Abstract
Famous trials not only generate immense popularity and intrigue, they also have the power to change history. Surprisingly, little research examines the use of complex language during these culturally-significant trial outcomes. In the present study, we helped fill in this gap by evaluating the relationship between attorneys’ use of integratively complex language and trial outcomes. Using the well-validated Automated Integrative Complexity scoring system, we analyzed the complexity of language in the opening and closing statements of famous trials. We found that higher levels of integrative complexity led to a significant increase in winning outcomes, but only for the prosecution. Further, this effect was driven by elaborative forms of complexity and not dialectical forms of complexity. Taken together, these results fill a large gap in our understanding of how language might influence the outcomes of culturally-significant legal proceedings.
Sixty years ago, millions of people tuned in to watch the first televised courtroom trial of war criminal Adolf Eichmann. Since then, other famous trials have captivated American interest. About 5.2 million people watched the acquittal of Casey Anthony (Joyella, 2011), 2.1 million people tuned in for the live “Guilty” verdict of Dr. Conrad Murphy for the death of Michael Jackson (Schutte, 2011), and an astonishing 100 million viewers watched as the “Not Guilty” verdict was delivered to O.J. Simpson (Wood, 2013).
Many famous trials generate immense interest, but more importantly, famous trials can change the course of history, such as Brown v. Topeka Board of Education or United States v. Susan B. Anthony. Given their cultural significance, it is worth evaluating what factors might have been associated with particular sides winning and losing those famous trials. In the present study, we use a well-validated linguistic variable—integrative complexity—to evaluate the relationship between language and trial outcome in famous trials.
What do we know so far about the factors influencing culturally-significant trial outcomes? Surprisingly, no research that we know of has systematically examined factors influencing the outcomes of these famous trials. However, there is considerable research that attempts to understand the outcome of trials more broadly. This includes exploring pre-trial attitudes (Griffin, 2016; Lundrigan et al., 2013; Peter-Hagene & Bottoms, 2017), testimony style (Conley et al., 1978; Freedman et al., 1996; Hildebrand-Edgar & Ehrlich, 2017; O’Barr, 1982), and the influence of social media (Bakhshay & Haney, 2018; Battaglia, 2012; Taylor & Tarrant, 2019). Furthermore, many factors have been examined related to what influences a juror member’s verdict: factors related to jury decision making (Rodriguez et al., 2019; Weinstock & Flaton, 2004) and the race of the defendant (Leippe et al., 2016; Skolnick & Shaw, 2010; Sommers, 2007). Thus, there are many important elements within “common” (i.e., not-famous) trials that influence the outcome.
Importantly for our present purpose, there is additionally quite a bit of research that examines the effect of language at various points of trials. Indeed, a new methodological approach, corpus linguistics, provides sets of tools for researchers to analyze legal issues (Solan & Gales, 2017; Vogel et al., 2018). Although no work that we know of systematically explores the language of famous trials specifically, some work on linguistics is relevant to this topic. For example, a body of work examines language used in trials from the Supreme Court of the United States (SCOTUS). Although some of these trials are of course famous, not all Supreme Court trials end up being culturally-significant, widely-discussed, or widely-watched events. Nonetheless, they are instructive for our present purpose. Previous research examines how SCOTUS and Supreme Court justice (e.g., opinion pieces) language is predictive of a variety of the Court’s processes and outcomes (Black et al., 2011; Brenner & Heberlig, 2002; Carlson et al., 2015; Coleman & Phung, 2010; Feldman, 2016, 2017; Hawes et al., 2009; Hazelton et al., 2019; Johnson, 2014; Long & Christensen, 2013; Mazzi, 2010; McGuire et al., 2007; Savchak & Bowie, 2016; Skinner & Pludwin, 2013). Moreover, the use of complex and simple language by Supreme Court justices can provide insight into SCOTUS decisions and behavior (Gruenfeld, 1995; Gruenfeld & Preston, 2000; Hansford & Coe, 2019; Owens & Wedeking, 2011; Tetlock et al., 1985). For instance, higher levels of complexity were exhibited in Supreme Court justice writings regarding decisions to uphold legal precedent (Gruenfeld & Preston, 2000). Collectively, this research indicates that language—and in particular the complexity of language—can be used to understand SCOTUS trials.
As suggested by this review so far, there is a burgeoning literature on linguistic analysis of trials, but little work has evaluated the type of language used by some of the most important players in any trial: The attorneys for the prosecution and defense. Previous research related to the prosecution/defense has explored attorney ability (Abrams & Yoon, 2007; Anderson & Heaton, 2011), emotional expression in the opening and closing statements (Hahn & Clayton, 1996), and discussing the race of the defendant (Bucolo & Cohn, 2010). However, there is a relative lack of research on the linguistic approach of attorneys on trial outcome more broadly—and no research that we are aware of examining the type of language used by attorneys in culturally-significant famous trials specifically.
In the present paper, we help fill in these gaps in two ways. (a) First, we evaluate the relationship between the language used by trial attorneys and the outcome of famous, culturally-significant trials. No research has yet examined the language of attorneys in these trials. (b) Second, we evaluate a particular linguistic quality of the prosecution and defense that is known to be associated with outcomes in other areas (e.g., Abe, 2011, 2012), as well as to predict important aspects of SCOTUS trials (Gruenfeld, 1995; Gruenfeld & Preston, 2000; Hansford & Coe, 2019; Owens & Wedeking, 2011; Tetlock et al., 1985): The simplicity or complexity of language used during the trial. Specifically, we use a well-validated measurement of Automated Integrative Complexity (Conway et al., 2014, 2020; Houck et al., 2014) to evaluate the relationship between opening/closing statements’ complexity and success in multiple famous trials. In so doing, we can understand the relationship between complex (or simple) language and successful trial outcomes in culturally-significant famous trials. Thus, the present study adds to an already rich literature on linguistic analysis of trials by evaluating (for the first time) the linguistic complexity of attorneys in relation to the outcome of culturally-significant famous trials.
Integrative Complexity
What is complex language? Although many different ways to evaluate the complexity of language exist (e.g., Abe, 2011, 2012; Hermann, 1980; Jordan et al., 2019), we focus here on a widely-used and widely-validated measurement of linguistic complexity: Integrative complexity. Developed in its current instantiation by Peter Suedfeld (e.g., Suedfeld et al., 1977), integrative complexity is used to assess the complexity of spoken or written communications according to their basic structure. Passages are coded and assigned a score between 1 and 7 based on the level of differentiation (i.e., the extent to which differing dimensions are used to describe a given topic, scores 1–3) and, if more than one dimension is present, integration (i.e., the joining of these multiple dimensions to form new ideas about or approaches to the topic, scores 4 to 7; see Baker-Brown et al., 1992, for integrative complexity scoring details).
In assigning complexity scores, the particular position argued for by the speaker/writer is irrelevant; the score is based on the structure of the passage rather than its substance or meaning. As such, this construct is able to capture the underlying mechanisms of the complexity of thought on a broad level regardless of which variables may have influenced the cognitive strategies used in formulating the passages.
Elaborative Versus Dialectical Complexity
Due to its breadth, two linguistic passages that describe the same topic in very different ways can receive the same integrative complexity score. Although this feature is one of the strengths of the construct, it also limits theory-building because knowing why a particular passage is complex is often difficult. In response to this problem, Conway et al. (2008, 2011) developed two subordinate constructs under the rubric of the Multiple Complexity Model. The model, designed as a supplement to and not as a replacement for the integrative complexity construct, considers the different routes by which complex thinking arises. In particular, it considers whether complexity assigned through integrative complexity coding arose in an elaborative complexity (argument of a singular perspective complexly) form, a dialectical complexity (recognizes the tension of multiple competing perspectives complexly) form, or a simultaneous elaborative and dialectical complexity form.
Elaborative complexity is achieved when a topic is described using two or more differentiated points supporting either a positive or a negative dominant theme, but not both. Consider the statement “whale watching boats are not fun. I feel fearfully anxious when I am in line for them, and once I get off, I am nauseous for days.” Both of these differentiated elements are used to support the negative argument that whale watching boats are not fun. In contrast, dialectical complexity is achieved when both negative and positive aspects of the same topic are differentiated, giving validity to each side. Consider the statement “I both dislike and like whale watching boats. I get very scared when I’m in line, but I love seeing the whales up close.” Both a negative element and a positive element are used to describe the topic. A passage may contain both elaborative and dialectical complexity (“I both dislike and like whale watching boats. I get scared while I’m in line, and always am nauseous once I get off the ride; but during the ride I enjoy the seeing the whales up close”).
The Multiple Complexity Model is incorporated into the integrative complexity construct, and the passages are coded and scored on the same 1 to 7 scale. Therefore, all passages coded under the Multiple Complexity Model receive three scores: One for overall integrative complexity, and one each for the two subcomponents (elaborative, dialectical; see, Conway et al., 2008; Houck et al., 2014).
Will Simple or Complex Rhetoric Win a Trial?
There is ample theoretical and empirical evidence that suggests the strategic manipulation of linguistic complexity can be beneficial for success (e.g., Conway et al., 2012; Repke et al., 2018; Suedfeld & Rank, 1976; Tetlock, 1981; Thoemmes & Conway, 2007). However, although integrative complexity research examines expansive topics ranging from political speeches (e.g., Conway et al., 2012; Conway & Zubrod, 2020; Houck & Conway, 2019) to terrorism (e.g., Houck et al., 2017; Putra et al., 2018), no integrative complexity research that we know of has examined attorneys’ linguistic style during trials (although, as discussed earlier, some work has examined Supreme Court judges, Gruenfeld, 1995; Gruenfeld & Preston, 2000; Hansford & Coe, 2019; Owens & Wedeking, 2011; Tetlock et al., 1985). As a result of this lack of research on attorney courtroom language complexity, we had no certain predictions entering the study. However, prior research in other areas offers some clues as to possibilities. We discuss those below.
Reasons Why Simple Language Might Be More Effective
Simplicity and complexity have often-competing strengths and weaknesses (e.g., Conway et al., 2012). On the one hand, simple rhetoric is often easier to understand and perceived as more powerful. Indeed, there is some work that suggests individuals prefer simplicity over complexity. Within the world of pop culture, there is evidence to suggest that award-winning (e.g., Oscars, Golden Globes) films show lower integrative complexity than their losing counterparts (McCullough & Conway, 2018a). In political domains, some evidence similarly suggests that decreases in complexity during the election season lead to more electoral success (Conway et al., 2012, Study 1; Thoemmes & Conway, 2007). The authors (Conway et al., 2012) suggested this may partially be because people view too much complexity as a sign of weakness and indecision. Further, research has tied the effectiveness of terrorist propaganda to its simplicity (Houck et al., 2017). Even in legal domains, plain English (opposed to legalese; Flammer, 2010) and clear opinions (Benson & Kessler, 1987; Owens & Wedeking, 2011) are preferred among courts and provide less opportunity for ambiguity. For instance, an experimental study demonstrated that complex decision language can hinder decision acceptance and detract from basic cues regarding decision judgment (Hansford & Coe, 2019).
Reasons Why Complex Language Might Be More Effective
On the other hand, complex language can also be viewed as a marker of high ability (for discussion, see Conway et al., 2012), and complex arguments contain more flexibility in their potential to persuade wider audiences (see Repke et al., 2018). Indeed, some research suggests that complexity can sometimes be more powerful. For example, in the popular world of fanfiction, there is a preference of complexity relative to simplicity for popular and successful fanfiction (McCullough, 2018a; see also McCullough, 2018b). Further, one study (Conway et al., 2012, Study 2) that gave complex and simple rhetoric for political candidates to participants found that participants showed more of a positive shift in voting tendencies after reading a complex (vs.) simple passage from the candidate—but only for the candidate who was likely viewed as simple-minded to begin with (McCain). Although no work directly ties this work to trial outcomes specifically, some work suggests it may apply: Majority opinion Supreme Court justices wrote more complexly regarding decisions to uphold legal precedent (Gruenfeld & Preston, 2000). This work does not directly show a clear line between outcome and complexity, but it does suggest that justices on the winning side write more complexly when upholding legal precedent. Collectively, along with additional work on attitudes we cover in the discussion (e.g., Carpenter, 2015), these results suggest higher levels of complexity, in some message contexts, might produce more positive outcomes when higher quality is expected and warranted—such as in fanfiction (or, possibly, in a courtroom).
A Multiple Complexity Perspective
In this context, then, it might be especially important for attorneys to present their arguments using language that heightens the likelihood that both the attorneys and their arguments are perceived as intelligent. However, both types of complexity in the Multiple Complexity Model discussed above may not be equally useful for this purpose. Indeed, it may be that higher levels of elaborative complexity—complexity used to defend one side of an argument—would be more effective for attorneys to win. On the flip side, it may be inadvisable for attorneys to present a dialectically complex approach that might imply reasons to doubt the truth of one’s case.
If these reflections are true, we ought to see that winning sides are associated with higher elaborative complexity—but not dialectical complexity. Some evidence from the persuasion literature is consistent with this notion: For example, one meta-analysis suggested that refutational two-sided arguments—that is, arguments that both refute the opponents’ side and build up one’s own side—are more persuasive than one-sided appeals alone (O’Keefe, 1999). This suggests, consistent with the elaborative complexity = more effectiveness perspective, that providing independent dimensions pointing in the same direction (both refutations of opponent arguments and new arguments that build one’s case are equally in favor of a single position) is more effective than providing a simpler one-sided argument. It is worth noting, however, that this effect is complicated and potentially qualified by multiple factors (for additional work, see, e.g., Allen, 1991; Crowley et al., 1994; Eisend, 2007; Lawson, 1970; Niederdeppe et al., 2012). Work in complexity has similarly suggested that integrative complexity effects are often situationally-specific (Conway et al., 2016). Given this, and the fact that there is currently limited prior evidence from courtrooms (and no evidence from attorney language) to draw on, we make no specific predictions. Indeed, this lack of prior research highlights the need for the current study: We know very little about the relationship between attorney language and culturally-significant courtroom outcomes.
Methods
Overview of Design
In this study, we examined the integrative complexity of opening and closing statements from famous, culturally-significant trials. Due to controversy about power estimates (e.g., Levine, 2011, 2013; Levine & Hullett, 2002; Levine et al., 2008; Morrison, 2010), we follow previous SCOTUS trial and language research norms relevant to power and report the descriptive information relevant to that issue (e.g., Black et al., 2011; Mazzi, 2010). Notably, because our study detected some significant effects at p ≤ .001, we were clearly powered enough to detect the key reported effects below. The current study is comprised of 233,475 words and 1,891 paragraphs, drawn from 28 opening and closing statements. 1 Using AutoIC (Conway et al., 2014, 2020), integrative complexity, dialectical complexity, and elaborative complexity were assessed. We looked at AutoIC levels by the outcome of the trial (win/loss for the speaker). We also assessed two other key factors: The type of statement (opening vs. closing statement) and who said the statement (prosecution vs. defense).
Automated Integrative Complexity
Designed by integrative complexity (IC) researchers, the Automated Integrative Complexity scoring system (AutoIC) is an automated scoring system specifically designed to measure the IC construct (Conway et al., 2014; Houck et al., 2014). Similar to the human-scored IC system, the AutoIC scores differentiation (distinguishing different dimensions) and integration (the recognition of the interplay of differentiated dimensions). In addition, AutoIC scores both dialectical complexity (the recognition of the tension between multiple competing perspectives) and elaborative complexity (illustrates a singular perspective; see Conway et al., 2008). In sum, three scores were assessed from the same 1 to 7 scale produced by AutoIC: (1) an overall IC score, (2) elaborative complexity, and (3) dialectical complexity.
Since its inception, AutoIC has shown higher correlations with expert human scorers than other attempts to automate linguistic complexity (Conway et al., 2014, 2020). In the original validity paper, there was an average correlation of r = 0.46 between AutoIC and human-scored IC paragraphs (Conway et al., 2014). Further, AutoIC has also replicated effects from human-scored IC in various materials (e.g., presidential debates, early Christian writings, and smoking/health domains; Conway et al., 2014).
More recent research also provides a great deal of further evidence that AutoIC is a valid measure of IC (see Conway et al., 2020, for a summary). For instance, more recent studies comparing expert human scorers to AutoIC have shown higher correlations than the original validity paper (Conway et al., 2020; Houck et al., 2018; McCullough & Conway, 2018a; Prinsloo, 2016). Further, there is ample evidence across various domains to suggest that AutoIC produces theoretically interpretable findings: terrorism (Houck et al., 2017; Putra et al., 2018), fictional versus real dialogue (McCullough & Conway, 2018b), decision-making (Prinsloo, 2016), the film industry (McCullough & Conway, 2018a), religion (Houck et al., 2018), and social media ( McCullough & Conway, 2019). In addition, recent tests (Conway et al., 2020) reveal successful AutoIC replication attempts of prior hand-scored IC studies in health, ideology, and presidential leadership. They further show that AutoIC passes basic validity tests such as comparing Donald Trump’s rhetoric to that of famous philosophers (Conway et al., 2020).
AutoIC avoids the often labor intensive and time constrained reality of human scoring IC and thus allows for scoring of all the available materials from each respective trial (human-scoring codes only a small subset), increasing the power and ecological validity of the study. In the present study, for each opening and closing statement, we scored materials separately for each trial (removing all dialogue from the Court, other attorneys, and/or the judge). AutoIC automatically divides each opening/closing statement into paragraphs of equal word length and provides a score for each paragraph as well as provides an overall document score. Although we focus our narrative on document-level analyzes, we also performed paragraph-level analyzes and report a summary in a footnote. As will be seen there, the results are not dependent on the level of analyzes used.
Selection of Opening and Closing Statements
We collected materials from the Famous Trials database that contains materials (e.g., primary documents, images, essays) related to various famous trials. Created by Professor Douglas O. Linder, the first installation of the Famous Trials database appeared on the web in 1995 and was most recently updated in 2017. Linder had two criteria that guided his choice to include a particular trial for the database: (1) the trial must have garnered public attention (at least in its own given time and place) and (2) the trial must have significantly shaped history or the trial must provide evidence for observing and understanding a given time in history. The database currently contains over 80 trials with varying degrees of available information for each respective trial. The Famous Trials database is a resource for various official platforms, such as Pace University’s Pace Law School Library. Thus, for the purposes of the current study, we deemed this an acceptable source from which to retrieve opening and closing transcript statements.
Due to the fact that the information available for each trial is variable in the database, in the present study, opening and closing statements were used when there was at least one full statement available per trial in each category (i.e., either a complete opening and/or closing statement from the prosecution or defense). Specifically, to be used in the current study, the opening and closing statements had to meet the following criteria: (1) we had to have access to either (or both) the opening and closing statement for a given trial, (2) we had to know if the statement was from the prosecution or the defense, and (3) the outcome of the trial had to be known. Furthermore, the main analyzes consisted of all the available documents; however, not all the documents were parallel (i.e., we did not always have opening and closing statements for each side in every trial). Thus, we also ran analyzes looking at paired opening and closing statements only.
Results
We ran a series of 2 (outcome: win or loss) × 2 (speaker: prosecution or defense) × 2 (type of statement: opening or closing) Factorial ANOVAs on integrative complexity, elaborative complexity, and dialectical complexity. A summary of key analyzes is presented in Table 1.
Means and Standard Deviations of the Trial Outcome and Prosecution/Defense.
Note. p-value = interaction between trial outcome and prosecution/defense. Standard deviations are in parentheses.
Integrative Complexity
For integrative complexity, results revealed a main effect for the outcome of the trial: Trial winners (M = 1.89) used higher integrative complexity than trial losers (M = 1.74), F(1, 20) = 6.01, p = .024, d = 1.10. There was no main effect for statement type, p = .993 or prosecution/defense, p = .141. Further, there were no significant interactions between the variables, although an interaction between outcome and prosecution/defense emerged that was in the same direction as reported below for elaborative complexity, p < .20, d = 0.61 (see Table 1; Figure 1).
Integrative complexity: interaction between trial outcome and prosecution/defense.
Elaborative and Dialectical Complexity
Additional results suggest the above effect is being driven by elaborative (and not dialectical) complexity. Specifically, for elaborative complexity, results revealed a main effect for the outcome of the trial: Trial winners (M = 1.44) used higher elaborative complexity than trial losers (M = 1.27), F(1, 20) = 13.90, p = .001, d = 1.67. This main effect was qualified by an interaction between prosecution/defense and the outcome of the trial. There was no difference on elaborative complexity related to the outcome of the trial for the defense (win M = 1.36, loss M = 1.35); however, there was a difference for the prosecution in the outcome of the trial (win M = 1.52, loss M = 1.20), interaction F(1, 20) = 9.66, p = .006, d = 0.57 (Figure 2). Further, there were no main effects of type of statement (p = .787) and prosecution/defense (p = .488). There was a three-way interaction between statement type, who said the statement, and the outcome of the trial that approached significance (p = .055). This three-way interaction suggested that the interaction between outcome and prosecution/defense was stronger for opening statements than closing statements.
Elaborative complexity: interaction between trial outcome and prosecution/defense.
For dialectical complexity, results suggested that there were no main effects for statement type (p = .593), prosecution/defense (p = .101), and the outcome of the trial (p = .277). In addition, there were no interactions among the variables. 2
Paired Statements
We ran analyzes on the paired opening and closing statements from the same corresponding trial to see if there were any paired differences. These analyzes remove any cases where there was no corresponding opening or closing statement from the same attorney. We used the same ANOVA design to examine the differences among the factors in relation to integrative, elaborative, and dialectical complexity from the paired statements. These results revealed that, although effects for integrative complexity were weaker, the key effects for elaborative complexity remained significant (and effect sizes remained comparable).
Specifically, for elaborative complexity, there was a main effect for the outcome of the trial: Trial winners (M = 1.43) used higher elaborative complexity than trial losers (M = 1.29), F(1, 11) = 4.94, p = .048, d = 1.34. Further, as in the analyzes with all the documents, this main effect was qualified by an interaction between prosecution/defense and the outcome of the trial. There was no difference on elaborative complexity related to the outcome of the trial for the defense (win M = 1.32, loss M = 1.35); however, there was a difference for the prosecution in the outcome of the trial (win M = 1.56, loss M = 1.22), interaction F(1, 11) = 7.50, p = .019, d = 1.65 (Figure 3). There was a three-way interaction between statement type, who said the statement, and the outcome of the trial, interaction F(1, 11) = 5.17, p = .044, d = 1.37. Mirroring analyzes on the whole sample, the interaction between outcome and prosecution/defense was stronger for opening statements than closing statements. Further, there were no main effects of type of statement (p = .854) and prosecution/defense (p = .894), or any other significant interactions among the variables. As in the main analyzes, for dialectical complexity, there were no main effects for statement type (p = .406), prosecution/defense (p = .456), and the outcome of the trial (p = .717), or interactions.3,4
Paired statements: interaction between trial outcome and prosecution/defense.
Discussion
Taken together, these results reveal that higher integrative complexity is associated with winning in famous trials, especially for the prosecution (and not the defense). In both the whole sample and paired sample analyzes, integrative complexity was significantly higher for prosecuting trial winners than prosecuting trial losers. Equally as important, subsequent analyzes further suggest that this effect is driven by elaborative and not dialectical forms of complexity. Our results consistently demonstrate that higher elaborative complexity leads to more famous trial wins for the prosecution. Thus, prosecutors (but not the defenders) who elaborate their argument complexly win more often than they lose in famous trials. Below, we discuss the relationship of elaboratively complex language with trial outcomes. We then discuss why this effect might occur for prosecuting (but not defense) attorneys. We conclude by discussing what the present study adds to literature on the social psychology of language.
The Relationship of Elaborative Complexity to Trial Outcomes
Why might elaboratively complex language be associated with more positive trial outcomes? We explore here three possibilities.
Elaborative complexity might serve as a heuristic cue
First, it may be that elaborative complexity serves as a cue to the jury that the speaker is intelligent and capable, the arguments are good arguments, and/or taps into the message length heuristic. Consider two statements. In statement 1, an attorney says, “this person is guilty; the physical evidence supports that conclusion, and one example is the glove at the scene.” In statement 2, an attorney says, “this person is guilty; not only does the physical evidence support that conclusion, but also, independently, in contradiction of what my opponent says, the eyewitness testimony supports this conclusion.” According to AutoIC, Statement 1 scores a 1 on elaborative complexity, while statement 2 scores a 3. And it is possible that part of the reason that Statement 2—and similar statements—are more successful in trial contexts is that they function much like heuristic cues (e.g., Petty et al., 1981) suggesting to the jury that there are multiple dimensions supporting the argument—and thus both the attorney and the argument are judged a higher quality. Indeed, work suggests that arguments in favor of a position—whether the arguments are strong or weak arguments—consistently increases persuasion (Petty & Cacioppo, 1984). This is potentially conceptually parallel to elaborative complexity, which measures the consistency with which independent dimensions are used in rhetoric. Thus, it is possible that increasing elaborative complexity functions in famous trials much like the message length heuristic.
Elaborative complexity might increase the perceived quality of the argument
Of course, one hopes that juries go beyond heuristic processing to more effortful processing as well. To the degree that is the case, it is possible that elaboratively complex arguments work through a different mechanism: Namely, that such arguments are higher quality arguments that, as a result, are more likely on average to persuade an audience in this context that is likely to produce effortful consideration.
In most research paradigms, strong, high-quality arguments are classified as those that invoke the most positive elaboration in favor of an attitude and the least negative elaboration opposed to it (see, e.g., Carpenter, 2015; Hoeken et al., 2019; O’Keefe, 2013). As other researchers have long noted, establishing criteria for what makes a quality argument is a complicated task that needs to incorporate normative criteria for argument strength that are independent of invoked elaboration (e.g., Fishbein & Ajzen, 1981; Hahn, 2020; Hoeken et al., 2019; O’Keefe, 2013). Further, the relationship between the contents of an argument and the likelihood of processing deeply are complicated (e.g., Liu et al., 2016). That said, for our purposes, it is worth noting the clear conceptual overlap between traditional conceptualizations of a strong/quality argument and elaborative complexity. However defined, it is empirically the case that most research associates strong/quality arguments with greater invoked positive elaboration on one side of an argument (Carpenter, 2015). Because elaboration in these studies often is operationalized as the number of generated points on one side of the argument, it suggests that prior work on strong arguments is functionally similar to elaborative (but not dialectical) complexity. Elaborative complexity, like the often-used criterion for strong/quality arguments, involves multiple points discussed on one (but not the other) side of a discussion or debate. 5
Although no work directly ties our measurement of elaborative complexity to argument quality, this clear conceptual overlap does provide a foundation from which to understand our effects. In trials—and especially in famous trials—the stakes might be so high that the targets of the persuasive message, the jury, might be especially likely to process elaboratively with effortful thought. If so, factors that influence argument quality might play a more important role than those based in heuristic processing. And indeed, much prior work on argument quality and attitudes suggests that is the case. Argument quality is often more influential in persuasion if people are able and motivated to assess the message (Petty & Cacioppo, 1986), and a meta-analysis revealed that messages are more persuasive with higher argument quality, if an argument is processed through the central route of persuasion (as opposed to the peripheral/heuristic route; Carpenter, 2015). 6 To the degree that elaboratively complex arguments are especially persuasive to the person who is motivated to process effortfully, this suggests that the mechanism by which elaborative complexity is associated with successful outcomes might run through effortful processing. In courtroom contexts especially—and particularly for famous trials where so much is at stake—a jury’s attention to such details might outweigh many of the potential disadvantages of complex rhetoric discussed in the introduction.
Indeed, one of the clear theoretical suggestions of this work is that the value of complex language in producing successful outcomes likely goes up as the stakes go up. Much like work on attitudes (e.g., Petty & Cacioppo, 1984; Petty et al., 1981) suggests people rely more on argument quality and less on source heuristics as issues become more important, it may be that in immersive contexts such as a high-profile courtroom trial, elaborative (but not dialectical) complexity may have a higher value for potential success. Thus, with higher stakes, individuals might be more likely to rely on message quality that is expressed through rhetoric. This also suggests that argument quality and elaboratively complex rhetoric may be less important in trials involving smaller stakes—and even less important in more anonymous personal opinion votes (e.g., McCullough & Conway, 2018a).
Strength of initial case might produce more elaborative complexity
Finally, and relatedly, it is possible that attorneys who actually had stronger cases—and thus were more likely to win—were likely to use elaborative complexity to begin with. Elaborative complexity requires multiple dimensions in favor of one’s case in order to have something to elaborate on. Probabilistically speaking, if one attorney has four good independent arguments and another attorney has two, the attorney with four may be more likely to use elaborative complexity (to discuss more of the independent points at any time) than the attorney with two. Of course, even if true, this would not clearly distinguish the causal mechanism behind the effect. For example, if this were true, it could be the case that the attorneys with the stronger case were going to win anyway (elaborative complexity or not). But it also could be the case that this probabilistic advantage helped attorneys with a stronger case win due to its effect on increasing elaborative complexity, and that if the opposing attorneys had used their fewer points more elaboratively in a strategically complex way, those attorneys would have increased their odds of success. 7
Prosecuting Versus Defending Attorneys
Why might the effect of elaborative complexity hold only for prosecuting attorneys? At this point, we can only speculate. We consider three overlapping possibilities. First, it is possible the context of the typical trial favors elaborative complexity for prosecutors more than for defense. In particular, defense attorneys have an intrinsic “presumption of innocence” which shifts the burden of proof to the prosecutors. As a result, a successful outcome may require a compensatory elaborative argumentative orientation by the prosecution to counteract this presumption.
Relatedly, it may be that elaborative complexity only works for prosecuting attorneys because juries perceive that if defense attorneys use too much complexity, they are appearing too defensive and thus the elaborative complexity backfires. Thus, nonrefutational two-sided arguments (arguments that acknowledge but do not attack the other side; see Niederdeppe et al., 2012) might be utilized more by the defense than the prosecution to avoid the label of being too “defensive,” which would equate to higher dialectical complexity. Furthermore, high elaborative complexity is associated with more extreme attitudes (Conway et al., 2008); and it is possible that such a position simply works better for the prosecution than the defense. Attackers generally need to use more complexity for success than defenders (see, e.g., Suedfeld, 1992).
A final possibility is that utilizing high levels of elaborative complexity could be effective for winning for both the prosecution and the defense, but the prosecution is more likely to utilize this strategy. A prosecution attorney more naturally might lean towards high forms of elaborative complexity attacking the defense with one key argument that is extensively expanded on in a multi-dimensional way: guilt. An attack strategy marked by a singular perspective that is elaboratively defended has also been found in terrorist rhetoric before a violent act (Conway et al., 2011). This interpretation in the present data is partially supported by the mean pattern for defensive complexity: Overall, defense attorneys had fairly low levels of defensive complexity (the ratio of elaborative to dialectical complexity) whether they won or lost—levels comparable to losing prosecutors. Clearly some defense attorneys did use higher levels of defensive complexity, but there may simply not have been enough power at these levels to detect an effect. So it is possible that, if more defense attorneys had more consistently used a defensive complexity pattern in line with prosecuting attorneys (or we had a larger sample to include more such cases), we would see a similar pattern for defense attorneys as well. Thus, this finding may be due to the fact that defense attorneys are less likely to actually use defensive complexity approaches in their rhetoric—either because they on average have fewer dimensions to draw from, or that they (erroneously) perceive it will be less effective, or both. 8
Limitations
Like all studies, this study has limitations. First, we used a selective group of famous trial opening and closing statements that were available and selected by contributors on the Famous Trials database. Further, selected trials are only from one cultural context and given in one language. In general, the exploratory nature of the current study and monolingual materials resulted in a concentrated sample, which does limit the scope of our findings. Thus, similar to much prior work, our study uses a somewhat condensed Western, English-speaking sample. However, this does not invalidate the basic conclusion of the study on this sample. First, we were purposefully interested in famous and culturally-significant trials. Not only are these trials of historical importance, they also provide one vital context for building theories about complex language and outcomes. Second, all studies have selective samples in varying degrees, and this should not stop research on the social psychology of language. Our work at a minimum suggests that, on this sample of historically important trials, high levels of integrative complexity—and especially driven by elaborative complexity—is associated with more winning in trials for prosecutors.
Concluding Thoughts
These results offer insight into understanding the impact of language on the outcome of famous trials. More broadly, they also represent an advance for our understanding of the social psychology of language. Indeed, the present study highlights a new application for the well-documented linguistic construct of integrative complexity to the criminal justice domain. There is ample evidence from other domains suggesting that strategic manipulation of linguistic complexity can be beneficial for success, in work that includes U.S. presidents (Conway et al., 2012; Tetlock, 1981; Thoemmes & Conway, 2007), revolutionary leaders (Suedfeld & Rank, 1976), and simulated lying scenarios (Repke et al., 2018). The present results build on that prior work by suggesting that such strategic alteration by prosecuting attorneys could affect the outcome of famous trials. They further offer up an excellent starting point for additional work that (a) expands the scope of trials beyond famous trials, (b) expands the scope of materials beyond opening and closing statements, and (c) uses more controlled experimental settings to directly parse out some of the explanations offered here. Perhaps, the use of linguistic mastery in the courtroom, in the form of complexity, could be another tool used to win a case.
Footnotes
Acknowledgements
The authors thank the two reviewers for their constructive and well-articulated set of critiques.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
