. At the time in 1991, you had these automatic dialers that generated huge numbers of random or sequential numbers and just dialed them. It's like blitz dialing to tie up cell phone networks and bring them down, to reach unlisted phone numbers, to reach hospitals, 911 lines, things like that. And that's the problem that Congress was targeting, and that's why it wrote an ATDS definition that has as its dominant phrase, using a random or sequential number generator. How long after a clicker agent presses the make call button does it take for the call to be dialed? Can it be done immediately? It's virtually instantaneous, yes. And in that sense, it's very much like, for example, if I use my iPhone and I tell my iPhone to make... Is it possible that it may not be done immediately? I think, I don't think it's technically possible, I suppose, but the way this system works, I think is undisputed, that the make call button is clicked and that immediately initiates the call and that this all happens instantaneously. I think the best citation for that is in the Supplemental Appendix at 38 to 41. But the way this all works is that it's instantaneous dialing when you click make call. And there's no random or sequential number generation involved with any of this. And so the reading that the TCPA can be understood to cover just dialing from a list is inconsistent with the text. It makes no sense. So just, I mean, I have to say, I find it a hard statute to read for what it's worth, but that may say more about me than the statute. I mean, the generator does connect with produce more naturally than store. I mean, so that's what bothers me. So I know what the rules of grammar are. I know when these things apply and don't. I get all that stuff. And that's why they're right. We do look at context to try to make sense out of it. So generator does suggest it's about the producing function, unless you say that whole clause, the using clause, is just modifying what's to be called. And then I find that a little funny, too. I mean, I mean, what's being said, you use this. In that sense, it's regenerating what's going to be called next. Is that what's going on? So I don't think anybody would say that this is the most clearly written statute in the world. So I don't feel stupid. Good. Thank you. So what's your preferred approach to it? Is your preferred approach to it to say, pull the storing together with the clause? Or is your preferred approach, in your view, the fine. We store and produce these telephone numbers to be called. In a weird sort of way, forget how we even do it. But then when we get them ready to be called, I guess when we put them in the queue, that's when the clause comes into effect? So I'll answer your question, of course, and tell you what my preferred approach is. But I also think that the overarching point is that either of the two approaches that you've laid out are far more defensible. Get rid of Armageddon. Get rid of the Armageddon scenario. Are far more defensible than reading this to constitute list dialing. Now, if you're asking me what my preferred approach is, I think probably the better reading as between the two is that using a random or sequential number generator modifies both store and produce. But I think you also could read it to modify to be called. If you read it that way, then what it's doing is it's describing a quality of the numbers that an ATTS has to have the capacity to store or produce. But I think the better way, consistent with rules of grammar, is that this post positive modifier modifies the entire series of verbs that came before it, especially when you have the comma there indicating just as much. And if you look at the Marx interpretation, the only way that the Marx court read this differently was to completely rewrite the statute. I mean, what the Marx court said was it read this language to mean, quote, to store numbers to be called, or two, to produce numbers to be called using a random or sequential number generator. So it had to separate these two paired verbs, store or produce. It had to take their shared object and shove it between the two verbs. And then it had to add a second copy of that same shared object a second time in order to make clear. I get that. But your preferred approach is to store using a random or sequential number generator. That's funny, too. I mean, that is your preferred. That's why I was trying to get away from that to be called. And the way that works is because I guess it's true. If you produce, you're going to store. That seems incredibly right. It's weird to think of producing and then not storing it. But then you ask, why did you sequence store or produce as opposed to produce or store? I actually don't think it's weird to think of producing and not storing. The way Congress might have thought of these devices, and bear in mind it was 1991 and computer technology was early on and this statute was not written by a bunch of computer scientists. The way Congress might have thought of this is that the machines could work one of two ways. They can produce and immediately dial. So produce, dial, produce, dial. No need to store. Or you can produce a whole long list, store it in the machine, and then dial it automatically. You're covering the waterfront by saying that you can produce it using the random or sequential number generator and then immediately dial. Or you can store it using the random or sequential number generator. But either way, the dominant phrase that's in this statute is using a random or sequential number generator. And to read this statute just to cover any sort of a list dial, list dialing scenario, writes that out of the statute. It also creates the constitutional problem with smartphones. In addition to that— Mr. Sarek, can I just ask you—I'm sorry to interrupt. You're developing my momentum. I hated that when I was interrupted when I was developing the momentum, but it happens. Were there any devices in 1991 that would—if you knew about the device or you knew about the commercial practice, you would look at these words and go, aha, I know what they were thinking about. Do you understand what I'm saying? Were there types of devices or commercial practices, robocall practices that make these words a little more sensible, even though today they don't seem as sensible? Well, I think the devices and the commercial practices reflected random and sequential number generation, which today is not something that's sensible because nobody does that today. Nobody has any reason anymore today to randomly sell anything just by making random calls. Instead, the way marketing tends to work today is through targeted marketing with lists. In 1991, when you were dealing with computers that—an entire large computer in 1991 had a fraction of the memory of today's iPhone. The ability of machines then to engage in the kind of targeted dialing that is commonplace today was simply not as widespread, but the machines that did exist there generated random and sequential numbers and then called them in mass quantities, and that's the problem that Congress was targeting and the reason that it would have used this kind of random or sequential number language. Is store or produce the most artful way to say it? Perhaps not, but the alternative, I think, does great violence to the statute, creates constitutional problems with smartphones, and again, is inconsistent with that text in history. Judge Okuda was also worried that if you adopted your approach, it made a mess of the exemptions or was hard to reconcile with some of the exemptions. Do you want to respond to that? Yeah, let me address that. There is a consent exception in the statute. If you get consent, you can use an automatic telephone dialing system. That's not superfluous under our approach because you can consent to be called by equipment that has the capacity to engage in random and sequential number generation, even if you yourself are called with that equipment as part of a list. So that takes care of the consent exception. Judge Okuda in Marx pointed to the government debt collection exception. There are a few explanations for that that are consistent with our theory. One, again, the same capacity point. Government debt collectors might use equipment to call people on a list, but that equipment might have the ability to engage in random and sequential number generation, and you can consent to that. In addition to that, Congress passed this government debt collection exception in November of 2015. This was a few months after the FCC had come out with its 2015 order, which the D.C. Circuit later set aside for being inherently self-contradictory, just mush. It's hardly surprising that Congress at that point would have done what any private defendant wishes it could have done, which is save itself. No matter how all this turns out, give us an exception so that we don't get caught up in this TCPA mess. And then I think the last point that Judge Okuda made was that through the government debt collection exception, Congress ratified the FCC's understanding of an ATDS. Ratified what? Again, this 2015 order was set aside by a unanimous D.C. Circuit panel for being essentially incomprehensible and self-contradictory. There was nothing to ratify when Congress passed that in 2015. So with due respect to the Ninth Circuit, marks shouldn't be followed. The textual analysis in marks essentially consists of recounting both sides' arguments, which I think we can agree there's no perfect reading of this statute, but the textual analysis was basically recounting both sides' arguments and then throwing the court's hands up in the air and saying, okay, both sides have points, so now let's go to context. First, I don't think you should give up on the text that easily. I think if you follow the rules of grammar, the better reading is to give effect to random or sequential number generation. But even the contextual points that Judge Okuda was making I think ultimately are unpersuasive and shouldn't be followed here. So the D.C. Circuit, you know, there's a couple different ways to read that opinion. On one hand, there's parts of the opinion where it seems Judge Srinivasan is saying either interpretation, in other words, the debate today, there's a point at which you can see him suggesting anyway either approach might be reasonable, and then the criticism punchline is you didn't even pick one, that's his real. But then when he talks about all of this covering Facebook, or excuse me, not Facebook, but cell phones, all cell phones, then he seems to suggest that can't possibly be a reasonable interpretation. So is anyone reading the D.C. Circuit is saying, well, no, there weren't two reasonable interpretations. There's only one, because if one of the one of the interpretations on the table covers all cell phones, that's eye popping coverage, blah, blah, blah. Which is it? What are they saying? I think, Judge, the D.C. Circuit stopped one step short of where it could have gone if it logically followed its own reasoning. The main issue in that D.C. Circuit appeal was capacity. That was what dominated the briefing and the argument. And so on the capacity question, the D.C. Circuit said, look, the one thing we know the statute can't mean is that every smartphone is an ATDS. When it then got to the functions question, which was properly before it, I agree with you, Judge Sutton, the D.C. Circuit could have applied the same logic and said the functions of an ATDS have to be used, list dialing as well, or to require random or sequential number generation. It just didn't go that far because it had a narrower ground on which to set aside the FCC's order, which was that it was inherently self-contradictory. I think as for whether there are two reasonable readings of this, look, I think if you just focus on this one provision, store or produce telephone numbers to be called using a random or sequential number generator, as we've discussed, it's a tough provision to bring any other considerations. Perhaps there could be two interpretations, but the better interpretation is to require random or sequential number generation. And I think when you look at the context and the constitutional concerns that we're talking about, that there is only one reasonable interpretation of it and that dialing from a list can't possibly be covered. In the very little bit of time that I have left, if the court were to not agree with what I just argued and instead to think that the previous orders of the FCC do still govern, we would still have a human intervention defense in this case, which is that this system could not dial without a human dialing agent clicking make call for each and every call. There was a one to one correlation between clicking make call and the connection being initiated and that connection was initiated virtually instantaneously. That's a voluntary act by human dialing agents which would satisfy the human intervention requirement under the FCC's pre-2015 orders. But the better way to resolve this case, I would submit, is to provide some clarity to this area of the law by holding that the TCPA does not cover list dialing, rather it requires random or sequential number generation. Thank you. Thank you. Your Honors, I'll start right there. He's wrong. It was not a one to one click to dial. It was more than four to one. In fact, Hilton got complaints that there were so many calls going out, there was no available agents and they changed the programming. This was not one to one and that's in the record. But the call doesn't occur without that clicker. That the call doesn't occur without the list being created. And that's what Blow and the Seventh Circuit and Marks held, that the human intervention is not in the creation of lists or even here passing through a list, it's on a dialing. Along that line, you asked, is it immediate? It's not. When they reprogrammed their system... Where in the record does it establish that? I can find the site, Your Honor, in a moment. But we did put this out there that when after it's clicked, the computer then determines whether there's enough available agents before it dials. So it does not go out automatically. Your agent says that, I mean, your expert says the intelligence system qualifies as an automatic dialer because after the agent clicks, may call the numbers forwarded to another part of the system to be dialed and the agent doesn't hear a dial tone. But there's nothing in that testimony that says the function is delayed or not instantaneous. Is there? There is, Your Honor. There's also their expert, Mr. Taylor, said that the agent doesn't dial the phone. There's no dialing capability and there's no phone system there. Does it go to whether it's instantaneous or not from the point at which the click agent presses the button? It has to go through the software, Your Honor, to see if there's an available agent. If there's no available agent, it does not go through. And, Your Honor, it's page 13 of our brief, our opening brief, footnote five, has the sites, there's DOC 104-1, there's a whole litany of them where it discusses that delay. Further, this panel asked Hilton, in 1991, what types of systems was Congress worried about? And he answered, they're all systems that only random and dial the numbers. But that's not true. We have over five pages in our brief citing forth that in 91, Congress heard testimony about targeted lists. And up to 40% of those calls were targeted lists. They were both. It wasn't one or the other. They were both. And they created a law that dealt with that. In fact, you know, the quote, one of the co-sponsors of the act, Representative Markey, the reason for the proliferation of such unsoliciting advertising over the nation's telecommunications network is that companies can now target their marketing. And it goes on, the reason why these unsolicited telemarketers can target individual homes is simple. Corporate America has your number. They're dying from lists. That was a concern in 91. And it's a concern now. So when I asked you about the iPhone, and I had a specific example, you said, well, to dial such numbers in my specific example. And I just want to make sure I'm getting the point because the word is capacity. And I thought the way this worked is once you decided you had the ATDS capacity, it didn't matter if it was one call at a time. Am I wrong about that? I think, Your Honor. Otherwise, I thought there was liability for one call as long as it was from an ATDS. There's liability for one call, but when the system is making mass calling, automatic calling. So if it's a series of calls, it's plural. It's not just the one call. But I thought that was I thought that was about the definition of it. Like, so it has to have the capacity to dial plural numbers. But once you establish that with the piece of technology, then it's used even for one call created liability. Aren't I right about that? I don't think so, Your Honor. And I think that's why. Tell me why. Tell me linguistically from the statute where the exemption is. We're assuming it's an ATDS, and they use it to call Jeff Sutton. Because it's not automatically dialing that number, Your Honor. And that's what Mark's held, that you're not turning every iPhone into an automatic dialer for dialing one number. It's only when you automatically dial without human intervention. And that's where the FCC orders come in. Looking at human intervention at the time of the call. And that's where. So the way to deal with this problem is assume the 03 and 08 order is just in place. Well, that's one way to do it, Your Honor. Just tell me from the statute. Just tell me the section of the statute. When I read it, I'm like, got you. The store, I think it's the plural. I think, I don't think it's just the capacity. I think it has to be the use. And I don't think, I think the systems, the way it's used is what matters. And the way it was used here was to automatically dial numerous numbers. And I think the modifier does not modify store. And I think that the panel asked the question, well, what's the example? Is there examples of when store can be used to randomly or sequentially dial numbers? And he said, yes. He can't give you a single example. And I believe Your Honor mentioned points that the Facebook petition might have some of those that I'm not familiar with. But I have yet to see a real-life example. Why isn't a customary production piece of technology going to store it? I mean, it's not. I thought the answer was you could have a situation where you produce and call. And you produce, store, and then call. And it seems like storing is almost automatic. Again, technology is advanced. That's the problem, is that interpretation, restore, superfluous to produce. You don't need it. You throw out store. If you say, you call numbers that are produced randomly or sequentially, you don't need store. Well, so, but isn't it possible you can have produce and call? Produce, store, and call? No. Produce and call, not store. Because they're always going to store. You don't need a function to produce. And there's not a single real-life example where that takes place. So the example might be produce, you store for call at 5 p.m. when they come home. That would be a list then. You're done dialing from a list. So I think that's the problem, is you're interpreting such a way to remove the context and the meaning of the statute. And you're doing it in a way that would allow telemarketers to end run it. I can have a list of every single number in this country. And if you say that I have to randomly or sequentially dial it, I can dial that list millions of times a day. I can sit in my basement with a couple computers, and I can blast every single person in this country and have no liability under the TCPA if you require a random or sequentially dial generator. And that is, as the Fourth Circuit said, dispiriting if you apply a statute that's supposed to protect consumers in a way that turn it on its head. Thank you. Okay. We understand your case. And now we can go for a reprise. Thank you. Evans.