Audience Responses and the Context of Political Speeches

Previous studies showed that cultural dimensions (individualism and collectivism) are related to audience behavior in responding to political speeches. However, this study suggests that speech context is an important issue to be considered in understanding speaker-audience interaction in political speeches. Forms of response, audience behavior, and response rates were analyzed in three speech contexts: acceptance speeches to nomination as political parties’ candidates for presidential election, presidential election campaign speeches, and presidential inauguration speeches in the Korean presidential election of 2012. We found that audience response forms and behavior were distinctive according to the three speech contexts: in-group partisan leadership, competitive, and formal contexts. However, there was no relationship between the affiliative response rate and electoral success in the election. The function of the audience response is popularity and support of a speaker in acceptance and election campaign speeches, while it is conformity to social norms in inauguration speeches.

assessments of the speaker's popularity (Bull & Feldman, 2011;Bull & Miskinis, 2015;West, 1984).They are immediate feedback from the internal audience members in the speech venues, and also an important component for the mass media to evaluate of the speaker's performances to mass audiences by the way they report the speech appearances (West, 1984).The dominance of attention by the audience members and the mass media is a fundamental indicator of leadership (Stewart, 2015).Moreover, the generation of collective audience responses is a characteristic skill of charismatic speakers (Atkinson, 1984a).Correspondingly, the audience members' collective responses to the speaker are a means to display the identity of the group (Bull & Feldman, 2011;Stewart, 2015).Thus, the audience members' collective responses play an important role in the oratorical setting.
In studies of Japanese political speeches (Bull & Feldman, 2011;Feldman & Bull, 2012) based on two general election campaigns (2005 and 2009), the scholars identified six forms of audience response: applause, laughter, cheering, applause followed by cheering, applause and cheering, and verbal responses.Of these six forms, applause was the predominant form of collective audience response in Japanese political speeches.Notably, there was no isolated applause.Thus, Japanese audience members responded to the speeches collectively at every single turn.
A recent study of American speeches in the 2012 presidential election campaigns (Bull & Miskinis, 2015) reported three important findings.First, like British political speeches, both collective and isolated responses occurred during the speeches.Second, there was a greater diversity in the forms of audience response in the American speeches than those for the British and Japanese, with chanting, booing, cheering, applause, and laughter.Third, American speakers invited booing as both affiliative and disaffiliative responses to their speeches.Consequently, audience members displayed not only collective affiliative responses but also collective disaffiliative responses to the speeches.Bull and Miskinis (2015) evaluated the cultural differences in audience response to these political speeches in the context of Hofstede's distinction between individualist and collectivistic cultures (2001 and 2010).
An individualistic society is defined as a culture in which "the ties between individuals are loose; everyone is expected to look after him or herself and his or her immediate family" (Hofstede et al., 2010, p. 92).Individualists focus on individual goals, initiatives, and achievements, and also emphasis an "I" identity.Hence, the individualistic society is a vertical culture in which (1) freedom is important and equality is not valued, (2) individual opinion is regarded as a characteristic of an honest person, and (3) committing crimes make individuals feel guilt and loss of self-respect rather than shame and loss of face.
A collectivistic society is defined as one in which "people from birth onwards are integrated into strong, cohesive in-groups, which throughout people's lifetime continue to protect them in exchange for unquestioning loyalty" (Hofstede et al., 2010, p. 92).Collectivists require that individuals fit into their group, emphasizing a "We" identity.
They focus on the goals, needs, and views of the in-groups.The collectivistic society is a horizontal culture in which (1) individual freedom is not valued but equality is, (2) harmony of the in-group is more important than individual opinion, and (3) people have a strong sense of losing face and maintaining face not only for the individual but also for the group.Therefore, value is placed on cooperation with in-group members.
According to the Hofstede ratings on the cultural dimension, Japan (collectivist) and the USA (individualist) are regarded as polar opposites: individualism scores -USA 91, UK 89, and Japan 46.The UK is categorized as an individualistic culture but not as stridently individualistic as the USA.Given the cultural similarities and differences between the three cultures, Bull and Miskinis (2015) pointed out that individualistic cultures (USA and UK) allow audience members more freedom of activity in response to the political speeches than a collectivistic culture (Japan).As a consequence, there is a diversity of forms of response (both affiliative and disaffiliative), and collective and isolated responses in American speeches, while there are only affiliative responses and collective responses in Japanese speeches.
Moreover, there was a significant positive correlation between affiliative response rate and electoral success.
Obama who generated a higher affiliative response rate than his opponent (Romney) received a higher percentage of the votes than Romney.In contrast, there was no relationship between response rate and electoral success in Japanese speeches (Feldman & Bull, 2012).Hence, while audience responses are indicators of speaker's popularity and electoral popularity amongst audience members in an individualistic society (USA), they do not play such roles in a collectivist society (Japan) but might be seen as indicators of conformity to social norms.Thus, they argued that the cultural differences in response to the political speeches can be understood in terms of individualism and collectivism.However, there is an important limitation in comparing the characteristic features in audience responses between the three cultures due to different speech contexts between the three data sets.Studies of British speeches (e.g., Atkinson, 1984aAtkinson, , 1984b;;Bull, 2000;Bull & Noordhuizen, 2000;Bull & Wells, 2002;Heritage & Greatbatch, 1986) are based on speeches to party political conferences, whereas the studies of Japanese speeches (Bull & Feldman, 2011;Feldman & Bull, 2012) are based on general election campaign speeches.Hence, the two speech contexts can be distinguished in terms of the functions of the political meetings and the audiences.British speeches were delivered to the speakers' political party members at their party conferences, for the purpose of discussing complicated political issues, policies, and events.The Japanese speeches were delivered to the speakers' supporters, for the purpose of expressing appreciation to them and giving them the opportunity to express their support for the speaker.As the speeches in the two contexts were both delivered to in-groups (party members and supporters), affiliative responses were expected.
The study of American speeches (Bull & Miskinis, 2015) is based on presidential election campaign speeches.
Although the both Japanese and American speeches are election campaign speeches, there are differences between the two speech contexts.As described above, the Japanese speeches were delivered only to the supporters of the speaker in expressing appreciation at a kind of community social event rather than seeking to win the support of uncommitted voters.In contrast, the American speeches were delivered at informal public meetings without a pre-selected audience in swing states where no specific candidate or party has overall support.Moreover, there are different levels of importance for the political meetings between the two contexts.The Japanese context is a general election campaign for electing MPs for the general election, whereas the American context is a presidential election campaign which is a much bigger political event in which the speeches were delivered by candidates who might become the future leaders of the nation.Thus, the two speech contexts differ in election campaign atmosphere, levels of enthusiasm of the audience members, and relative speaker status.
For these reasons, although the previous studies show notable differences between the three cultures, it is uncertain whether these distinguishing features can be understood in terms of cultural differences or political contexts.
Therefore, it seems that the previous studies did not give sufficient consideration to the communicative context as a variable in studying speaker-audience interaction in political oratory.The aim of this study is to address this issue.

Speech Contexts and Presidential Election
In order to investigate whether there are distinguishing features in audience response according to the different contexts of political speeches within one culture, this study is based on analyses of Korean political speeches in three different speech contexts: (1) presidential election candidature nomination acceptance speeches in 2012, (2) presidential election campaign speeches in 2012, and (3) presidential inauguration speeches (from the twelfth to the present presidents).The three speech contexts may be distinguished according to the purpose of the speeches and the nature of the audience.
The context of acceptance speeches is complex.The speeches are delivered to party members at the presidential nomination conventions after in-party competitions (nomination contests) and before the main competition (the official presidential election campaign).Thus, the function of the political meeting is to nominate a candidate for the presidential election and launch the election campaign.The purpose of the speech is to accept the nomination, to show appreciation for it, to convey the speaker's visions and pledges for policies, to ask the party members for solidarity to win the presidential election, and to swear to do their best to win the election.Due to these reasons, the context is (1) formal but less formal than presidential inauguration speeches, (2) both post-competition and pre-competition, and (3) different in atmosphere from the annual party political conferences.
The context of election campaign speeches is informal and highly competitive.The nominated candidates representing each party deliver their campaign speeches to voters in various cities during the election campaign tour.
The purpose of the speeches is to win the presidential election.The function of these political meetings is to rally decided voters, to persuade undecided voters, to help them evaluate the speaker's and the opponent's competence and capacity to be a president, to praise one's own party, to condemn the opponent's party or the government, and to convey pledges.
Election events generally contain (1) performances of election campaign songs by the event teams or supporting songs by singers, (2) supporting speeches by the candidates' political colleagues, and (3) main speeches by the candidates.The events normally take place at outdoor locations such as public squares, rail station squares, streets, markets, and university campuses.The candidates deliver their speeches on special stages built for big events or stages built on campaigning cars for small events.Although the election campaign speeches target the wider electorate, typically it is primarily the supporters and the decided voters who attend the campaign speech events.The undecided voters may attend the events; however, the supporters of political opponents hardly ever attend.
The context of presidential inauguration speeches differs from the other two contexts.Although the inaugural speech is a political speech, it is also a ceremonial speech for the inauguration of the national president.Thus, the context of inaugural speeches is more formal than that of acceptance speeches.After the presidential election, the winner delivers his or her presidential inauguration speech to invited representatives from foreign countries, domestic politicians, and the Korean people.The purpose of the speech is clearly distinguishable from the other two contexts.It is to (1) convey a president's vision, directions in managing his/her government, and general polices for each political sphere, (2) pledge to do his/her best in running the government for the nation and country, and (3) ask cooperation for making a hopeful new era.
So far, we have distinguished the three speech contexts according to the explicit speech context: the purpose of the speeches, the function of the political meetings, audience members, atmosphere, formality, competition/noncompetition, and venues.Implicit speech context to consider is leadership.The speakers in this study were all political leaders."Leadership is about…power and influence to set agenda, define identity, and mobilize people to achieve collective goals" (Hogg, 2001, p. 188)."Group members conform to, and thus are influenced by, the prototype…More prototypical members tend to identify more strongly and thus display more pronounced group behaviors; they will be more normative, show greater ingroup loyalty and ethnocentrism, and generally behave in a more group serving manner" (Hogg, 2001, p. 189).Their behavior is based on perceived prototypicality of their leaders (Hogg, 2001;Platow & van Knippenberg, 2001).Political party is an important form of social identity (Green, Palmquist, & Schickler, 2004).Social identity is defined as "the individuals' self-concept which derives from their knowledge of their membership of a social group (or groups) together with the value and emotional significance attached to that membership" (Tajfel, 1982, p. 2).Hence, it is closely related to in-group conceptualization and behavior.
Acceptance speeches were delivered to the members of their political party as leaders of their groups.There were strong in-group leadership and partisan identity due to the party members' collective goal, winning the presidential election.However, in campaign speeches, the speakers have yet to win power; their position is to ask audiences for their support and for their votes.Although the audience are generally supporters of the speakers, they are not in-group party members.Thus, the speakers vocalize their political endorsements to individual identities.Inauguration speeches were delivered to the nation as a whole as leaders of their country.The group identity and collective goals are broader than in acceptance speeches of political parties.Thus, speaker-audience (or leader-follower) status and leadership varies in terms of the three speech contexts.
The acceptance speeches are comparable to the British data because the speeches were delivered to the speakers' party political conventions and indoor venues.The campaign speeches are comparable to the Japanese and American data because the speeches were delivered to the speakers' supporters and citizens during election campaigns.In terms of culture, Japan and Korea are regarded as collectivist (Hofstede et al., 2010), whereas the USA as an individualist society is a polar opposite from Korea.
The analysis of audience response in three different political contexts in Korean speeches presented in this paper may give an opportunity to study whether (1) audience responses occur similarly in a collectivist culture regardless of speech context, whether (2) the absence of isolated response in Japanese speeches is a characteristic feature of collectivist cultures in general, and whether (3) there is a relationship between audience responses rates and electoral success in Korea.Overall, the study was focused on audience behavior in three political speech contexts on three dimensions: forms of response, collective and isolated responses, and response rate (the frequency of collective responses over time).It should be noted that all response forms and incidents identified in this study were displays of affiliation.

Method Sample of Speeches
A total of 21 Korean political speeches from three different contexts were analyzed: (1) four nomination acceptance speeches from the 18th Korean presidential election of 2012, (2) ten speeches during the presidential election campaign of 2012, and (3) seven presidential inauguration speeches from 1981 to 2012.The speeches lasted a total of 7 hours 17 minutes.The 21 speeches are listed in the Appendix; durations are summarized in Table 1.The four acceptance speeches were delivered at indoor venues and lasted a total of 74:04 minutes (mean 18:31 minutes).Of the four nominated candidates, two candidates (Lee JH and Sim SJ) withdrew their candidature during the 22 days of the official campaign period.Hence, the analyses of campaign speeches were based on the other two candidates: Park GH (82 speeches); Moon (79 speeches).Ten speeches were selected for analysis (5 for each candidate), as the same speech contents were often repeated in different cities.Dates, cities, and the average number of audience members were taken into account, in order to select speeches comparable in content and context.The 10 speeches were delivered in outdoor venues and lasted a total of 163:14 minutes (mean 17:00 minutes).
In the inaugural speeches of the first to eleventh presidents, audience responses did not occur except for applause prior to and at the end of the speeches.
i Thus, only the remaining seven speeches (from the twelfth president to the present day) were selected for analysis.The twelfth inaugural speech was delivered at an indoor venue, the other six at outdoor locations, on special stages for the inaugural ceremonies in front of the National Assembly building.The seven inaugural speeches occupied a total of 193:22 minutes (mean 27:66 minutes).

Procedure
Videos and full transcripts of the speeches were collected from websites such as the Presidential History Museum website, party political websites, broadcast archives, online newspapers, and YouTube.The collected videos were compared with other videos on different websites in order to check whether the obtained videos recorded the full duration of the speeches, and whether there had been any editing.It was confirmed that the full duration of each speech was recorded, and that there had been no editing.Each speech was transcribed verbatim into a word processing package and checked against the video data for accuracy.
Content analysis was conducted in terms of two dimensions: forms of response (e.g., applause, cheers, or chanting) and collective/isolated responses.Each audience response was identified and marked on the transcript according to the two dimensions, which were then collated into one coding system sheet for statistical analysis.
The criteria of the coding system are presented below in terms of the two dimensions, together with some illustrative examples.

The Criteria of the Coding System
Feldman and Bull (2012) identified two types of affiliative audience response: unitary and composite.Unitary refers to one form of response, such as applause, cheers, laughter, chanting, and verbal responses.Composite refers to a combined response in which two forms of response co-occur within one audience turn, such as applause and laughter, applause and cheers, and applause followed by cheering.
Applause, cheers, and laughter were coded as simply applause, cheers, and laughter, respectively.In the case of verbal responses, each verbal response was identified and transcribed (e.g., "Yes", "No", "That's right", "President", "By voting", and so on).Where two of these response forms co-occurred, these composites were coded as a combined response using "+".For example, a co-occurrence of applause and cheers was coded as "applause + cheers".
It is noted that one of the forms in a composite response may occur earlier than the other.For example, in "applause + cheers", the cheering may occur first, followed immediately by applause which overlaps with the cheering, the whole incident lasting for eight seconds until it dies down.As applause generally begins within 0.3 seconds of a speaker's completion point and physical initiation of clapping takes approximately 0.2 seconds (Heritage & Greatbatch, 1986), it is possible that cheering is easier to initiate, hence precedes the collective applause.However, cheering may also occur after applause, and if the two forms overlap for an extended period, they are also coded as a composite.
Chanting is coded as chanting.In the Korean context, it is often accompanied by hand movements (e.g., power grip, displaying printed material, or rhythmic clapping).Korean names are usually a compound of three syllables, for example, in Extract 1 (Lines 5-7), when audiences chant a name of a speaker, Moon Jae-In, they make a clap (X) at Moon and no clap (-) at Jae then a clap (X) at In.In this example, the chanting of the name and the co-occurring rhythmic claps last for 5.5 seconds.applause + cheers ((5.9 seconds)) → chanting ((5.5 seconds)) Audience: 5 Note.Single parentheses indicate duration of pause; double parentheses indicate duration of audience response or transcriber's descriptions of events.
In coding a chant, the content was transcribed in italics, together with a translation in parentheses.It should be noted that the audience can also shout the speaker's name without chanting.In chanting, the audience repeats the speaker's name with rhythm and claps, whereas, in verbal responses, the audience simply shouts the name once without rhythm and claps.In order to distinguish between verbal responses and chanting, the rhythmical claps for chanting were presented by "X".Verbal responses do not have this additional annotation.
The above extract was a typical example in which the audience chanted collectively.In some incidents, chanting occurred (1) without the rhythmical claps, but with the characteristic three syllables rhythm, (2) in cooperation between an MC (Master of Ceremonies) and the audience, or (3) in cooperation with an audience divided into two groups.For example, the audiences chanted the speaker's name and then the MC chanted "president".In so doing, the chanting occurred rhythmically between the audience and the MC.In some cases, half of the audience chanted the speaker's name and then the other half chanted "president".All the examples displayed characteristic features of chanting in Korean political speeches: repetition, rhythm, syllables, rhythmical claps, affiliation, and collective responses.
In addition to unitary and composite responses, it is a necessary to introduce a third category: sequential responses.
It was observed that audience members displayed one form of response (either unitary or composite) then extended their turn by shifting to another form of response.For example, they might begin with "applause + cheers" (first form) then move on to chanting (second form).
Where two or more of the forms of audience response occurred one after another within one turn, the sequence is represented by the symbol "→".For example, in Extract 1 above, audience members responded immediately after the speaker's completion point (Line 4, within 0.2 seconds) with "applause + cheers" (Line 5) for 5.9 seconds, and then moved on to chanting ("applause + cheers → chanting"); the chanting lasted for a further 5.5 seconds.
Extract 2 below illustrates a more complex audience turn, involving a sequence of three forms of responses: applause + cheers → chanting → applause.After the speaker's greeting (Lines 1-3), the audience responds with "applause + cheers" overlapping with the speaker's next greeting sentence "I greet you with a bow" (Line 4), and then moves on to chanting (Lines 5-6).They chant the speaker's name six times with rhythmical clapping, then continue their turn by transferring the chanting to applause (Line 6).The whole audience turn lasts for 22.4 seconds.In sequential responses, incidents in which transition from one form of response to another form of response occurred only through an isolated response were coded according to the predominant collective responses.For example, in audience turn (Line 4) of Extract 3 below, "applause + cheers" occur collectively within 0.3 seconds of the completion point of the speaker's message (Line 3), then only two or three audience members moved to chanting while the majority of audience members completed their turn with "applause + cheers".This incident was coded as "applause + cheers" not as "applause + cheers → chanting".Although this incident was coded as a collective response for statistical analysis, isolated forms of response in the sequential response were also identified and marked using "(i)" for further analysis.This particular incident is referred to as a heterogeneous response.Heterogeneous responses were considered to occur as follows: (1) a majority of audience members display one form of response while one or two audience members display an alternative form of response (either unitary or composite); (2) the collective response is followed by isolated response (e.g., collective applause followed by isolated cheers) in sequence.There were 58 incidents of heterogeneous responses: no incidents in acceptance speeches, 9 incidents in inauguration speeches, 49 incidents in campaign speeches.
All response incidents were then categorized in terms of the 3 dimensions as described above (unitary, composite, and sequential).As a result of this procedure, all response incidents were categorized into 12 forms: 5 unitary (applause, laughter, cheers, chanting, and verbal responses), 2 composite (cheers + verbal; applause + cheers), and 5 sequential responses (cheers → chanting; applause + cheers → chanting; applause + cheers → various; applause + cheers → chanting → various; and verbal response → various).It is noted that (1) categorizing sequential responses was conducted by focusing on the first action, and (2) "various" means that there were various further responses after the first or second response.The 12 forms are listed below together with brief descriptions (Table 2).(2) Verbal other (e.g., name of speaker, party, or government; positive or negative response -'that's right', 'that's not right'; campaign slogans) Verbal response 11.
Verbal occurred then moved to other forms

Reliability
A random sample (N = 100) of audience response incidents (10% of the total sample, N = 964, from 21 speeches) were coded by an independent rater, a native speaker of Korean.There was a high level of agreement between the main coder and the independent coder for the audience response forms (k = .91).

Forms of Audience Response in Three Speech Contexts
Initially, five basic forms of response were identified: applause, laughter, cheers, chanting, and verbal responses.
Audience responses were displayed with (1) one of the basic forms (unitary responses), (2) combined forms of the basic forms (composite responses), or (3) transferring one form to another form (sequential responses).Table 3 shows the relative proportions of collective audience responses for the 12 categories according to the three speech contexts.The results showed that there were three distinguishing features for each of the three speech contexts.First, acceptance and campaign speeches showed a greater diversity of response forms than inauguration speeches, where only two forms of response occurred (applause; applause + cheers).In acceptance speeches, seven forms of response occurred (applause; cheers; applause + cheers; applause + cheers → chanting; applause + cheers → chanting → various; verbal response; and verbal response → various).In campaign speeches, twelve forms of response occurred (applause; laughter; cheers; chanting; cheers → chanting; cheers + verbal; applause + cheers; applause + cheers → chanting; applause + cheers → various; applause + cheers → chanting → various; verbal response; and verbal response → various).Thus, the degree of diversity of response form can be illustrated as follows: inauguration (2 forms) < acceptance (7 forms) < campaign (12 forms).
Further analyses of each of the 12 response forms were conducted, using Kruskal Wallis H Tests and Mann Whitney U Tests.These non-parametric tests were utilized, because the data were not normally distributed, and because there were unequal Ns between each of the 3 conditions (4 acceptance speeches, 7 inaugural speeches, and 10 campaign speeches).Because a large number of tests were carried out (multiple comparisons), only results significant at the .01level were accepted.
The results of the Kruskal Wallis H Tests for the 12 response forms are shown in Table 4.There were four results significant at the .01level: applause + cheers → chanting; verbal → various; verbal response; and applause.
Pairwise comparisons using Mann Whitney U Tests were then carried out for these four response forms, the results of which are shown in Table 5.Three response forms occurred with significantly greater frequency in campaign than in inauguration speeches: applause + cheers → chanting (U = 3.5, p < .002);verbal → various (U = 0, p < .001);verbal response (U = 0, p < .001).In contrast, applause occurred with significantly greater frequency in inauguration than in campaign and acceptance speeches (U = 0, p < .001;U = 0, p < .008).Third, collective audience behaviors in extending a turn are also displayed differently according to the speech contexts.Table 6 shows the 12 forms grouped into non-sequential responses (seven forms; 1-4, 6, 7, and 11) and sequential responses (five forms; 5, 8, 9, 10, and 12).From this table, it can be seen that sequential responses did not occur in the inauguration speeches, audience members completing all of their turns within the first action.On the other hand, in campaign speeches, over one third of the audience responses were sequential.The audience members extended 34.74% of their turns, by moving the first response form to the second and third response forms.Hence, the frequency of turn extension (sequential response) behavior can be illustrated as follows: inauguration (0%) < acceptance (8.75%) < campaign (34.74%).It shows that there are significantly different sequential and non-sequential response behaviors between inauguration and acceptance contexts (U = 3.5, p < .012),and inauguration and campaign contexts (U = 0, p < .001),while there were no significant differences between acceptance and campaign contexts.

Collective and Isolated Responses
As shown in Table 7, Korean audiences responded to the speeches more collectively (95.24% in the acceptance context; 91.75% in the campaign context; 98.88% in the inauguration context) than with isolated responses.Isolated responses also occurred in all three contexts.However, these incidents accounted for small proportions of the total response events.Among the three contexts, campaign speeches received the highest proportion of isolated responses (8.25%) while acceptance and inauguration speeches received 4.8% and 1.1% of isolated responses respectively.Thus, the frequency of isolated response can be illustrated as follows: inauguration (1.12%) < acceptance (4.76%) < campaign (8.25%).It shows that there are significantly different collective and isolated response behaviors between inauguration and campaign contexts (U = 4.5, p < .002),while there were no significant differences between the other two paired contexts.All collective and isolated responses occurred at the end of statements in acceptance and inauguration speeches, whereas in campaign speeches, there were 13 incidents of audience interruption (responses occurred in the middle of the statements).

Response Rate
Table 8 shows overall collective audience response rates per minute of speech according to the three contexts.
The table indicates that audience members responded to campaign speeches more frequently than the other two speech contexts.When calculated as a rate per minute, campaign speeches received an average of 3.6 collective responses, while acceptance and inauguration speeches received 2.1 per minute and 0.9 per minute collective responses respectively.Hence, the response rate can be illustrated as follows: inauguration (0.9) < acceptance (2.1) < campaign (3.6).It shows that response rates were significantly higher in campaign context compared to inauguration context (U = 0, p < .001),while there were no significant differences between acceptance and campaign contexts.Table 9 shows response rates and percentage of votes received in the each region according to the two speakers.

Electoral Success
Although Moon generated more collective audience responses than Park in the each region, he received less votes than Park except in Seoul.The results indicate that there was no relationship between audience response rate and electoral success in the Korean presidential election 2012.

Characteristic Features of Audience Responses in Each Context
From the results, it can be seen that inauguration speeches and election campaign speeches are significantly different genres of political oratory.There were distinctive audience behaviors for each context, in particular with regard to their relative degree of formality, competitiveness, and in-group leadership.First, there is a characteristic predominant form of response for each context: applause in formal and ceremonial context of inauguration speeches; applause + cheers in partisan in-group context of acceptance speeches, and verbal responses in informal and competitive context of election campaign speeches.
Second, the more formal the speech context (inauguration > acceptance > campaign), the fewer types of audience response, the fewer isolated responses, the fewer sequential response incidents, and the fewer audience responses.
In inauguration speeches, only two response forms occurred: applause and applause + cheers.Conversely, in the more informal and competitive context of election campaign speeches, there were a much greater diversity of response forms, more frequent audience responses, and more frequent isolated responses.Furthermore, both collective and isolated responses were more likely to be interruptive of speaker statements in campaign speeches than in acceptance speeches, where audience members were more likely to respond collectively at the end of speakers' messages.Thereby, in that in-group, partisan context, audience members arguably displayed clear collective action in support of their leaders.
Third, chanting and sequential responses are particularly characteristic of audience behavior in Korean political speeches, but only in acceptance and campaign speeches.Incidents of chanting were more frequent in campaign speeches (1.0 responses per minute) than in acceptance speeches (0.2 responses per minute), hence there is a clear association of chanting with more informal and competitive settings, thereby supporting political leaders and issues, and affirming the audiences' group identity.
Notably, all the incidents of chanting occurred as the second or third actions in a sequential response (apart from six incidents of unitary chanting in campaign speeches).As illustrated in Extracts 1 and 2 in the Method section above, audience members responded initially with "applause + cheers" in their first action, then extended their turns with chanting as the second action.Arguably, audience members displayed their approval of or agreement with the speakers' messages in their first action.When the messages are popular and audience members approve or agree strongly, they extended their turn with chanting in the second action.Thereby, they can support and encourage the speakers, and display their enthusiastic support.
Acceptance and campaign speeches are characterized by much more active participation from the audience, especially in campaign speeches, where they respond with various verbal responses to support the speaker, to agree or disagree with political issues, and to attack the opponent.They then extend their turn by shifting to another form of response.However, in the inauguration context, sequential responses and chanting did not occur.
Why did sequential responses and chanting occur in acceptance and campaign speeches (competitive contexts), but not in inauguration speeches (uncompetitive and ceremonial context)?As nomination conventions are generally reported through the media, nominating candidates and their acceptance speeches are announced and delivered not only in the convention hall but also to distant audience members.Moreover, during an election campaign, the media report the speech events every day.Thus, collective audience behaviors, the group-presentation, are important factors in showing the popularity of speakers to distant audience members through the media.Such collective audience behaviors are crucial tools in supporting and encouraging the candidates and in presenting solidarity of the supporters in the competitive context.Hence, it can be suggested that (1) the collective audience responses are a crucial part of the speeches in acceptance and campaign speeches, and (2) the speeches are generated by cooperation between a speaker and the audience members to reach their shared goal.

Cultural Differences and Speech Context
Table 10 presents a summary of the cultural dimension and audience responses according to the four countries: USA, UK, Japan, and Korea.It is noted that the Korean section is presented based on the election campaign context in order to compare with American and Japanese election campaign contexts.As presented in the table, Korea and the USA might be seen as even more polar opposites than Japan and the USA.However, unlike the results from studies of Japanese campaign speeches, isolated responses did occur in Korean speeches: 1.12% in inauguration, 4.76% in acceptance, and 8.25% in campaign contexts.In American election speeches, isolated responses occurred throughout (Bull & Miskinis, 2015).
The predominant audience response form is different for each country.In the USA, it was cheering (66.95%, a mean of the two speakers, in 2012 presidential election campaign speeches), whereas, in Japan, it was applause (58.66%) in 2005, and applause (39.72%) and laughter (38.89%) in 2009 general election campaign speeches.
In Korea, it varied according to the three speech contexts: verbal response categories (47.80%) was the predominant form in election campaign speeches.
Although only affiliative responses occurred in Korean election speeches, there was a greater diversity of collective audience response forms and behaviors than in American election speeches.Furthermore, Korean audience members responded to campaign speeches more frequently than American audience members: 3.6 responses (per minute) in Korean speeches, 2.4 responses in American speeches (Bull & Miskinis, 2015), and 0.5 responses ii in Japanese speeches (Feldman & Bull, 2012).
The analysis showed no relationship between response rate and electoral success in the Korean presidential election of 2012, unlike the American context but like the Japanese context.However, there are notable speech contextual differences between the Japanese and Korean election campaign speeches: the purpose of the political meetings, election events, venues, audience members, and speakers.
In addition, the Japanese speakers received an average 0.5 responses per minute, whereas, Korean speakers received an average 3.6 responses per minute.Thus, audience response in Korean speeches occurred seven times more frequently than in Japanese speeches.These mean that speaker-audience interaction in Korean speeches occurred more actively and enthusiastically than in Japanese speeches due to the specific political events of Korean data.Also, the more the battle was heated, the more the audience response rate was higher.As shown in the Figure 4, both candidates' response rates reached the highest rates at the last day of the campaign with 4.6 (Park) and 5.7 (Moon).Hence, we suggest that the function of the audience responses in the Korean election campaign speeches is not conformity to social norms but to show speaker's popularity, support the speaker, and achieve a collective goal in the informal and competitive political event.However, in an inauguration speech context which is high formal and ceremonial political event, 99% of audience responses were collective behavior, and chanting and sequential response did not occurred.Thus, it is possible that the function of audience response in inauguration is conformity to social norms.
There are similarities between USA and Korea election campaign speech context: presidential election campaign, venues, speakers, open-air, and the function of the political meetings.However, there are also different presidential election voting system between USA (electoral vote system) and Korea (popular vote system).In USA, it is the Electoral College that votes in the presidential election, which is whoever wins a state, takes all the Electoral College votes for that state.Therefore, it is critical to win each state, especially, the swing states.Due to the system, the composition of the audience members in the swing states may be different from the other states where a single candidate or party has overwhelming support.As the rallies in the USA are open meetings, both affiliative and disaffiliative responses occurred, hence affiliative response rate may be significant.
In Korea, as it is a popular vote system, the total percentage of votes received is critical.Hence, the size of the audience at each campaign event is important.During the election campaign, it is an important issue for the media to report the number of people attending the events, enthusiastic collective audience responses to the speakers, and atmosphere of the events.They are measurements of the popularity of the speakers in the media and they may influence the media audiences.This may be a possible reason why supporters attend the events and opponents hardly attend the event in Korean political culture.Thus, the different election systems and campaign culture may affect the composition of the audience in the speech events between USA and Korea.Moreover, this can be a crucial reason why only affiliative responses occurred in Korean speeches, while there were both affiliative and disaffiliative responses in American speeches delivered in the swing states.
Like the USA, there are also regions that support one party strongly in Korea.However, regardless of the regions, Moon received a higher response rate in each region than Park who won the election.This can be understood by group identities.The Korean presidential election of 2012 was regarded not only as a battle between the progressive (Moon) and the conservative (Park), but also between the younger generation and the older generation.Moon was supported more by progressive younger generations (20's, 30's, and 40's) than conservative older generation (50's and over 60's): the results of exit poll showed clearly that Moon and Park received higher percentage of votes from the younger generation and the older generation, respectively.It was also observed that the composition of the audience differed in the speech events of the two sides: younger generation of audience members in Moon's speeches and older generation of audience members in Park's speeches.However, the older generation participated more in voting than the younger generation.As a result, Park won the election.Notably, coding results in this study showed that audience members in Moon's speeches displayed more isolated responses and interruptive responses than audience members in Park's speeches.Thus, it seems that there were different audience inclination and behavior on the group polarization between the younger generation (progressive) and the older generation (conservative) in responding to the speakers.Arguably, the younger and progressive audiences are more individualistic and free in reacting to the political leaders than the old and conservative audiences, while the old and conservative audiences display more collective behavior.
Therefore, we suggest that although the relationship between speech rate and electoral success differ between USA and Korea, this may be not understood by the cultural dimensions solely but political culture, election system, and group polarization (partisans and generation identities).Accordingly, it is necessary to conduct future research in audience response in various political speech contexts.In so doing, we can understand speaker-audience interaction in political oratory further.

Conclusion
This study has demonstrated micro-analysis on audience behavior to political oratory and a number of distinctive features in audience responses for each of three speech contexts and cultures.As each context shows characteristic audience responses, we propose that political speech context is an important factor in studying audience behavior which possibly overrides cultural dimensions.From this perspective, the results of previous studies on cultural differences in political oratory are limited by their focus on only one speech context, hence their findings are incomplete.
Social and political psychology has placed relatively little focus on the study of audience behavior in political speeches.Studies of political speeches have been conducted predominantly in a western political context, emphasizing speaker's speech content and rhetoric.However, the systematic micro-analysis on audience behavior in Korean political oratory shows important findings and implications in social and political action.(1) The function of audience response is different in terms of in-group partisan leadership, competitive, and formal contexts.(2) There is no relationship between response rate and electoral success in Korea.However, audience behavior is an important tool in analyzing group identity, inclination, and polarization in presidential election campaign speeches.(3) Speech context, political culture, and the election systems are closely related to audience behavior beyond collectivistic and individualistic cultural dimensions.In conclusion, we propose that this detailed microanalysis of audience behavior in a previously unresearched culture has given us novel insights into the speaker-audience interaction in three different contexts of political speech-making.

Notes
i) It is noted that audios (or videos) of the third and the eighth inaugural speech were not archived in the Presidential History Museum.Due to this limitation, it cannot be confirmed whether there was an absence of audience response in the two speeches.
ii) This response rate was not presented in the original paper; the corresponding author calculated from the available data (p.389).
Audience Responses in Political Speeches 620 Note. [ ] indicates overlaps between audience response and a speaker's utterance.

Figure 1 .
Figure 1.Collective audience responses in acceptance speeches.

Figure 4
Figure 4 shows collective audience response rates in each candidate's speeches according to the date (first day to last day).As presented, Moon generated more collective audience responses (average 4.3 responses per minute) than Park (average 2.8 responses per minute) at all speech events.However Park won the election: the election results, Park 51.6% and Moon 48.0%.

Figure 4 .
Figure 4. Response rate by date in campaign speeches.

Table 1
Speeches and Duration by Three Contexts

Table 2
Forms of Response

Table 3
Forms of Audience Response by Three Speech Contexts

Table 4
Significant Tests Between Three Contexts

Table 5
Significant Tests Between Two Speech Contexts

Table 6
Non-Sequential and Sequential Response by the Three Speech Contexts

Table 7
Collective and Isolated Response by the Three Speech Contexts

Table 8
Collective Audience Response Rate by the Three Speech Contexts

Table 9
Response Rate and Electoral Success.

per minute (percentage of votes received)
a Additional speeches (see Appendix).b Mean of 08, 15, 18 December in Seoul.

Table 10
Summary of Cultural Dimensions and Audience Response by CountriesThe scores are based on the research ofHofstede at el. (2010, pp.95-97).bIndicateslimitations in the previous study. a