Hand gestures in Roman rhetoric and public speaking

Public speakers use gestures in a number of different ways, including making deictic references, as a way of representing objects or actions, and as a way of punctuating, marking, or showing aspects of the structure of their speech.Footnote ⁹ These types of gestures are sometimes called illustrators.Footnote ¹⁰ Gestures can also express emotions and feelings and often play a very important role in key moments of social interaction, sometimes even replacing words.Footnote ¹¹ In the latter case, they become kinesic codes or emblems, in the terminology of Ekman and Friesen;Footnote ¹² the ultimate example – where signs replace all words – is sign languages.Footnote ¹³

The significance of gestures in communication among the Romans is indicated not only by rhetorical treatisesFootnote ¹⁴ but also by anecdotes found scattered through the work of various ancient authors regarding this aspect of body language.Footnote ¹⁵ According to Lucretius,Footnote ¹⁶ there was even a pre-speech stage of human history when communication was based on gestures and grunts alone. While there is no evidence of codified sign languages in Roman times, the concept of gestural communication was not foreign to ancient thought – Plato in the Cratylus has Socrates ask: “Suppose that we had no voice or tongue, and wanted to communicate with one another, should we not, like the deaf and dumb, make signs with the hands and head and the rest of the body?”Footnote ¹⁷ Though this passage is often cited as the oldest reference to sign language, it likely describes ad hoc non-verbal communication rather than a formal system like modern sign languages, which emerged only about 300 years ago.Footnote ¹⁸ In the Roman context, gestures had a wide range of communicative functions, including expressing emotion, indicating, mimicking, signaling, and displaying rhythm.Footnote ¹⁹

The use and significance of gestures in Roman rhetoric and art have already been the subjects of numerous studies.Footnote ²⁰ We rely on Quintilian because his work represents the most comprehensive treatise on ancient rhetoric, in which the matter of hand gestures is also treated most extensively. Since he was active in the second half of the 1st c. CE, it is necessary to address scholarly debates regarding the evolution of gestural communication between the times of Cicero and Augustus, and those of Quintilian. Some scholars believe that the gestural repertoire developed considerably and became more formalized over this period,Footnote ²¹ or that the gestures may have had different meanings than in everyday life, or even that the Quintilian system had to be taught because it was artificial and unknown to average Romans.Footnote ²² Others, however, argue that gestures did not very much change and that Quintilian’s work was firmly rooted in tradition.Footnote ²³ Jon Hall draws attention to the nature of the texts – Quintilian’s “textbook-ness” – while observing that Cicero advocated more spontaneous and intuitive gestures, something the textbook author could not afford to do.Footnote ²⁴ While the precise interpretation of specific gestures may be contestable, our study focuses primarily on the visual presence of gestural communication, independent of semantic nuance. Regardless of the specific historical period of Cicero or Quintilian, we can confidently assert the existence of diverse manual gesticulations involving finger and hand(s) manipulations.

When speaking, Roman orators communicated with their listeners in two languages: verbal and non-verbal. The latter category included mainly vocal non-verbal factors (vox), gestures (gestus), and facial expressions (vultus).Footnote ²⁵ The aim was to help stimulate the emotions of the listeners, which in turn was supposed to influence their judgement.Footnote ²⁶ Romans believed that a speech delivered brilliantly, even if mediocre in content, held more power than a poorly presented great speech.Footnote ²⁷ Quintilian emphasizes the importance of gestures in his analysis of Cicero’s Pro Ligario, highlighting that a skilled speaker would incorporate up to eight distinct gestures when delivering a sentence consisting of merely 17 words.Footnote ²⁸

Gestures held great significance, particularly at public gatherings such as contiones, quaestiones, and other occasions where speakers addressed the people or soldiers,Footnote ²⁹ as well as during public religious rites.Footnote ³⁰ Public oratory played a crucial role in Roman politics, at various types of contiones and other meetings, including legislative, electoral, judicial, military, and funerary gatherings.Footnote ³¹ Legislative contiones provided an opportunity for influential politicians to discuss laws that had been proposed and present their views to the Roman people. According to some scholars,Footnote ³² contiones played a pivotal role in determining the fate of such legislation. They argue that proponents would withdraw a law if they sensed a lack of support from the participants, effectively avoiding a substantial blow to their reputation at the People’s Assembly.

In Republican times, contiones played a vital role in the elections of officials, allowing candidates to present themselves and seek public support from former consuls and other influential figures. Funerary or laudatory speeches also played an important role in Roman public life, constituting an essential component of funerary rites. They involved family members praising the deceased or officials speaking on behalf of the community. The oration delivered by Caesar at the funeral of his aunt, and the wife of the late C. Marius,Footnote ³³ is considered an important step in his political career.Footnote ³⁴

Following Caesar’s assassination, the Roman Republic was in turmoil, with public sentiment oscillating between grief and outrage. In this charged atmosphere, Mark Antony and Marcus Brutus, among others, delivered a series of historic speeches. Antony, a staunch ally of Caesar, swayed the masses by portraying Caesar as a benevolent ruler unjustly murdered, revealing Caesar’s will and his blood-stained cloak to incite emotions. In contrast, Brutus argued that the assassination was necessary to save the Republic from tyranny, appealing to the citizens’ sense of justice and republican values. These speeches showcased the political acumen and persuasive prowess of both men, setting the stage for the power struggle that would end the Roman Republic and usher in the Roman Empire.Footnote ³⁵

Although political oratory may have diminished in significance during the Imperial period, especially in the context of public debate, given the declining role of the people in governance, the practice persisted. It remained an essential part of certain official acts, such as instructing voters before elections (for as long as elections continued) and providing platforms where the public could be informed about matters of state. Despite their reduced political weight, contiones and other speech-giving events still functioned as arenas for public communication between the government and the people.Footnote ³⁶ Crucially, they were increasingly controlled by the emperor, who used them to communicate and reinforce his authority. We have an especially large number of examples from the Julio-Claudian dynasty, including from the time of Augustus.Footnote ³⁷ Many of these speeches took the form of funeral eulogies, which played a significant role in reinforcing imperial authority and public memory.Footnote ³⁸ During political crises, the role of contiones in propaganda and public persuasion often intensified again; this is exemplified in the rivalry between Galba and Otho in 69 CE, where oratory in contiones played a key part in securing military and public support.Footnote ³⁹ Evidence from the correspondence between Marcus Aurelius and Fronto suggests that imperial public rhetoric retained a crucial administrative function, with the emperor expected to address not only the senatorial body but also the broader populace on diverse matters during public assemblies.Footnote ⁴⁰ Although the sources are limited, it appears that consuls, praetors, and later urban prefects also presided over public-speaking events, particularly those with ceremonial purposes, such as funerals or oath-taking. By the Late Empire, their use was further restricted, likely becoming the exclusive domain of the emperor and his representatives. Ultimately, while contiones no longer shaped political decision-making as they had in the Republic, they persisted as instruments of imperial authority, public communication, and legitimacy.Footnote ⁴¹

Crowd size and composition

The substantial political and ceremonial significance of contiones and other speech-giving events justifies investigating the identity of the individuals who assembled to listen to Roman politicians and how many of them there were. The question of crowd composition has been addressed many times before.Footnote ⁴² The composition of the crowd itself, however, tells us little about its size. The written sources provide little assistance in this matter, as they rarely, if ever, go beyond describing the size of the crowd, as either small or large at most.Footnote ⁴³ Consequently, attempts to estimate crowd sizes are based solely on the dimensions of the venues. For the Forum Romanum, these estimates – made in relation to elections rather than contiones – put the number of people present at 6,000, 10,000, or 15,000–20,000.Footnote ⁴⁴ However, given the area of the Forum Romanum and a reasonable measure of crowd density, this estimate could be raised to at least 30,000.Footnote ⁴⁵

Both the size and the composition of the crowd probably varied considerably depending on the issue being discussed, the commitment of those with a stake in the debate, the time for preparation, and even the level of the politician’s popularity across different social circles.Footnote ⁴⁶ This does not change the fact that the audience (and its number) was an integral element of a speech, as Cicero made clear in De oratore: “A speaker can no more be eloquent without a large audience than a flute player can perform without a flute.”Footnote ⁴⁷

Thus, the maximum number of participants in these public meetings can only be approximated. As the meetings served as platforms for conveying information, expressing opinions, and persuading others during elections and votes, certain limitations regarding the visibility of the speaker and their gestures, as well as the audibility of their speech, could have affected the maximum capacity of a specific space to accommodate people. In this paper, we present the results of analyses of the visibility of rhetorical gestures for different speaking platforms located in the Late Republican and Early Imperial Forum Romanum. While other venues were used as speaking platforms in Rome – written sources indicate that speeches were also delivered from the podium of the Capitoline Temple, the podium of the Temple of Bellona, and within the Circus FlaminiusFootnote ⁴⁸ – we focus on the Forum Romanum because the other venues and their immediate surroundings are much less well understood archaeologically, making reconstructions difficult.

The main objective is to estimate, for the different case studies, the maximum number of people who could see the rhetorical gestures as a proxy for maximum crowd size. The reason for taking this approach is that written sources depict contiones in which certain members of the audience were able to see the speaker’s gestures but unable to discern the words being spoken.Footnote ⁴⁹ This suggests that, in some cases, the number of those present did not fall within the range within which the speech was intelligibly audible. The approach yields measurable and reliable results that are less influenced by uncertain factors than acoustic simulations. In particular, it offers an easier way to grasp the overall crowd layout than methods focused on speech intelligibility in open areas. Although based on Roman rhetorical gestures, the methodology, in our view, is universal and could also be applied to the analysis of various public events in other cultures. Apart from the fact that the results can be viewed in the context of well-known historical events (such as Cicero’s delivery of the pro lege Manilia from the rostra vetera in 66 BCE, the contio on Metellus Nepos’s proposal to recall Pompey from the East convened at the rostra aedis Castoris in 62 BCE,Footnote ⁵⁰ the speeches delivered by Mark Antony and M. Aemilius Lepidus on March 16, 44 BCEFootnote ⁵¹ from the rostra Caesaris, or the delivery of the Sixth Philippic by Cicero from the same place on January 4, 43 BCE),Footnote ⁵² they allow us to compare the different speaking platforms and to examine the oratorical taskscape of the Forum and changes to it more objectively. The choice of chronological range was prompted by the fact that, during this relatively short period, the Forum underwent the most dynamic changes to its spatial geometry, including the relocation of speaking platforms, while simultaneously, public assemblies experienced functional shifts driven by political changes and the rise of the Principate.Footnote ⁵³

We shall also use the results to verify some of the hypotheses that might explain the changes that occurred in the oratorical taskscape of the Forum Romanum between the Late Republican period and the reign of Augustus.Footnote ⁵⁴ Based on the results obtained, we shall present our own hypothesis regarding the desire to separate participants in contiones taking place simultaneously in different venues at the Forum.

Object of research

The subject of our study is contiones that took place in the Forum Romanum during three different – though chronologically close – periods of the forum’s development:

• – Late Republican (LR): ca. 54 BCE (Fig. 1a)

  

  Fig. 1. Forum Romanum, top views: (a) ca. 54 BCE: rostra vetera marked red, Temple of Castor and Pollux marked blue; (b) ca. 27 BCE: rostra Augusti marked red, Temple of Castor and Pollux marked blue, Temple of Divus Iulius marked yellow; (c) ca. 14 CE: rostra Augusti marked red, Temple of Castor and Pollux marked blue, Temple of Divus Iulius marked yellow.

• – Early Augustan (EA): ca. 27 BCE (Fig. 1b)

• – Late Augustan (LA): ca. 14 CE (Fig. 1c)

For the case of the Forum Romanum ca. 54 BCE, we analyzed two speaking platforms (see Fig. 1a):

• – rostra vetera

• – rostra aedis Castoris (podium of the Temple of Castor and Pollux)

For the case of the Forum Romanum ca. 27 BCE and ca. 14 CE, on the other hand, we analyzed three speaking platforms (see Fig. 1b–1c):

• – rostra Augusti

• – rostra aedis Castoris

• – rostra aedis divi Iulii (podium of the Temple of Divus Iulius)Footnote ⁵⁵

The virtual 3D reconstructions prepared for research purposes replicate the spatial geometry, the appearance of buildings, and the interrelations between those buildings as known from the products of the Digitales Forum Romanum project,Footnote ⁵⁶ with minor modifications (see Supplementary Files S1–S3).Footnote ⁵⁷

The rostra vetera (Republican rostra), which was adjacent to the Comitium, was the primary platform for public speaking in Rome. However, Roman politicians also used alternative venues when necessary.Footnote ⁵⁸ We know that between 161 and 100 BCE, there was a transfer of the comitia from the Comitium to the area of the Temple of Castor and Pollux.Footnote ⁵⁹ The speaking platform likely emerged there due to the remodeling of the Temple of Castor and Pollux around the middle of the 2nd c. BCE, referred to as Temple IA.Footnote ⁶⁰ This platform was further developed during a major reconstruction initiated by L. Caecilius Metellus Dalmaticus (Temple II), which began around 117 BCE.Footnote ⁶¹ Historical records indicate that Scipio Aemilianus delivered a speech from the temple’s podium as early as 142 BCE.Footnote ⁶² We know of several contiones, during which crowds gathered in front of the temple, that took place in the late 60s and early 50s of the 1st c. BCE.Footnote ⁶³ A series of contiones was certainly held there in November 44 BCE: first, Mark Antony spoke against Octavian,Footnote ⁶⁴ then a contio was convened by the tribune of the plebs, Ti. Cannutius, where both the tribune and Octavian spoke against Antony.Footnote ⁶⁵

The oratorical taskscape of the Forum Romanum underwent major changes during and just after the civil wars. The initial phase of this transformation can be attributed to Caesar’s interventions, which, as DaviesFootnote ⁶⁶ aptly observes, “overwhelmed constraints on architectural propaganda” in parts of the Forum. This architectural reconfiguration represented a calculated effort to reshape the physical and symbolic taskscape of Rome’s central public space, thereby consolidating Caesar’s political authority through the built environment. The Forum underwent a significant functional and symbolic transformation, transitioning from its traditional role as predominantly a political arena – a space of dignitas – to an environment increasingly oriented towards the aggrandizement and promotion of charismatic individuals.Footnote ⁶⁷ According to Cassius Dio’s account,Footnote ⁶⁸ the decision to build the new rostra in a different location was Caesar’s, and it was under his dictatorship that its construction began. Coarelli believes that the date of 44 BCE given by Dio refers to the completion of the construction of the speaking platform and its dedication by Antony and that the project had begun perhaps in 45 BCE.Footnote ⁶⁹ Muth hypothesizes that Caesar’s decision stemmed from the fact that the Basilica Aemilia not only limited the number of people who could gather in front of the old rostra, but even restricted the number of people who could hear the speeches delivered from there intelligibly.Footnote ⁷⁰ Written sources mention several occasions on which speeches were delivered from this speaking platform.Footnote ⁷¹

A new rostra (rostra aedis divi Iuli) appeared in the Forum Romanum at the dedication of the Temple to Divus Iulius on August 18, 29 BCE.Footnote ⁷² The construction of the temple had begun as early as 42 BCE;Footnote ⁷³ therefore, the available space in front of the Temple of Castor and Pollux was reduced. Most scholars believe that the rostra was an integral part of the temple podium,Footnote ⁷⁴ as was the case with the Temple of Castor and Pollux. Coarelli argues that it was a free-standing platform in front of the temple,Footnote ⁷⁵ but his arguments are not convincing (the traces he interprets as remnants of this rostra are more likely of Late Antique origin).Footnote ⁷⁶ Although it is likely that many speeches of different natures were delivered from this place, we have secure evidence only for two funeral eulogies.Footnote ⁷⁷

The fate of the rostra aedis Castoris is debated. One view holds that with the dedication of the temple (or perhaps even with its commencement),Footnote ⁷⁸ the platform went out of use, as the Temple of Divus Iulius somewhat limited the space in front of it. While we know that the speaking platform was eventually removed altogether, there is disagreement about the timing. Taylor argues that the platform remained in place during Tiberius’s restoration after the fires of 14 and 9 BCE,Footnote ⁷⁹ interpreting this as evidence that the podium maintained its role as a rostra – a view also supported by Ulrich.Footnote ⁸⁰ However, the platform must have been removed by the beginning of the 3rd c. CE at the latest, since the surviving fragment of the Forma Urbis Romae (18bc) shows the temple without the speaking platform.Footnote ⁸¹

Thus, we have a complete picture of the physical setting for the oratorical taskscape of the Forum Romanum in the Late Republican and Early Imperial periods and can compare the visibility of rhetorical gestures for individual rostra while considering the changes that occurred over time. As the entire Forum Romanum underwent a rather far-reaching transformation during the period in question, we will also trace hypothetical scenarios for the rostra vetera and the rostra Augusti in which we examine the visibility of gestures from the rostra Augusti, but in ca. 54 BCE, and from the Republican rostra, but in ca. 27 BCE and ca. 14 CE.

Visibility analysis

Methods

The analysis of gesture visibility relies on the findings of a field experiment conducted by Reference Kopij, Głomb and PopławskiKopij, Głomb, and Popławski,Footnote ⁸² which aimed to establish the maximum distance from which rhetorical gestures are observable. The study involved categorizing gestures sourced from ancient Roman rhetorical treatises into three classes based on their level of detail and the size of the gesturing body part (Fig. 2).

Fig. 2. Examples of gestures used in the experiment: (a–b) gestures of the first class; (c–d) gestures of the second class; (e–f) gestures of the third class (Aldrete Reference Aldrete1999, figures 1, 5, 7, 14, and 15. © 1999 Johns Hopkins University Press. Reprinted with permission of Johns Hopkins University Press.)

The first category involves the fingers, the second the hand, and the third the arm (or arms). The task for participants in the experiment was to identify, at different distances, one of the five gestures in each class, as demonstrated by a “mock speaker.” The procedure was based on similar experiments in psychology.Footnote ⁸³ According to the results, the recognition of gestures in the first class (which depend on the arrangement of the fingers), becomes statistically random at distances ranging from 80 to 85 m from the speaker. Gestures in the second class (which involve recognizing the arrangement and orientation of the entire hand), cease to be visible between 250 and 260 m from the speaker. The limit of visibility of gestures in the third class was not determined in the study because it exceeded the extent of the space in the Forum. In addition, a regression analysis was performed to identify four zones of gesture visibility: Complete/Near-Complete Visibility – C/NCV (1–0.9); High Visibility – HV (0.9–0.7); Average Visibility – AV (0.7–0.4); and Low Visibility – LV (0.4–0.2). Due to insufficient data for milestones close to the speaker, in the case of the first class of gestures it was only possible to identify two of the four visibility zones: the Average Visibility (AV) zone (or rather the combined Complete/Near-Complete Visibility, High Visibility and Average Visibility zones, called the “AV zone” for simplicity’s sake), which reached up to 66 m from the speaker, and the Low Visibility (LV) zone, which lay between 66 and 83 m from the speaker (see Table 1). It is worth noting here that between 25 and 80 m, recognition remained constant (about 40–50%).Footnote ⁸⁴ The visibility zones resulting from the regression analysis can be seen as a proxy for the overall visual acuity of the Roman population, who were unaware of effective ways to improve their vision.

Table 1. Visibility zone intervals for a given gesture class. C/NCV = Complete/Near-Complete Visibility, HV = High Visibility, AV = Average Visibility, LV = Low Visibility. (Kopij et al. Reference Kopij, Głomb and Popławski2023, table 9.)

Visibility analyses were conducted using the Viewshed function in Exploratory 3D Analysis in ArcGIS Pro 2.9, utilizing the virtual 3D reconstructions of venues that had been custom made for each period of the Forum Romanum (Supplementary Files – S4).

In designing the analyses, we relied on the methodology proposed by Reference Kopij, Głomb and PopławskiKopij, Głomb, and Popławski.Footnote ⁸⁵ Accordingly, each study involves two overlapping analyses for points that differ in height by 0.6 m, which reflects the fact that gestures can be displayed at various heights, from abdomen to head level, or even slightly above. We assumed the heights of the points analyzed to be 1.15 m and 1.75 m from the level of the construction upon which the gesturing individual was standing. We adopted this range of heights based on estimates of the average height of male inhabitants of Italy during the Roman period, which Kron estimates at 1.683 m.Footnote ⁸⁶ To maximize the realism of the analyses, the pavement and other areas in the Forum Romanum where spectators might have stood were raised by 1.50 m relative to reality (without simultaneously raising the level of buildings), to ensure that the results of the analyses represented the eye level of the gathered Romans. This was necessary because the software used in the visibility analysis requires a plane to demonstrate visibility or its absence from a specified height. To reveal the potential spaces beneath the porticoes, typically concealed by a roof, we made a horizontal cut, trimming the virtual reconstructions at a specific point and thereby removing the roofs and upper parts of buildings.

We assumed in our analyses that the gestures could be seen distinctly at up to an angle of 75 degrees (especially as the orator could turn his body to maximize the audience), resulting in a total horizontal viewing angle of 150 degrees. This allowed us to consider a speaker’s ability to rotate around their axis, not only exposing hand gestures to people directly in front of them, but also aligning with the “maximum visibility model” we employed (the limitations of the model will be further discussed in the Conclusions).

The visibility analysis, as mentioned, included two groups of hand gestures, which differed in their level of detail. The first group included gestures where it is important to see the arrangement of the fingers of the hand. The second group included gestures where it is important to recognize the arrangement and orientation of the whole hand.Footnote ⁸⁷ The analysis took place in three steps: the first two steps were related to the first class of gestures and involved exploring their maximum visibility and examining them by means of the results of a regression analysis, divided into the zones of visibility. Reference Kopij, Głomb and PopławskiKopij, Głomb, and Popławski were unable to capture the initial two areas of visibility for the first class of gestures, resulting in only the last two levels being utilized in the analysis. In contrast, in the third step, the broader spectrum of visibility levels resulting from the results of the regression analysis was used to study the visibility of the second class of gestures. In this case, the Low Visibility zone was excluded due to the distances obtained extending beyond the venues.

Finally, to examine our hypothesis regarding intentionally separating participants at two simultaneous events to the greatest extent possible, we conducted visibility analyses aimed at determining the level of overlap between hypothetical crowds during various time periods. Additionally, we considered the circumstances that existed after the end of Caesar’s dictatorship and before the commencement of the construction of the temple dedicated to him. These analyses were conducted for both the total visibility of the first-class gestures and for their visibility in the Average Visibility zone.

Results

The first step of the gesture visibility analysis was to study the visibility of only the first class of gestures (Fig. 2a–b), whose maximum visibility distance was determined in the experiment to be between 80 and 85 m.Footnote ⁸⁸ For practical purposes, we adopted the mean of 83 m from the speaker as a threshold.

The area in front of the Republican rostra vetera (ca. 54 BCE) from which the gestures were visibleFootnote ⁸⁹ covered approx. 5,601 m² (Fig. 3). The relevant area in front of the rostra Augusti, on the other hand, covered approx. 5,653 m² for the Forum ca. 27 BCE (Fig. 4) and 5,403 m² for the Forum ca. 14 CE (Fig. 5). The difference between the “capacity” of the rostra vetera (ca. 54 BCE) and the “capacity” of the rostra Augusti (ca. 27 BCE) is not large, amounting to barely 52 m², but the “capacity” of the rostra Augusti (ca. 14 CE) stands out more clearly from both, as the available area is smaller by approx. 200 m² and 250 m², respectively.

Fig. 3. Results of the first class of gesture visibility analysis for the ca. 54 BCE rostra vetera with regression AV for Average Visibility zone; LV for Low Visibility zone; NV for the nonvisible area.

Fig. 4. Results of the first class of gesture visibility analysis for the ca. 27 BCE rostra Augusti with regression AV for Average Visibility zone; LV for Low Visibility zone; NV for the nonvisible area.

Fig. 5. Results of the first class of gesture visibility analysis for the ca. 14 CE rostra Augusti with regression AV for Average Visibility zone; LV for Low Visibility zone; NV for the nonvisible area.

The differences detected for the rostra aedis Castoris are much greater. The space available in front of the temple in the Forum ca. 54 BCE was approx. 5,963 m² (Fig. 6), while for the Forum ca. 27 BCE and the Forum ca. 14 CE it was 4,368 m² (Fig. 7) and 4,465 m² (Fig. 8), respectively. The difference remains very clear even if we exclude the space occupied by the Gradus Aurelii, since then the available space is approx. 5,219 m². In any case, such a step does not seem justified, as arguably the vast majority of that area was available to the participants in contiones. The reduction in available space was primarily due to the construction of the Temple of Divus Iulius (and later also the Arch of Augustus/Parthian Arch). It also seems that the rebuilding of the Basilica Aemilia, especially the addition of the Porticus of Gaius and Lucius, had no negative impact on the amount of space available. On the contrary, the removal of the tabernae novae slightly increased the available space in this part of the Forum.

Fig. 6. Results of the first class of gesture visibility analysis for the ca. 54 BCE rostra aedis Castoris with regression AV for Average Visibility zone; LV for Low Visibility zone; NV for the nonvisible area.

Fig. 7. Results of the first class of gesture visibility analysis for the ca. 27 BCE rostra aedis Castoris with regression AV for Average Visibility zone; LV for Low Visibility zone; NV for the nonvisible area.

Fig. 8. Results of the first class of gesture visibility analysis for the ca. 14 CE rostra aedis Castoris with regression AV for Average Visibility zone; LV for Low Visibility zone; NV for the nonvisible area.

As mentioned earlier, the construction of the Temple of Divus Iulius began in 42 BCE, and the inauguration of both the building itself and the associated rostra took place in 29 BCE. In ca. 27 BCE, the available space in front of the rostra aedis divi Iulii was approx. 5,022 m² (Fig. 9), while ca. 14 CE, it was slightly more, at approx. 5,077 m² (Fig. 10). The increase in available space of approx. 55 m² was related to the demolition of the tabernae novae and the construction of the Porticus of Gaius and Lucius.

Fig. 9. Results of the first class of gesture visibility analysis for the ca. 27 BCE rostra aedis divi Iulii with regression AV for Average Visibility zone; LV for Low Visibility zone; NV for the nonvisible area.

Fig. 10. Results of the first class of gesture visibility analysis for the ca. 14 CE rostra aedis divi Iulii with regression AV for Average Visibility zone; LV for Low Visibility zone; NV for the nonvisible area.

In the next step, we performed a study of the visibility of the first class of gestures using the results of the regression analysis.Footnote ⁹⁰ The visibility zone (with a radius of 83 m), was divided into two sub-zones corresponding to Average Visibility and Low Visibility and covering areas of radius 1–66 m and 66.1–83 m from the speaker, respectively.

For the rostra vetera (ca. 54 BCE), the Average Visibility area covered approx. 4,580 m², while the Low Visibility area covered approx. 1,021 m² (Fig. 3). In ca. 27 BCE, for the rostra Augusti, the Average Visibility area equated to approx. 4,269 m² and the Low Visibility zone to approx. 1,384 m² (Fig. 4), while ca. 14 CE, the Average Visibility zone was approx. 4,186 m² and the Low Visibility zone was approx. 1,227 m² (Fig. 5). The difference of approx. 83 m² between the two periods in the Average Visibility zone is due to the demolition of the tabernae novae and the construction of the Porticus of Gaius and Lucius in front of the Basilica Aemilia. The same developments were responsible for the Low Visibility zone being slightly reduced (by less than 50 m²).

In the case of the Temple of Castor and Pollux, the Average Visibility zone for the Forum ca. 54 BCE equated to approx. 4,823 m² and Low Visibility zone to approx. 1,141 m² (Fig. 6); for the Forum ca. 27 BCE, these zones covered approx. 3,495 m² and approx. 873 m² (Fig. 7), respectively; and for the Forum ca. 14 CE, approx. 3,472 m² and approx. 993 m² (Fig. 8), respectively. As in the analysis of the total visibility of the first class, the differences between the Forum ca. 54 BCE and in later periods are the most apparent. The area of the Average Visibility zone has decreased by as much as approx. 1,328 m² relative to the Forum ca. 27 BCE and approx. 1,351 m² relative to the Forum ca. 14 CE. This difference would be significantly reduced (to approx. 584–607 m²) if the area of the Gradus Aurelii was excluded from the analysis, but, for the reasons we outlined above, such a step does not appear to be justified.

For the case of the rostra aedis divi Iulii, the Average Visibility zone was approx. 3,875 m² ca. 27 BCE (Fig. 9) and 3,934 m² ca. 14 CE (Fig. 10), while the area of the Low Visibility zone covered approx. 1,147 m² ca. 27 BCE and 1,143 m² ca. 14 CE. As can be seen, therefore, the area of the Average Visibility zone increased slightly before ca. 14 CE (by approx. 60 m²), which was probably due in part to the replacement of the tabernae novae with the Porticus of Gaius and Lucius, and to a slight rebuilding of the front of the podium of the speaking platform itself, which allowed the speaker to position himself closer to the center of the rostra. The negligible difference between the areas of the Low Visibility zones, however, suggests that the possible slight shift in the speaker’s position between ca. 27 BCE and ca. 14 CE was of less significance than the changes associated with the construction of the portico adjacent to the Basilica Aemilia.

The final step of the visibility study was the investigation, using the results of regression analyses, of the second class of gestures.Footnote ⁹¹ In this case, due to the limited space, we only considered the results for Complete/Near-Complete Visibility (C/NCV), High Visibility (HV), and Average Visibility (AV) zones, with the plots for the last of these already partly outside the modeled area, meaning they should only be considered a rough approximation.

If we study the rostra vetera (ca. 54 BCE), the area of the Complete/Near-Complete Visibility zone is approx. 246 m², the High Visibility zone approx. 5,915 m², and the Average Visibility zone approx. 2,016 m² (Fig. 11). For the rostra Augusti ca. 27 BCE, the area of the Complete/Near-Complete Visibility zone equals approx. 207 m², the High Visibility zone approx. 6,035 m², and the Average Visibility zone approx. 2,835 m² (Fig. 12). For the rostra Augusti ca. 14 CE, by contrast, the Complete/Near-Complete Visibility zone equals approx. 208 m², the High Visibility zone approx. 5,805 m², and the Average Visibility zone approx. 2,291 m² (Fig. 13).

Fig. 11. Results of the second class of gesture visibility analysis for the ca. 54 BCE rostra vetera with regression C/NCV for Complete/Near Complete Visibility zone; HV for High Visibility zone; AV for Average Visibility zone; NV for the nonvisible area.

Fig. 12. Results of the second class of gesture visibility analysis for the ca. 27 BCE rostra Augusti with regression C/NCV for Complete/Near Complete Visibility zone; HV for High Visibility zone; AV for Average Visibility zone; NV for the nonvisible area.

Fig. 13. Results of the second class of gesture visibility analysis for the ca. 14 CE rostra Augusti with regression C/NCV for Complete/Near Complete Visibility zone; HV for High Visibility zone; AV for Average Visibility zone; NV for the nonvisible area.

As we can see, the demolition of the old speaking platform and the construction of a new one in a different location slightly reduced even the area of the Complete/Near-Complete Visibility zone (by approx. 40 m²). This difference is due not so much to the change of location, and thus the immediate vicinity of the rostra, as to differences in the shapes of the fronts of the platforms. The Republican rostra was constructed on a circular plan (as part of the Comitium), while the front of the rostra Augusti is straight (the rostra Caesaris, not analyzed by us, seems to have duplicated the plan of the Republican rostra, but this was changed after the dictator’s death). The changes in the High Visibility zone are interesting. First, between ca. 54 BCE and ca. 27 BCE, the area of this zone increased (by approx. 120 m²), then it decreased between ca. 27 BCE and ca. 14 CE (by approx. 230 m²). The first development can primarily be explained by the displacement of the rostra to a different location, causing thus, together with other changes in the surrounding built environment, a complete alteration in the geometry of the venue. On the other hand, the decrease in the area of the High Visibility zone for the second class of gestures between ca. 27 BCE and ca. 14 CE is mainly due to the rebuilding of the Basilica Aemilia and the construction of the Porticus of Gaius and Lucius. The changes in the Average Visibility zones – estimates of the area of which, we stress again, must be treated with caution – are due to changes in the location of the speaking platform between ca. 54 BCE and ca. 27 BCE, but also to changes in the built environment: the construction of Basilica Iulia and the Temple of Divus Iulius. The differences between the Forum in ca. 27 BCE and in ca. 14 CE, on the other hand, are due to the construction of both the Porticus of Gaius and Lucius and the Parthian Arch of Augustus.

In the case of the Temple of Castor and Pollux, the area of the Complete/Near-Complete Visibility zone is identical for all periods studied and covers approx. 240 m². Differences become apparent only in zones of lower visibility. The area of the High Visibility zone ca. 54 BCE equals approx. 6,336 m²,Footnote ⁹² ca. 27 BCE approx. 4,622 m², and ca. 14 CE approx. 4,732 m². The area of the Average Visibility zone, on the other hand, (the estimation of which should, however, be treated with caution) equals approx. 5,277 m² ca. 54 BCE, approx. 4,580 m² ca. 27 BCE and approx. 4,273 m² ca. 14 CE (Figs. 14–16).

Fig. 14. Results of the second class of gesture visibility analysis for the ca. 54 BCE rostra aedis Castoris with regression C/NCV for Complete/Near Complete Visibility zone; HV for High Visibility zone; AV for Average Visibility zone; NV for the nonvisible area.

Fig. 15. Results of the second class of gesture visibility analysis for the ca. 27 BCE rostra aedis Castoris with regression C/NCV for Complete/Near Complete Visibility zone; HV for High Visibility zone; AV for Average Visibility zone; NV for the nonvisible area.

Fig. 16. Results of the second class of gesture visibility analysis for the ca. 14 CE rostra aedis Castoris with regression C/NCV for Complete/Near Complete Visibility zone; HV for High Visibility zone; AV for Average Visibility zone; NV for the nonvisible area.

The significant reduction in the High Visibility area between ca. 54 BCE and 27 BCEFootnote ⁹³ was primarily due to the construction of the Temple of Divus Iulius. This building not only took the place of the earlier Gradus Aurelii, which had presumably been available for contionales in front of the Temple of Castor and Pollux, but also obstructed the view for all spectators who might have stood between the Gradus Aurelii/Temple of Divus Iulius and the Basilica Aemilia, as well as to the southeast of the basilica. The slight increase in the area of the High Visibility zone between ca. 27 BCE and ca. 14 CE (approx. 110 m²) is instead explained by the construction of the Porticus of Gaius and Lucius. The area of the Average Visibility zone between ca. 54 BCE and ca. 27 BCE is reduced (by approx. 697 m²) due to the construction of the Temple of Divus Iulius, the construction of the Curia Iulia closer to the Basilica Aemilia than its predecessor, the Curia Cornelia, and the erection of the Arch of Augustus next to the new Imperial rostra. A further reduction in this zone between ca. 27 BCE and ca. 14 CE (by approx. 307 m²) was mainly due to the rebuilding of the Temple of Concordia.

In the case of the rostra aedes divi Iulii, the area of the Complete/Near-Complete Visibility zone is similar for ca. 27 BCE and ca. 14 CE, at approx. 230 m² and 223 m², respectively. Minor differences here may be due to a slight change in the speaker’s position (related to the remodeling of the front of the temple podium) and we should disregard them. There are greater differences between the areas of the High Visibility zone in these periods. In ca. 27 BCE it covers approx. 5,663 m² and in ca. 14 CE, approx. 5,442 m². The difference of approx. 221 m² is due to the construction of the Porticus of Gaius and Lucius. It is worth noting that this is the only case where the replacement of tabernae novae by this structure negatively affected the visibility of gestures. The area of the Average Visibility zone, instead, increased between ca. 27 BCE and ca. 14 CE. In the former case, it covers approx. 5,901 m² and in the latter, approx. 6,102 m². This difference (of approx. 201 m²) is due to the reconstruction of the Temple of Concordia. It is interesting to note that, depending on the location of the speaking platform, this remodeling could have a positive effect on the area of the Average Visibility zone – as in this case – or a negative one, as in the case of the Temple of Castor and Pollux discussed above. This can be attributed to the different geometry of the two venues (Figs. 17–18).

Fig. 17. Results of the second class of gesture visibility analysis for the ca. 27 BCE rostra aedis divi Iulii with regression C/NCV for Complete/Near Complete Visibility zone; HV for High Visibility zone; AV for Average Visibility zone; NV for the nonvisible area.

Fig. 18. Results of the second class of gesture visibility analysis for the ca. 14 CE rostra aedis divi Iulii with regression C/NCV for Complete/Near Complete Visibility zone; HV for High Visibility zone; AV for Average Visibility zone; NV for the nonvisible area.

In addition to the above, we decided to test two alternative scenarios. First, we checked what the visibility of gestures would look like if the Republican rostra remained in place through the Forum ca. 27 BCE and ca. 14 CE, but all other changes occurred as reflected in our reconstructions (Figs. 19–20).

Fig. 19. Results of the first class of gesture visibility analysis for the hypothetical scenario of rostra vetera at the ca. 27 BCE Forum with regression AV for Average Visibility zone; LV for Low Visibility zone; NV for the nonvisible area.

Fig. 20. Results of the first class of gesture visibility analysis for the hypothetical scenario of rostra vetera at the ca. 14 CE Forum with regression AV for Average Visibility zone; LV for Low Visibility zone; NV for the nonvisible area.

Second, we checked what the results of the visibility analysis would look like for the Roman Forum ca. 54 BCE with the Imperial rostra (Fig. 21). The first class of gestures would have been visible from the rostra vetera, ca. 27 BCE, at approx. 6,134 m², and from the rostra vetera, ca. 14 CE, at approx. 5,990 m². In contrast, during ca. 54 BCE, the first class of gestures would be visible from the rostra Augusti over an area of approx. 6,337 m².

Fig. 21. Results of the first class of gesture visibility analysis for the hypothetical scenario of rostra Augusti at the ca. 54 BCE Forum with regression AV for Average Visibility zone; LV for Low Visibility zone; NV for the nonvisible area.

If we consider the results of the regression analysis, the Average Visibility zone of the first class of gestures for the rostra vetera at the Forum Romanum ca. 27 BCE would cover approx. 4,669 m² and at the Forum ca. 14 CE, approx. 4,550 m². In contrast, the Low Visibility zone area would be approx. 1,573 m² and 1,545 m² respectively. For the rostra Augusti on the Forum Romanum ca. 54 BCE, the Average Visibility zone would cover approx. 4,644 m² and the Low Visibility zone approx. 1,693 m².

Estimating crowd sizes

Based on the results presented above, we estimated the maximum sizes of the crowds that could attend the contiones held in the Forum and read the speakers’ gestures intelligibly.Footnote ⁹⁴ To do so, we used the method proposed by Reference Kopij, Głomb and PopławskiKopij, Głomb, and Popławski.Footnote ⁹⁵ First, we used only the results of the maximum visibility analysis of gestures in the first class, taking into account the different crowd density factors of between one and five persons/m². Admittedly, research has shown that the limit of five people/m² could be exceeded under appropriate circumstances, but this was usually only for a short time.Footnote ⁹⁶ The framework we have adopted appears to be the most reasonable in terms of estimating the maximum number of participants at contiones. As shown in Table 2, the rostra vetera could accommodate between approx. 5,601 and 28,005 people, depending on the density of the crowd. In contrast, the rostra Augusti could accommodate between approx. 5,653 and 28,265 people ca. 27 BCE and between approx. 5,403 and 27,015 people ca. 14 CE.

Table 2. Calculation of crowd size for the Republican and Imperial rostra for maximum visibility of the first class of gestures.

Table 3 shows the results of the same analysis for a crowd gathered in front of the Temple of Castor and Pollux. As can be seen, in ca. 54 BCE, the contiones taking place there could be witnessed by between approx. 5,963 and 29,815 people. For the Forum ca. 27 BCE, the crowd decreases to a range between approx. 4,368 and 24,840 people, only to increase again slightly for the Forum ca. 14 CE to between approx. 4,465 and 22,325 people.

Table 3. Calculation of crowd size for the Temple of Castor and Pollux for maximum visibility of the first class of gestures.

In Table 4, we include the results of the crowd estimation for contiones held in front of the Temple of Divus Iulius. Differences between the Forum in ca. 27 BCE and in ca. 14 CE are not significant. In the former case, a contio taking place here could have been witnessed by a crowd of between approx. 5,022 and 25,110 people, while in the latter case, the figures are approx. 5,077 to 28,385 people.

Table 4. Calculation of crowd size for the rostra aedis divi Iulii for maximum visibility of the first class of gestures.

Finally, for comparison, hypothetical scenarios involving the rosta vetera in the Augustan “settings” and the rostra Augusti in the Republican “setting” were also considered. The first scenario aims to demonstrate how the results of the analysis would appear if all changes occurred in the Forum except for the replacement of the rostra vetera with the rostra Caesaris/Augusti. The second scenario assumes that the construction of the rostra Caesaris/Augusti was predominantly planned within the context of the Forum ca. 54 BCE. This will help us discern whether Caesar’s intention was to increase the venue’s capacity, even if, in reality, the rostra Augusti was primarily utilized in the context of the Forum ca. 27 BCE and its subsequent phases. The results of crowd-size estimations for these two cases are shown in Table 5. The hand gestures of a speaker presenting from the rostra Augusti would have been seen by a crowd size ranging from approx. 6,337 to 31,685 people. On the other hand, approx. 6,134–30,670 people could have been in front of the rostra vetera ca. 27 BCE, and approx. 5,990–29,950 ca. 14 CE.

Table 5. Calculation of crowd size for the hypothetical scenarios of the rostra Augusti at the Forum Romanum ca. 54 BCE and the rostra vetera at the Forum Romanum ca. 27 BCE and ca. 14 CE for maximum visibility of the first class of gestures.

However, it is important to acknowledge that research on crowd behavior at similar public events in modern times indicates that crowd density is not the same throughout a venue. Usually, the density is higher closer to the speaker and decreases with distance.Footnote ⁹⁷ Therefore, we decided to take into account the decrease of the crowd density coefficient with the decrease of the gesture visibility resulting from the regression analysis.Footnote ⁹⁸ We assumed that for the area of the Complete/Near-Complete Visibility zone, crowd density reaches a theoretical maximum of five people/m². Then, as the degree of visibility decreases, we reduced the crowd density by one person/m², giving us four people/m² for the High Visibility zone, three people/m² for the Average Visibility zone, and two people/m² in the Low Visibility zone.

The results of estimating crowd size using this method for the rostra vetera and rostra Augusti are shown in Table 6. As can be seen, this approach changes the picture somewhat as, theoretically, the largest crowd able to intelligibly read the gestures of the speakers would have been witnessing a speech from the rostra vetera. In subsequent phases, this number slightly but steadily decreases (by approx. 210 and 550 people ca. 27 BCE and ca. 14 CE, respectively, relative to the Forum ca. 54 BCE). Although the difference in total between the rostra vetera ca. 54 BCE and the rostra Augusti ca. 27 BCE is not large, it is worth noting that the number of people who would theoretically fit into the Average Visibility zone falls by approx. 950 people. The small difference in the total is, therefore, due to the expansion of the Low Visibility zone after the position of the rostra was shifted. This shows that, although the gestures of the speakers were observed by a similar number of people in both periods, in Republican times, many more attendees saw them at least to an average degree.

Table 6. Calculations of crowd sizes for the rostra vetera and rostra Augusti for the first class of gestures including the results of the regression analysis (AV = Average Visibility, LV = Low Visibility).

Table 7 shows the results of a similar analysis for contiones held in front of the Temple of Castor and Pollux. In this case, unsurprisingly, the number of people able to see the gestures drops significantly (by approx. 4,000 people) between ca. 54 BCE and the Forum ca. 27 BCE and ca. 14 CE. In this case, the decrease is noticeable in both the Average Visibility and the Low Visibility zones.

Table 7. Calculations of crowd sizes for the podium of the Temple of Castor and Pollux in all periods for the first class of gestures including the results of the regression analysis (AV = Average Visibility, LV = Low Visibility).

We also estimated the crowd size taking into account the results of the regression analysis for the rostra aedes divi Iulii. The results are included in Table 8. In this case, the number of people who could see the gestures increased slightly between ca. 27 BCE and ca. 14, CE by approx. 169 people. At the same time, it is worth noting that the number of people in the Average Visibility zone increased slightly more, by approx. 177 individuals.

Table 8. Calculations of crowd sizes for the rostra aedes divi Iulii in all periods for the first class of gestures including the results of the regression analysis (AV = Average Visibility, LV = Low Visibility).

The next step in estimating the crowd size for each speaking platform was to use the results of the regression analysis for the visibility of the second class of gestures, which is characterized by a lower level of detail. We reiterate that, in this case, we did not determine the area of the Low Visibility zone, as it extended beyond the area we were modelling (and arguably beyond the area of the Forum Romanum itself), and that the results for the Average Visibility zone should be treated with caution, as parts of this zone may also have extended beyond the modelled area.

The results for the rostra vetera and the rostra Augusti can be seen in Table 9. The trends remain similar to those already observed when estimating crowd size with the first class of gestures. The number of people able to see gestures in the second class increases between ca. 54 BCE and ca. 27 BCE (alongside the change in the location of the rostra) from approx. 30,940 to approx. 33,680 people (i.e., by approx. 2,740 people). Later, however, between ca. 27 BCE and ca. 14 CE, it decreases to approx. 31,130 persons (i.e., by approx. 2,550 persons). It is, however, worth noting that the differences are much smaller if we consider the Complete/Near-Complete Visibility and High Visibility zones as most certain given the limitations of our models. Then, the difference in “capacity” between the rostra vetera and the rostra Augusti is barely ca. 285 people (increasing from approx. 24,890 to. 25,175 people over time). It then drops to approx. 24,260 persons before ca. 14 CE; that is, to a level below even that of the rostra vetera ca. 54 BCE.

Table 9. Calculations of crowd sizes for the rostra vetera and rostra Augusti for the second class of gestures including the results of the regression analysis. C/NCV = Complete/Near-Complete Visibility, HV = High Visibility, AV = Average Visibility.

Table 10 presents the results of estimating the maximum crowd size for the rostra aedis Castoris in all periods. Again, trends resembling those for the first class of gestures are evident. We observe a large decrease in the “capacity” of this venue between ca. 54 BCE and ca. 27 BCE, by as much as approx. 8,950 individuals (from approx. 42,375 to 33,430 persons). Between ca. 27 BCE and ca. 14 CE, there is a much smaller decrease of another approx. 480 individuals. The same is true if we consider only the Complete/Near-Complete Visibility and High Visibility zones: the decrease in the size of the maximum crowd is still very pronounced (approx. 6,860 people); however, in this case, there is an increase in capacity between ca. 27 BCE and ca. 14 CE (by approx. 445 people).

Table 10. Calculations of crowd sizes for the podium of the Temple of Castor and Pollux for the second class of gestures including the results of the regression analysis. C/NCV = Complete/Near-Complete Visibility, HV = High Visibility, AV = Average Visibility.

The results of the same analysis for the rostra aedes divi Iulii, ca. 27 BCE and ca. 14 CE are shown in Table 11. For this venue, the change in capacity is negligible (approx. 320 individuals in favor of the ca. 27 BCE rostra). However, the difference increases if we exclude the Average Visibility area from the analysis, favoring the speaking platform from ca. 27 BCE by approx. 920 persons. In this case, therefore, the trend of change is opposite to that for the crowd estimation based on analyses of the visibility of more subtle gestures.

Table 11. Calculations of crowd sizes for the rostra aedes divi Iulii in all periods for the second class of gestures including the results of the regression analysis. C/NCV = Complete/Near-Complete Visibility, HV = High Visibility, AV = Average Visibility.

To better understand changes in the oratorical taskscape of the Forum Romanum between the decline of the Republic and the Early Imperial period, we also estimated crowd sizes for hypothetical scenarios assuming the embedding of the rostra Augusti in the setting of the Forum Romanum, ca. 54 BCE and the rostra vetera in the setting of the main square of Rome ca. 27 BCE and ca. 14 CE. The results are shown in Table 12. In this case, the largest “capacity,” amounting to an average crowd density factor of 3 (approx. 19,000 people), is the rostra Augusti at the Forum Romanum ca. 54 BCE. Embedding the rostra vetera in the Forum ca. 27 BCE results in a crowd that is smaller by approx. 600 people (“capacity” approx. 18,400), and embedding it in the Forum ca. 14 CE results in a crowd that is reduced by a further approx. 430 people (to approx. 17,970 people).

Table 12. Calculation of crowd size for the hypothetical scenarios of rostra Augusti at the Republican Forum and Republican rostra at Augustan Forum for the maximum visibility of the first class of gestures.

A similar trend can be observed for the same type of exercise taking into account the results of the regression analysis (see Table 13). In this case, however, the difference between the “capacities” of the rostra Augusti ca. 54 BCE and the rostra vetera ca. 27 BCE is much smaller (by approx. 160 individuals). It is worth noting here, however, that the rostra vetera ca. 27 BCE performs slightly better in terms of the “capacity” of the Average Visibility zone (by approx. 80 people), which is partially offset by a smaller Low Visibility zone. The differences between the rostra vetera ca. 27 BCE and ca. 14 CE, in contrast, are similar to the results of the maximum visibility analysis.

Table 13. Calculation of crowd size for the hypothetical scenarios of rostra Augusti at the Republican Forum and rostra vetera at ca. 27 BCE and ca. 14 CE Forum for the first class of gestures including the results of the regression analysis. AV = Average Visibility, LV = Low Visibility.

These results can be juxtaposed with the results for the real-life scenarios, as shown in Table 14, to highlight how the “capacity” of individual venues was affected by changes in the surrounding built environment. As can be seen, in each case, the number of people able to perceive the gestures made by the speakers is higher in the case of the hypothetical scenarios.

Table 14. Comparison of the “capacity” of real venues with hypothetical scenarios for the first gesture class. Hypothetical scenarios are highlighted in grey.

The results of estimating the crowd size are similar when we take into account the regression analysis for the first class of gestures (see Table 15). Again, in each successive period, the hypothetical venues have a larger “capacity” than the venues that were actually used. Using the rostra Augusti in the context of the Forum ca. 54 BCE would increase the size of the crowd who could perceive the gestures of the speakers by as many as approx. 1,530 people. It should be noted, however, that most of this increase (approx. 1,340) is in the Low Visibility zone. An even greater variation in the “capacity” of the two venues (approx. 1,580 persons) is observed for ca. 27 BCE. In that case, however, a large part of the difference is attributable to the Average Visibility zone (approx. 1,200 persons). A similar difference (approx. 1,510 people) in favor of the hypothetical scenario is observed ca. 14 CE. Again, a large part of this difference is attributable to the Average Visibility zone (approx. 1,100 persons).

Table 15. Comparison of the crowd-size estimates of real venues with hypothetical scenarios for the first gesture class. Hypothetical scenarios are highlighted in grey. AV = Average Visibility, LV = Low Visibility.

The charts in Figs. 22–24 show a comparison of the “capacity” of all venues by period. In ca. 54 BCE, the hypothetical rostra Augusti could accommodate the largest number of people theoretically able to perceive the gestures of the speakers (see Fig. 22). Of the real venues, the podium of the Temple of Castor and Pollux was slightly better than the rostra vetera in this respect.

Fig. 22. Comparison of the crowd-size estimates of real venues with hypothetical scenarios for the first gesture class, ca. 54 BCE. AV for Average Visibility zone, LV for Low Visibility zone.

Fig. 23. Comparison of the crowd-size estimates of real venues with hypothetical scenarios for the first class of gestures, ca. 27 BCE. AV for Average Visibility zone, LV for Low Visibility zone.

Fig. 24. Comparison of the crowd-size estimates of real venues with hypothetical scenarios for the first class of gestures, ca. 14 CE. AV for Average Visibility zone, LV for Low Visibility zone.

Furthermore, ca. 27 BCE, the largest “capacity” was a hypothetical speaking platform, this time, the rostra vetera (see Fig. 23). The amount of available space in front of the rostra aedis Castoris is significantly reduced, hence the most capacious venue becomes the rostra Augusti, followed by the rostra aedis divi Iulii. It is worth noting that if it had not been for the construction of the Temple of Divus Iulius, the podium of the Temple of Castor and Pollux would have rivalled the capacity of the rostra Augusti. Depending on the method used to estimate the crowd size, it would have had a capacity of approx. 930 more people (for total maximum visibility and an average crowd density factor of three people/m²) or slightly less, by approx. 400 people (for varying crowd density moderated by the results of the regression analysis). In the latter case, however, it can be noted that the podium of the Temple of Castor and Pollux would accommodate more people in the Average Visibility zone who were able to see the speaker’s gestures (approx. 14,470 compared with approx. 12,810 for the rostra Augusti).

As can be seen from Fig. 24, the hypothetical rostra vetera, ca. 14 CE, would have had the greatest capacity. Considering the changes associated with the construction of the Porticus of Gaius and Lucius, had the Temple of Divus Iulius not been built, the speaking platform on the podium of the Temple of Castor and Pollux would have competed with the rostra vetera for precedence. However, the most prominent venue in this period is the rostra Augusti, followed by the rostra aedis divi Iuli.