Humor in Supreme Court oral arguments

1 Introduction

In United States v. Alvarez, the Supreme Court of the United States considered the constitutionality of a law criminalizing false claims of having received military honors. Justice Sonia Sotomayor noted that merely being offensive was not usually enough to trigger criminal prosecution by saying “I, too, take offense when people make these kinds of claims, but I take offense when someone I’m dating makes a claim that’s not true,” sparking laughter in the courtroom (Supreme Court 2012: 23). The arguing attorney’s riposte, “As the father of a 20-year-old daughter, so do I, Justice Sotomayor,” continued the lighthearted tone (Supreme Court 2012: 23).

Humor might not be expected in the reputedly staid and solemn context of a court of law. Prior research has, though, suggested that humor, as a strategy, frequently appears in political settings, especially when someone seeks to tear down an opposing argument, and the Supreme Court is among other things a very political environment, shaping law and policy nationwide. Yet it is a distinctive political environment, without the electoral incentives or video transmission that might be conjectured to produce the jousting style of modern political humor. As such, the Court can help shed light on how crucial those contextual factors are to political actors’ use of humor.

This article discusses the circumstances that make oral arguments likelier to break into laughter in the Supreme Court from 2010 to 2016. Synthesizing several previous studies, the analysis considers hypotheses about how humor in oral arguments is affected by timing (when in the day and year a deliberation occurs), contentiousness (how divided the justices are about the issue), and subject matter (whether cases concern an individual’s civil rights). It also examines whether laughter is more common during the side of the argument ultimately disfavored by the justices. With its broad-stroke picture, this approach can help contextualize previous studies examining the use of humor in individual arguments, providing a window into the cognitive and social considerations that shape communication strategies more generally. Humor turns out to be more common earlier in the day, in cases less directly implicating a specific individual’s rights, and during arguments that ultimately lose the case.

2 Humor as political communication in the Supreme Court of the United States

Humor underpins communication in a variety of realms (e.g., Beach and Prickett 2017; Meyer 2000; Watson and Drew 2017), helping to persuade by heightening the salience of substantive arguments and creating fellow-feeling between speaker and listener. Unsurprisingly, then, humor is a vital tool in politics as well (Knobloch-Westerwick and Lewis 2017; Skurka et al. 2018; Whaley and Holloway 1997; Young et al. 2018). Studies of humor in political contexts, however, most often focus on efforts to signal to or engage the mass public, even when considering signals sent by elite actors (e.g., Baumgartner and Morris 2008; Bippus 2007; Juzefovičs and Vihalemm 2020; Takovski 2020; Verhulsdonk et al. 2021; Whaley and Holloway 1997). Connecting with the public – potential voters and consumers of policies – is, of course, a key function of political communication. Nevertheless, interactions between elites as they attempt to inform or convince one another can be equally pivotal to political communication processes (Naurin 2010; Proksch and Slapin 2012). Understanding when and how elites use humor to communicate with one another can reveal both politically strategic behavior and the workings of deliberation (Tsakona 2009; Yarwood 2001).

Political elites, like the general public, use humor for several purposes. It can spin one’s own image in electoral campaigns (Peterson 2018), enhance personal likeability (Smith and Voth 2002), or portray opponents or policies as irrational (Meyer 2000). Elites use humor among themselves, too. It is a key socializing tool (Yarwood 2003), often invoked during deliberations to lessen the pressure of weighty policy decisions at hand (Morreall 1991). These uses appear across time and political institutions (Speier 1998). Notably, humor recurs in judicial institutions, an important but relatively less-studied context in political communications (Malphurs 2013; Manzano and Ura 2013; Sill et al. 2013). This role for humor extends to the highest constitutional courts, including the United States’ Supreme Court (Jacobi and Sag 2019; Li and Pryor 2020). While traditional campaigning does not play into the needs of either advocates or justices of the Supreme Court, humor has considerable use there: the Court can be of interest both for its generalizable and for its politically distinctive features.

For example, Dorin (1983) noted that Supreme Court justices seemed to use humor to smooth over tensions among themselves, paralleling behavior in other groups attempting to foster collegiality (Fine and Corte 2017). More recently, Wexler (2005) explored differences across individual justices in propensity to be humorous in oral argument and Hobbs (2007b, 2011, in considering legal communication strategies more broadly, has discussed how humor helps justices and Supreme Court advocates probe each other. These strategies of communication and persuasion have particular practical implications because the oral arguments can be decisive. Justices have long insisted that Supreme Court arguments can sway or even reverse opinions (Wasby et al. 1976), and recent empirical analyses confirm this (Ringmuth 2015; Ringsmuth et al. 2013). While not every justice is open to persuasion in every case – highly partisan, entrenched issues leave less ground for the changing of minds – understanding the use of humor in a potentially decisive context can help show how those with policy responsibility communicate with one another when deliberations matter.

The Supreme Court context, meanwhile, creates little need for concern about the electability implications or sound-bite suitability of remarks. Its proceedings are as yet unrecorded on video. Even the most visible of Court functions, the oral arguments, historically offered audio recordings only some days after the fact, making them staler and less likely to be disseminated by media in search of recent news (Instant audio has been made available for occasional newsworthy cases and consistently during and since the 2020 coronavirus-pandemic lockdown). This relative insulation from public attention heightens the focus on sending messages within the elite. Indeed, analysts have generally seen the Court’s oral arguments as primarily about the justices trying to convince one another or to extract information from attorneys, with only a peripheral role for speaking to the broader public (e.g., Black et al. 2012).

Previous studies looking at humor in Supreme Court arguments have tended to focus on close readings of a few specific cases (though with exceptions like Jacobi and Sag 2019). The analysis here seeks to complement and build upon this literature by looking at broader patterns of humorous usage. As previous studies of communication suggest that humor is often a targeted, conscious rhetorical strategy, the question concerns what contexts would be expected to make it more or less likely at the Supreme Court.

3 Factors encouraging humor during oral arguments

Comedy sometimes arises opportunistically or in response to irrelevant events. In the Supreme Court, a presumably exogenous failure of the Court’s lighting system during the arguments for Nichols v. United States led to the Chief Justice saying “I knew we should have paid that bill,” to laughter from those present (Supreme Court 2016: 50). Such random occurrences, though, coexist with underlying patterns that make the use of humor more or less likely as justices and attorneys speak to one another about pending cases. Most often, humor is a conscious choice by a justice or an attorney. Potential factors raised by the literature include the effects of timing-related fatigue, internal controversy or contention, and the substantive issue under discussion.

Fatigue is important because exhaustion from sustained expenditure on judicial effort can influence how deeply judges engage with individual cases (Danziger et al. 2011).^[1] That is, judges may be less willing or able to put in effort on cases they encounter after a long day or with substantial time since a break. This might especially affect something like the effort to be witty, which is inessential to the central tasks of judging yet requires cognitive effort (Shammi and Stuss 1999). As marshalling such effort becomes more difficult over the course of a day, concentrating on the core elements of law rather than on humorous framing may be more compelling. Later arguments are especially likely to require more effort among a population with the seniority and achievement to become Supreme Court justices or advocates: older adults generally experience maximum alertness earlier in the day (Dijk et al. 2000). This suggests that earlier arguments in a given day will feature more humor than those later on.

Humor need not only be the result of laborious effort, though. It could also be a sign of attention depletion – that is, of the inability to maintain the effort of focusing on the task at hand, letting the mind wander to more amusing thoughts. In this case, the predictions of the effect of being later in the day reverse, so that cases later in the day would be expected to include more, not less, likely to feature humor.

Fatigue and tension operate at longer time scales as well. The Supreme Court uses a calendar with several overlapping cycles of activity. Weeks when the Court hears arguments have in recent decades typically involved cases on Monday, Tuesday, and Wednesday, with a private conference among the justices on Friday. Those weeks are themselves unevenly distributed, with two consecutive weeks in session followed by multiple weeks without arguments. There is also an annual rhythm as the justices attempt to wrap up all work for the term, notably including issuing all the year’s opinions, by the end of June, before dispersing for the summer. This annual cycle is particularly notable since the most contentious cases, requiring the most internal back-and-forth among justices and hence the longest time to decide, comprise a disproportionate share of those pending later in the term rather than earlier.

If longer breaks from the particularly intense work of hearing arguments help recharge justices and make it easier (or harder) for them to engage with wittiness from each other or from attorneys, then differences in the amount of humor may arise between Monday and Wednesday arguments or between the first week of a sitting and the following one. Similarly, as disagreements accumulate over the course of a sitting or within and across annual terms, the propensity for humor may change concomitantly.

This effect of disagreement, moreover, may also more narrowly apply to case being argued. Arguments where the justices anticipate close votes seem to generate distinctive communication styles during argument (Beňuš et al. 2014). Advocates and justices may resort to humor when they are pressing a case that seems closely contested, as a means of capturing the attention of pivotal voters. Alternatively, clear-cut cases where the justices sense they do not need to convince their colleagues may open up a more relaxed, congenial atmosphere conducive to humor among the justices (and the attorney arguing the case getting a warmer reception). It may after all be safer then to score rhetorical points, as the low risk of changing outcomes by flippantly alienating a vote may make humor more tempting.

Previous studies further suggest that this risk of topic-inappropriate flippancy may influence elites’, including high-court actors’, use of comedic styles. Humor can sometimes be a conscious strategy for defusing communication about serious or painful issues (Beach and Prickett 2017). Still, it is a risky approach in the setting of powerful people shaping policy, as it can appear glib, callous, and trivializing to the task of justice (Rushing 1990; Smith 1990). Commentators consequently often worry about humor in the courtroom being unbefitting to the solemnity of the judicial role (Hobbs 2007a; Oakley and Opeskin 2016).

Some cases amplify this risk of drollery. Cases in which specific individuals’ life or liberty are at stake likely impose tighter bounds on what might be considered in good taste (Rudolph 1989). The use of humor in such instances may come off as particularly cruel, risking the erosion of respect for institutions ranging from the court directly involved to the broader rule of law (Tsakona 2009). At least some Supreme Court personnel do seem to make distinctions along these lines, between cases where a light tone is more or less appropriate. For example, when Justices Scalia and Breyer lightheartedly sniped at one another over past opinions during the oral argument of Trevino v. Thaler, Justice Kennedy facetiously remarked, “This is very amusing in a capital case” (Supreme Court 2013: 43). This attitude suggests that cases involving issues of criminal-justice procedure or civil rights may be less hospitable for humor and accordingly less likely to see laughs than might cases involving, say, the scope of federalism or the award of attorney’s fees.^[2]

4 Humor as an indicator of attitudes towards arguments

Who is laughed at may matter as much as whether laughter occurs (see also Li and Pryor 2020). Supreme Court arguments have two adversarial sides, the petitioner hoping to overturn the lower-court ruling and the respondent in opposition.^[3] Communications scholarship suggests that the two sides in a given argument may not be equally likely to spur laughter. Notably, political debates, including those in courts, often use humor to belittle or dismiss opponents’ points: Signaling that a claim is to be laughed at implies that it lacks merit or will reflect poorly on those who adopt it (Hobbs 2007b, 2011; Tsakona 2009). Arguments that the hearer finds weak – or came to the conversation predisposed to reject – are accordingly likelier to get a humorous rejoinder, whether biting or dismissive. Not only does such derision undercut the argument under discussion, it may color attitudes towards the advocate’s other arguments. If one side of an argument triggers more laugh outbursts than its opponent, then, the higher-laugh side may struggle to win the case, whether because it signals ineffectual advocacy or because the justices were hostile to the merits of the case being made.

While this derisive use of humor is the one previously most attested in the political realm, in other contexts laughter has the opposite implication. That is, because humor seeks to connect with an audience, it helps form social capital (Navarro et al. 2016). Although such connections can be used to dismiss out-groups in favor of in-groups (paralleling the ways it dismisses opposing arguments), it can also build bridges within the in-group. Advocates and justices may thus feel more willing and able to joke when they are fundamentally sympathetic to each other’s arguments. Similarly, successfully impressing the audience with cleverness and wit, or causing the justices to relax into less formality, may make advocates and their arguments more compelling. This process would lead to a greater correlation between one side of the argument’s having more humor and that side being more, rather than less, likely to be successful.

5 Methods

The set of Supreme Court oral arguments from the start of the 2010 term, when Elena Kagan assumed her position as justice, until the death of Antonin Scalia in the middle of the 2015 term, offer a helpfully consistent context in which to examine the use of humor in arguments. Over these roughly five and a half terms, the Court heard arguments in just over 400 cases. While the attorneys and courtroom audiences differ from one case to the next, the justices themselves remain constant barring the occasional recusal (Omitting cases with recusals does not greatly change reported results). This consistency does, while increasing internal validity, suggest a need for caution when generalizing to other contexts, since idiosyncrasies of the people involved may influence outcomes. However, the hypotheses to be tested do not derive from any particular characteristics of the nine justices present in this data set, so there is little obvious reason to suppose that the results will not apply to other groups of justices, or of political elites.

The first set of hypotheses to be tested concern case characteristics – timing, decisiveness of result, and subject matter – that may increase or decrease the use of humorous styles in oral arguments. The measure here is drawn from the number of times the Supreme Court’s official transcript records that proceedings were interrupted by laughter.^[4] Such a measure necessarily omits failed attempts at humor that do not generate laughter raucous enough to merit noting in the transcripts (Bell 2015). This measure makes for a relatively narrow, rare use of comedy: over half the arguments (208 out of 402) produced two or fewer laughs, with almost 20 percent (72 out of 402) having none at all. On the other hand, thirteen cases saw ten or more laughs, with Yates v. United States (about whether a fish could be a “tangible object” for record-keeping purposes) tallying fifteen laughs and Florida v. Department of Health and Human Services (the last of a week of arguments about whether the Affordable Care Act involved unconstitutional coercion of states) eighteen. The coding accordingly allows for variation in the amount of humor observed. Negative-binomial regression models are used to predict this count-based dependent variable, as statistical tests reveal sufficient overdispersion to make Poisson models inappropriate.

Several of the predictor variables measure argument timing. One codes whether the argument was the first, second, or third of the day. Another codes the day of the two-week sitting in which the argument falls (in a typical sitting, this runs from 1 to 6, with 1 indicating Monday of the sitting’s first week and 6 indicating Wednesday of the setting’s second week). A third codes which sitting of the term the argument falls in, from the first, October, sitting, through the seventh, April, sitting. A final timing variable measures the year-to-year time trend.

As noted above, Supreme Court argument days vary in the number of arguments heard. Most argument days feature two arguments. Sometimes, though, a case being belatedly pulled from the Court’s calendar, or a holiday, affects the schedule. Under these circumstances, one- or three-argument days can arise. This pattern produces far fewer more first (217 instances) or second (173) case of the day than third cases (12; excluding these third cases does not greatly affect reported results). It also creates the possibility that generally busy days might shape the use of humor. Justices could be relatively focused and staid even on the first and second arguments on days featuring three arguments, simply because they have so many points and laws to mentally juggle. Third arguments of a day by definition occur on days packed with more arguments, so this sort of focus could create the false appearance that later arguments rather than busy days changed behavior. To account for this alternative possibility, a control variable counts the number of cases to be heard on the day of a given argument.

In testing the hypotheses regarding the effect of a close rather than blowout case on the adoption of a humorous communication style, analyses here focus on the size of the majority’s margin on the final vote. This is the number of justices who vote with or concur in the Court’s judgment, at least in part, minus the number of justices who do not concur with the Court’s judgment at all. By focusing on the “judgment,” this measure ignores the specific legal basis of the opinion issued: someone who concurs with the Court’s decision while disagreeing on the statutory or constitutional grounds for that conclusion counts as a member of the majority for this purpose.

The measure of cases’ subject matter derives from the Supreme Court Database (Spaeth et al. 2017), which codes the primary issue implicated in each case argued before the Court. In particular, to test whether cases that risk depriving individuals of personal liberty or other civil rights see less laughter, a variable indicates whether the database’s “issue Area” variable classified the case as dealing with “Criminal Procedure” or “Civil Rights.” Cases in either of these realms are coded with a value of 1, and those in other realms are coded 0.^[5]

All models predicting the total number of outbreaks of laughter include controls for the argument’s length. Having more time naturally allows for more time for witticisms. Although 97% of the arguments included here were officially allotted 60 min of argument, if advocates wrap up their presentations quickly, the argument may be shorter (as short as 36 min among the observations here). Conversely, the Chief Justice has discretion to allow arguments to run long, with ostensibly 60 min arguments lasting as long as 95 min. Controlling for this is particularly important when considering time-of-day effects, since Chief Justice Roberts is often more relaxed about time limits on days with only one argument, which is by definition the first of the day. Argument lengths are calculated from the official transcript, which provides a starting and ending time (in minutes) for each argument. These models also include nested random effects (i.e., random slopes) for term, sitting, and date, to account for any correlations within these temporal levels, with cases nested in dates, which are nested in sittings, which are nested in terms.

Testing whether laughs being concentrated during the argument of one particular side (petitioner or respondent) associates with rulings more likely to favor that side requires measuring which side a ruling favors. The most substantively important outcome is whether the petitioner won the case. Data used here follows the Supreme Court Database’s “partyWinning” variable in assuming that any decision except a dismissal of the case or an affirmance of the lower court’s ruling represents a win for the petitioner. Since this is a binary variable, it is predicted using logistic regressions. For a slightly less blunt measure that may better capture marginal effects of humorous argument styles, an alternative operationalization of this idea looks at the proportion of voting justices who supported an outcome favorable to petitioners, so that if a petitioner received the support of three out of nine justices the variable takes a value of 0.33… and the support of six out of eight justices (for example, in case of a recusal) would be 0.75. Fractional logistic regressions are used to predict this outcome. Whether predicting a win or a higher vote share for the petitioner, using probabilistic rather than logistic models produces substantively very similar estimates.

Hypotheses about differences in humor during the arguments also require assigning laughs to either side. This is mostly straightforward, since the argument transcripts are clearly divided into sections according to which attorney is arguing and which side – petitioner or respondent – the attorney is arguing in support of. Some arguments do, however, have friend-of-the-court arguments from third parties, most often the federal government, who are in support of neither side. The measure ignores (i.e., does not credit to either side) outbursts of laughter during these neutral arguments or during the presiding justice’s introductory spiel, simply subtracting the count of reported laughs during the respondent’s argument from those during the petitioner’s. To account for the effects of longer or shorter arguments, the difference in the number of laughs is adjusted to be the figure that, at the pace of laughs observed during the actual argument, would have occurred in a standard 60 min argument. The difference in observed laughs in a 120 min argument would hence be divided by two. The difference in a 48 min argument would similarly be multiplied by 1.25.

As a control variable, models include the Supreme Court Database’s measure of whether the lower-court ruling being appealed – and hence the respondent’s position – is liberal (coded 0) or conservative (coded 1). With the Court’s conservative tilt on most issues during the period examined, those petitioning against a conservative lower-court decision may face different communication styles from the justices and be less likely to win the case. This measure is unavailable for a small number of cases, mostly lawsuits between states for which the Supreme Court serves as the court of first instance rather than a forum for appealing lower-court rulings.

Models also include a measure of what proportion of the advocates arguing for a given side are “expert” Supreme Court litigators, using benchmarks established in previous studies of the Court (e.g., Johnson et al. 2006; Lazarus 2008). These measures include whether the advocate in question has argued at least five times before the Supreme Court, served as a clerk for a Supreme Court justice, or worked in the Office of the Solicitor General. Advocates who meet any of these criteria are deemed expert, and the measure considers the difference between the expert fraction of advocates supporting the petitioner’s position and the expert fraction of advocates support the respondent’s position. For example, if all counsel arguing for the petitioner are experts but only one of two arguing for the respondent is, the variable takes on the value of ½ (= 1 − ½), but if none of the petitioner’s arguing counsel are experts but all of the respondent’s counsel are, the value would be −1 (= 0 − 1).^[6] Expertise may both improve the likelihood of the advocate’s side winning and affect the likelihood of humor (if increased mutual comfort between attorneys and justices leads to more joshing, or conversely if inexperienced litigators are more likely to face sarcasm).

Connecting arguments to decisions raises an issue not encountered in simply considering features internal to arguments: the decision-makers who ultimately disposed of the case were in some cases not the same as those at oral argument. In particular, twenty-seven cases were argued while Justice Scalia was still alive but not decided before his death, so that he contributed to the atmosphere during oral arguments but did not vote in the case’s ultimate disposition. This is particularly important because Justice Scalia was typically one of the most loquacious justices in oral arguments (Johnson and Black 2017), the likeliest to be humorous (Wexler 2005), and particularly partial to the typically humorous mode of sarcasm (Hasen 2015). His ability to interject in some arguments without corresponding say in whether or not the petitioner won the case might therefore skew findings. The regression models predicting voting results are accordingly estimated both with and without these cases to ensure they do not drive any results.

6 Results

Table 1 shows the predictions of the regression models estimating total number of laughs, with standard errors of each estimate in parentheses. As the table’s first row indicates with its consistently negative coefficients, arguments later in the day involve less humor. This effect is substantively large. Holding control variables at their mean values, the coefficients imply that the typical first argument of the day has just over three humorous moments but that the day’s third argument has just over two. This is a decrease of more than 30%, roughly comparable in size to the predicted effect of the argument being 9 min shorter. The effect is also statistically significant at the p < 0.05 level, with z statistics ranging from −2.31 (in the first and third columns, p = 0.021) to −2.40 (in the second column, p = 0.018). Notably, humor continues to decrease in a day’s later arguments when controlling for the number of arguments per day (in the table’s second column), so those involved are not simply being less jocular on days when there are many arguments to get through. The coefficient on having more arguments in a day is in fact positive, implying more laughs in the average argument on a day with more arguments, though not statistically significantly so (z = 0.66, p = 0.510). Instead, it appears that arguments following multiple others on the same day have less humor.

Fatigue is not observed to have comparably large or statistically significant effects across larger time scales, as the second through fourth rows of coefficients indicate. Arguments later in sittings and in the annual term see if anything more laughs (i.e., their coefficients are positive), though neither these effects nor the effect of term are statistically distinguishable from zero: the smallest p-value among these variables, that on the “Term” variable in the table’s rightmost column, is 0.105.

Ultimate disagreement among the justices also has little relationship to the amount of humor during the oral argument, as seen in the “Margin of majority” row in Table 1’s third column of results. The predicted effect of a large margin is both trivially small and far short of any standard statistical-significance threshold (z = 0.05, p = 0.957). In fact, in alternative specifications the sign of the coefficient sometimes changes, though the effect remains tiny and indistinguishable from zero.

The case’s subject matter has a substantial effect, though. Cases concerning individuals’ rights, and thus potentially creating an identifiable victim of laughter, have less humor, as the negative sign on the relevant variable indicates (z = −2.74, p = 0.006). With control variables at their means, individual-rights cases see 2.4 humorous moments (95% confidence interval: 2.0–2.8), compared to other cases’ average of 3.2 (95% confidence interval: 2.8–3.5). This is roughly the same increase that the model predicts would occur if the argument were fourteen minutes longer.

Turning to the effects of one side of an argument generating more humor than the other does, Table 2 shows regression coefficients predicting petitioners’ success. In all specifications, more laughs occurring during the petitioner’s argument than the respondent’s associates with lower levels of success for the petitioner: the first row’s coefficients are all negative, with p statistics ranging from the second column’s 0.008 (z = −2.65) to the third column’s 0.045 (z = −2.00). That is, humor associates with less ultimate support among the justices, matching other studies’ view of judicial humor as most often sardonic and disapproving (Hobbs 2011). Neither control variable in Table 2 attains standard statistical significance levels, although the lower-court’s decision being conservative is marginally statistically significant in the second column (z = −1.72, p = 0.085). However, the control-variable coefficients are in the expected directions. The negative coefficients on the “direction of lower-court ruling” variable indicate that appeals against conservative rulings are less likely to win, and Supreme Court specialist litigators tend (as the positive coefficients indicate) to be somewhat likelier to win their cases.

The petitioner tends to win a majority of the vote no matter how many laughs arise during argument, echoing the established finding that the justices are more likely to review cases where they disagree with the lower court’s ruling (Bryan and Owens 2017). However, the probability of the favorable result has a clear association with the relative balance of humor in the argument. Petitioners who enjoy three fewer laughs than the respondent have a predicted probability of a favorable ruling around 70% and a predicted vote share of around two thirds (six out of the usual nine justices). Those enduring three more laughs than the respondent, however, eke out a predicted win probability just slightly over, and not statistically distinguishable from, 50%, with an expected vote share of approximately five-ninths. These estimates are generally similar whether or not Justice Scalia was alive to vote in the Court’s final decision.

The causal mechanism connecting jocular oral arguments to less-successful trial outcomes is somewhat ambiguous. The justices’ preexisting skepticism about an argument may lead them to ridicule it. Alternatively, any feature of an advocate’s style that provokes mockery from the judges may reduce that advocate’s convincingness in presenting arguments, leading to less favorable results. Publicly available data from oral arguments cannot distinguish between these two mechanisms, but in either case the conclusion holds that whichever side of the argument sparks more humorous episodes tends to have less success in the ultimate voting.

7 Conclusion

Humor appears in Supreme Court oral arguments in predictable patterns. It seems to be more common in earlier arguments in a given day, when the justices are likely to be fresher and more alert, and less frequently employed in cases involving an individual’s fundamental rights. Within arguments, laughter is likelier during presentations to which the justices ultimately prove unsympathetic. Like the number of questions directed at each side during oral arguments (Shullman 2004), humor can thus serve as a leading indicator of Supreme Court votes.

These results fit with emerging patterns in political- and judicial-communication studies. Although, when facing the public, politicians may use folksy humor to emphasize commonality, in deliberating among themselves about specific issues, political actors’ humor often has a sharper edge: Its function as a challenging or belittling technique corresponds to its use during arguments that the majority of the Court rules against, as observed in previous case studies of specific oral arguments (Hobbs 2007a, 2011). Court arguments do have some public-facing elements, with the courtroom spectators as well as the press reports and transcripts that can reach a broader audience, so these results do not absolutely reflect intra-elite communications. Still, they offer indications about behavior when decisionmakers are mostly talking to one another.

As communication style signals underlying patterns of cognition and norms, these findings are suggestive in many ways. First, deliberations and hearings earlier in the day may differ in quality and so, potentially, in outcome from those later on. Second, the subject under deliberation matters as well. Even among a set of elites – justices and attorneys – who are not subject to electoral retribution for remarks that seem disrespectful or rude, the tone of communication modulates in response to subject matter, becoming less comical when examining matters more likely to be personal to some member of the public. This solemnity may flow from the seemingly dismissive nature of oral-argument humor: a colloquy whose outcome controls an identifiable individual’s life or liberty may heighten the norm that ridicule is inappropriate.

While some features of the underlying theory here apply to other political contexts, future research might explore how much these results actually do extend to other contexts, such as legislative hearings. For example, to the extent livelier morning proceedings derive from participants’ advanced age, do other heavily gerontocratic institutions such as the Senate or the Soviet Politburo resemble the Supreme Court than do more youthful settings? Do the unique features of electoral politics, and the differences in personality that lead the ambitious to become judges rather than other sorts of officeholders, change the issues or circumstances in which humor is used? Does being in a setting that, unlike Supreme Court arguments, is typically broadcast on television or over the Internet matter? Conversely, within the Supreme Court, it may be worth comparing differences in communication style among individual justices with their distinct personalities and public profiles (Fogarty et al. 2020; Masood and Kassow 2020). Future research might also explore the extent to which humor serves as an effective rhetorical technique for persuasive communication rather than simply being a signal of preexisting attitudes. If the use of more humor during one side’s arguments correlates not merely with more votes against that side’s case but also more likelihood of justices’ changing their opinion from their pre-argument positions (Ringsmuth et al. 2013), it may help establish that comically made points land home in oral arguments.

Oral arguments are, alongside written opinions issued alongside judgments, one of the main glimpses that the Supreme Court offers of its own workings to analysts and to the general public. The arguments can moreover potentially exert a real influence on the outcome of cases and, by extension, on important statutory and constitutional issues. The communication techniques seen in those arguments, then, are of interest in helping to understand broader political outcomes. The choice to use humor there is a telling detail in a larger picture.