| { |
| "paper_id": "N12-1002", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T14:04:41.412014Z" |
| }, |
| "title": "Acoustic-Prosodic Entrainment and Social Behavior", |
| "authors": [ |
| { |
| "first": "Rivka", |
| "middle": [], |
| "last": "Levitan", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "Columbia University", |
| "location": { |
| "postCode": "10027", |
| "settlement": "New York", |
| "region": "NY", |
| "country": "USA" |
| } |
| }, |
| "email": "rlevitan@cs.columbia.edu" |
| }, |
| { |
| "first": "Agust\u00edn", |
| "middle": [], |
| "last": "Gravano", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "Universidad de Buenos Aires", |
| "location": { |
| "country": "Argentina" |
| } |
| }, |
| "email": "gravano@dc.uba.ar" |
| }, |
| { |
| "first": "Laura", |
| "middle": [], |
| "last": "Willson", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "Columbia University", |
| "location": { |
| "postCode": "10027", |
| "settlement": "New York", |
| "region": "NY", |
| "country": "USA" |
| } |
| }, |
| "email": "" |
| }, |
| { |
| "first": "Stefan", |
| "middle": [], |
| "last": "Benus", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "Slovak Academy of Sciences", |
| "location": { |
| "country": "Slovakia" |
| } |
| }, |
| "email": "sbenus@ukf.sk" |
| }, |
| { |
| "first": "Julia", |
| "middle": [], |
| "last": "Hirschberg", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "Columbia University", |
| "location": { |
| "postCode": "10027", |
| "settlement": "New York", |
| "region": "NY", |
| "country": "USA" |
| } |
| }, |
| "email": "" |
| }, |
| { |
| "first": "Ani", |
| "middle": [], |
| "last": "Nenkova", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "University of Pennsylvania", |
| "location": { |
| "postCode": "19104", |
| "settlement": "Philadelphia", |
| "region": "PA", |
| "country": "USA" |
| } |
| }, |
| "email": "nenkova@seas.upenn.edu" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "In conversation, speakers have been shown to entrain, or become more similar to each other, in various ways. We measure entrainment on eight acoustic features extracted from the speech of subjects playing a cooperative computer game and associate the degree of entrainment with a number of manually-labeled social variables acquired using Amazon Mechanical Turk, as well as objective measures of dialogue success. We find that male-female pairs entrain on all features, while male-male pairs entrain only on particular acoustic features (intensity mean, intensity maximum and syllables per second). We further determine that entrainment is more important to the perception of female-male social behavior than it is for same-gender pairs, and it is more important to the smoothness and flow of male-male dialogue than it is for female-female or mixedgender pairs. Finally, we find that entrainment is more pronounced when intensity or speaking rate is especially high or low.", |
| "pdf_parse": { |
| "paper_id": "N12-1002", |
| "_pdf_hash": "", |
| "abstract": [ |
| { |
| "text": "In conversation, speakers have been shown to entrain, or become more similar to each other, in various ways. We measure entrainment on eight acoustic features extracted from the speech of subjects playing a cooperative computer game and associate the degree of entrainment with a number of manually-labeled social variables acquired using Amazon Mechanical Turk, as well as objective measures of dialogue success. We find that male-female pairs entrain on all features, while male-male pairs entrain only on particular acoustic features (intensity mean, intensity maximum and syllables per second). We further determine that entrainment is more important to the perception of female-male social behavior than it is for same-gender pairs, and it is more important to the smoothness and flow of male-male dialogue than it is for female-female or mixedgender pairs. Finally, we find that entrainment is more pronounced when intensity or speaking rate is especially high or low.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstract", |
| "sec_num": null |
| } |
| ], |
| "body_text": [ |
| { |
| "text": "Entrainment, also termed alignment, adaptation, priming or coordination, is the phenomenon of conversational partners becoming more similar to each other in what they say, how they say it, and other behavioral phenomena. Entrainment has been shown to occur for numerous aspects of spoken language, including speakers' choice of referring expressions (Brennan & Clark, 1996) ; linguistic style (Niederhoffer & Pennebaker, 2002; Danescu-Niculescu-Mizil et al., 2011) ; syntactic structure (Reitter et al., 2006) ; speaking rate ; acoustic/prosodic features such as fundamental frequency, intensity, voice quality ; and phonetics (Pardo, 2006) .", |
| "cite_spans": [ |
| { |
| "start": 350, |
| "end": 373, |
| "text": "(Brennan & Clark, 1996)", |
| "ref_id": "BIBREF1" |
| }, |
| { |
| "start": 393, |
| "end": 426, |
| "text": "(Niederhoffer & Pennebaker, 2002;", |
| "ref_id": "BIBREF14" |
| }, |
| { |
| "start": 427, |
| "end": 464, |
| "text": "Danescu-Niculescu-Mizil et al., 2011)", |
| "ref_id": "BIBREF3" |
| }, |
| { |
| "start": 487, |
| "end": 509, |
| "text": "(Reitter et al., 2006)", |
| "ref_id": null |
| }, |
| { |
| "start": 627, |
| "end": 640, |
| "text": "(Pardo, 2006)", |
| "ref_id": "BIBREF15" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "Entrainment in many of these dimensions has also been associated with different measures of dialogue success. For example, Chartrand and Bargh (1999) demonstrated that mimicry of posture and behavior led to increased liking between the dialogue participants as well as a smoother interaction. They also found that naturally empathetic individuals exhibited a greater degree of mimicry than did others. Nenkova et al. (2008) found that entrainment on high-frequency words was correlated with naturalness, task success, and coordinated turn-taking behavior. Natale (1975) showed that an individual's social desirability, or \"propensity to act in a social manner,\" can predict the degree to which that individual will match her partner's vocal intensity. showed that entrainment on backchannel-preceding cues is correlated with shorter latency between turns, fewer interruptions, and a higher degree of task success. In a study of married couples discussing problems in their relationships, Lee et al. (2010) found that entrainment measures derived from pitch features were significantly higher in positive interactions than in negative interactions and were predictive of the polarity of the participants' attitudes.", |
| "cite_spans": [ |
| { |
| "start": 123, |
| "end": 149, |
| "text": "Chartrand and Bargh (1999)", |
| "ref_id": "BIBREF2" |
| }, |
| { |
| "start": 402, |
| "end": 423, |
| "text": "Nenkova et al. (2008)", |
| "ref_id": "BIBREF13" |
| }, |
| { |
| "start": 556, |
| "end": 569, |
| "text": "Natale (1975)", |
| "ref_id": "BIBREF12" |
| }, |
| { |
| "start": 988, |
| "end": 1005, |
| "text": "Lee et al. (2010)", |
| "ref_id": "BIBREF8" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "These studies have been motivated by theoretical models such as Giles' Communication Accommodation Theory (Giles & Coupland, 1991) , which proposes that speakers promote social approval or efficient communication by adapting to their interlocutors' communicative behavior. Another theory informing the association of entrainment and dialogue success is the coordination-rapport hypothesis (Tickle-Degnen & Rosenthal, 1990) , which posits that the degree of liking between conversational partners should be correlated with the degree of nonverbal coordination between them.", |
| "cite_spans": [ |
| { |
| "start": 106, |
| "end": 130, |
| "text": "(Giles & Coupland, 1991)", |
| "ref_id": "BIBREF4" |
| }, |
| { |
| "start": 389, |
| "end": 422, |
| "text": "(Tickle-Degnen & Rosenthal, 1990)", |
| "ref_id": "BIBREF18" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "Motivated by such theoretical proposals and empirical findings, we hypothesized that entrainment on acoustic/prosodic dimensions such as pitch, intensity, voice quality and speaking rate might also be correlated with positive aspects of perceived social behaviors as well as other perceived characteristics of efficient, well-coordinated conversations. In this paper we describe a series of experiments investigating the relationship between objective acoustic/prosodic dimensions of entrainment and manually-annotated perception of a set of social variables designed to capture important aspects of conversational partners' social behaviors. Since prior research on other dimensions of entrainment has sometimes observed differences in degree of entrainment between female-female, male-male and mixed gender groups (Bilous & Krauss, 1988; Pardo, 2006; Namy et al., 2002) , we also examined our data for variation by gender pair, considering female-female, male-male, and female-male pairs of speakers separately. If previous findings extend to acoustic/prosodic entrainment, we would expect female-female pairs to entrain to a greater degree than male-male pairs and female partners in mixed gender pairs to entrain more than their male counterparts. Since prior findings posit that entrainment leads to smoother and more natural conversations, we would also expect degree of entrainment to correlate with perception of other characteristics descriptive of such conversations.", |
| "cite_spans": [ |
| { |
| "start": 816, |
| "end": 839, |
| "text": "(Bilous & Krauss, 1988;", |
| "ref_id": "BIBREF0" |
| }, |
| { |
| "start": 840, |
| "end": 852, |
| "text": "Pardo, 2006;", |
| "ref_id": "BIBREF15" |
| }, |
| { |
| "start": 853, |
| "end": 871, |
| "text": "Namy et al., 2002)", |
| "ref_id": "BIBREF11" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "Below we describe the corpus and annotations used in this study and how our social annotations were obtained in Sections 2 and 3. We next discuss our method and results for the prevalence of entrainment among different gender groups (Section 4). In Sections 5 and 6, we present the results of correlating acoustic entrainment with social variables and objective success measures, respectively. Finally, in Section 7, we explore entrainment in cases of outlier feature values.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "The Columbia Games Corpus consists of approximately nine hours of spontaneous dialogue between pairs of subjects playing a series of computer games. Six females and seven males participated in the collection of the corpus; eleven of the subjects returned on a different day for another session with a new partner.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "The Columbia Games Corpus", |
| "sec_num": "2" |
| }, |
| { |
| "text": "During the course of each session, a pair of speakers played three Cards games and one Objects game. The work described here was carried out on the Objects games. This section of each session took 7m 12s on average. We have a total of 4h 19m of Objects game speech in the corpus.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "The Columbia Games Corpus", |
| "sec_num": "2" |
| }, |
| { |
| "text": "For each task in an Objects game, the players saw identical collections of objects on their screens. However, one player (the Describer) had an additional target object positioned among the other objects, while the other (the Follower) had the same object at the bottom of her screen. The Describer was instructed to describe the position of the target object so that the Follower could place it in exactly the same location on her screen. Points (up to 100) were awarded based on how well the Follower's target location matched the describers. Each pair of partners completed 14 such tasks, alternating roles with each task. The partners were separated by a curtain to ensure that all communication was oral.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "The Columbia Games Corpus", |
| "sec_num": "2" |
| }, |
| { |
| "text": "The entire corpus has been orthographically transcribed and words aligned with the speech source. It has also been ToBI-labeled (Silverman et al., 1992) for prosodic events, as well as labeled for turntaking behaviors.", |
| "cite_spans": [ |
| { |
| "start": 128, |
| "end": 152, |
| "text": "(Silverman et al., 1992)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "The Columbia Games Corpus", |
| "sec_num": "2" |
| }, |
| { |
| "text": "In order to study how entrainment in various dimensions correlated with perceived social behaviors of our subjects, we asked Amazon Mechanical Turk 1 annotators to label the 168 Objects games in our corpus for an array of social behaviors perceived for each of the speakers, which we term here \"social variables.\"", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Annotation of Social Variables", |
| "sec_num": "3" |
| }, |
| { |
| "text": "Each Human Intelligence Task (HIT) presented to the AMT workers for annotation consisted of a single Objects game task. To be eligible for our HITs, annotators had to have a 95% success rate on previous AMT HITs and to be located in the United States. They also had to complete a survey establishing that they were native English speakers with no hearing impairments. The annotators were paid $0.30 for each HIT they completed. Over half of the annotators completed fewer than five hits, and only four completed more than twenty.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Annotation of Social Variables", |
| "sec_num": "3" |
| }, |
| { |
| "text": "The annotators listened to an audio clip of the task, which was accompanied by an animation that displayed a blue square or a green circle depending on which speaker was currently talking. They were then asked to answer a series of questions about each speaker: A series of check questions with objectively determinable answers (e.g. \"Which speaker is the Describer?\") were included among the target questions to ensure that the annotators were completing the task with integrity. HITs for which the annotator failed to answer the check questions correctly were disqualified.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Annotation of Social Variables", |
| "sec_num": "3" |
| }, |
| { |
| "text": "Does Person A/B", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Annotation of Social Variables", |
| "sec_num": "3" |
| }, |
| { |
| "text": "Each task was rated by five unique annotators who answered \"yes\" or \"no\" to each question, yielding a score ranging from 0 to 5 for each social variable, representing the number of annotators who answered \"yes.\" A fuller description of the annotation for social variables can be found in .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Annotation of Social Variables", |
| "sec_num": "3" |
| }, |
| { |
| "text": "In this study, we focus our analysis on annotations of four social variables:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Annotation of Social Variables", |
| "sec_num": "3" |
| }, |
| { |
| "text": "\u2022 Is the speaker trying to be liked?", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Annotation of Social Variables", |
| "sec_num": "3" |
| }, |
| { |
| "text": "\u2022 Is the speaker trying to dominate the conversation? \u2022 Is the speaker giving encouragement to his/her partner?", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Annotation of Social Variables", |
| "sec_num": "3" |
| }, |
| { |
| "text": "\u2022 Is the conversation awkward?", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Annotation of Social Variables", |
| "sec_num": "3" |
| }, |
| { |
| "text": "We correlated annotations of these variables with an array of acoustic/prosodic features.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Annotation of Social Variables", |
| "sec_num": "3" |
| }, |
| { |
| "text": "We examined entrainment in this study in eight acoustic/prosodic features:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acoustic entrainment", |
| "sec_num": "4" |
| }, |
| { |
| "text": "\u2022 Intensity mean \u2022 Intensity max \u2022 Pitch mean \u2022 Pitch max \u2022 Jitter \u2022 Shimmer \u2022 Noise-to-harmonics ratio (NHR) \u2022 Syllables per second", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acoustic entrainment", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Intensity is an acoustic measure correlated with perceived loudness. Jitter, shimmer, and noise-toharmonics ratios are three measures of voice quality. Jitter describes varying pitch in the voice, which is perceived as a rough sound. Shimmer describes fluctuation of loudness in the voice. Noise-to-harmonics ratio is associated with perceived hoarseness. All features were speaker-normalized using z-scores.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acoustic entrainment", |
| "sec_num": "4" |
| }, |
| { |
| "text": "For each task, we define entrainment between partners on each feature f as", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acoustic entrainment", |
| "sec_num": "4" |
| }, |
| { |
| "text": "ENT p = \u2212|speaker1 f \u2212 speaker2 f |", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acoustic entrainment", |
| "sec_num": "4" |
| }, |
| { |
| "text": "where speaker[1,2] f represents the corresponding speaker's mean for that feature over the task.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acoustic entrainment", |
| "sec_num": "4" |
| }, |
| { |
| "text": "We say that the corpus shows evidence of entrainment on feature f if ENT p , the similarities between partners, are significantly greater than ENT x , the similarities between non-partners:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acoustic entrainment", |
| "sec_num": "4" |
| }, |
| { |
| "text": "ENT x = \u2212 i |speaker1 f \u2212 X i,f | |X|", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acoustic entrainment", |
| "sec_num": "4" |
| }, |
| { |
| "text": "where X is the set of speakers of same gender and role as the speaker's partner who are not paired with the speaker in any session. We restrict the comparisons to speakers of the same gender and role as the speaker's partner to control for the fact that differences may simply be due to differences in gender or role. The results of a series of paired t-tests comparing ENT p and ENT x for each feature are summarized in Table 1 . We find that female-female pairs in our corpus entrain on, in descending order of significance, jitter, intensity max, intensity mean, syllables per second and shimmer. They do not entrain on pitch mean or max or NHR. Male-male pairs show the least evidence of entrainment, entraining only on intensity mean, intensity max, and syllables per second, supporting the hypothesis that entrainment is less prevalent among males. Female-male pairs entrain on, again in descending order of significance, intensity mean, intensity max, jitter, syllables per second, pitch mean, NHR, shimmer, and pitch max -in fact, on every feature we examine, with significance values in each case of p<0.01.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 421, |
| "end": 428, |
| "text": "Table 1", |
| "ref_id": "TABREF2" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Acoustic entrainment", |
| "sec_num": "4" |
| }, |
| { |
| "text": "To look more closely at the entrainment behavior of males and females in mixed-gender pairs, we define ENT2 p as follows:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acoustic entrainment", |
| "sec_num": "4" |
| }, |
| { |
| "text": "ENT2 p = \u2212 i |P i,f \u2212 T i,f | |T|", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acoustic entrainment", |
| "sec_num": "4" |
| }, |
| { |
| "text": "where T is the set of the pause-free chunks of speech that begin a speaker's turns, and P is the corresponding set of pause-free chunks that end the interlocutor's preceding turns. Unlike ENT p , this measure is asymmetric, allowing us to consider each member of a pair separately. We compare ENT2 p for each feature for males and females of mixed gender pairs. Contrary to our hypothesis that females in mixed-gender pairs would entrain more, we found no significant differences in partner gender. Females in mixed-gender pairs do not match their interlocutor's previous turn any more than do males. This may be due to the fact that, as shown in Table 1 , the overall differences between partners in mixed-gender pairs are quite low, and so neither partner may be doing much turn-byturn matching.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 647, |
| "end": 654, |
| "text": "Table 1", |
| "ref_id": "TABREF2" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Acoustic entrainment", |
| "sec_num": "4" |
| }, |
| { |
| "text": "However, as we expected, entrainment is least prevalent among male-male pairs. Although we expected female-female pairs to exhibit the highest prevalence of entrainment, they do not show evidence of entrainment on pitch mean, pitch max or NHR, while female-male pairs entrain on every feature. In fact, although ENT p for these features is not significantly smaller between female-female pairs than between female-male pairs, ENT x , the overall similarity among non-partners for these features, is significantly larger between females than between females and males. The degree of similarity between female-female partners is therefore attributable to the overall similarity between females rather than the effect of entrainment.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acoustic entrainment", |
| "sec_num": "4" |
| }, |
| { |
| "text": "All three types of pairs exhibit entrainment on intensity mean, intensity max, and syllables per second. We look more closely into the gender-based differences in entrainment behavior with an ANOVA with the ratio of ENT p to ENT x as the dependent variable and gender pair as the independent variable. Normalizing ENT p by ENT x allows us to compare the degree of entrainment across gender pairs. Results are shown in Table 2 . Male-male pairs have lower entrainment than female-female pairs for every feature; female-male pairs have higher entrainment than female-female pairs for intensity mean and max and lower for syllables per second (p < 0.1). These results are consistent with the general finding that male-male pairs entrain the least and female-male pairs entrain the most.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 418, |
| "end": 425, |
| "text": "Table 2", |
| "ref_id": "TABREF4" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Acoustic entrainment", |
| "sec_num": "4" |
| }, |
| { |
| "text": "We next correlate each of the social variables described in Section 3 with ENT p for our eight acoustic features. Based on Communication Accommodation Theory, we would expect gives encouragement, a variable representing a desirable social characteristic, to be positively correlated with entrainment. Conversely, conversation awkward should be negatively correlated with entrainment. We note that Trying to be liked is negatively correlated with the like more variable in our data -that is, annotators were less likely to prefer speakers whom they perceived as trying to be liked. This reflects the intuition that someone overly eager to be liked may be perceived as annoying and socially inept. However, similarity-attraction theory states that similarity promotes attraction, and someone might therefore entrain in order to obtain his partner's social approval. This idea is supported by Natale's finding that the need for social approval is predictive of the degree of a speaker's convergence on intensity (Natale, 1975) . We can therefore expect trying to be liked to positively correlate with entrainment. Speakers who are perceived as trying to dominate may be overly entraining to their interlocutors in what is sometimes called \"dependency overaccommodation.\" Dependency overaccommodation causes the interlocutor to appear dependent on the speaker and gives the impression that the speaker is controlling the conversation (West & Turner, 2009) .", |
| "cite_spans": [ |
| { |
| "start": 1009, |
| "end": 1023, |
| "text": "(Natale, 1975)", |
| "ref_id": "BIBREF12" |
| }, |
| { |
| "start": 1430, |
| "end": 1451, |
| "text": "(West & Turner, 2009)", |
| "ref_id": "BIBREF19" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Entrainment and social behavior", |
| "sec_num": "5" |
| }, |
| { |
| "text": "The results of our correlations of social variables with acoustic/prosodic entrainment are generally consonant with these intuitions. Although it is not straightforward to compare correlation coefficients of groups for which we have varying amounts of data, for purposes of assessing trends, we will consider a correlation strong if it is significant at the p < 0.00001 level, moderate at the p < 0.01 level, and weak at the p < 0.05 level. The results are summarized in Table 3 ; we present only the significant results for space considerations.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 471, |
| "end": 478, |
| "text": "Table 3", |
| "ref_id": "TABREF6" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Entrainment and social behavior", |
| "sec_num": "5" |
| }, |
| { |
| "text": "For female-female pairs, giving encouragement is weakly correlated with entrainment on intensity max and shimmer. Conversation awkward is weakly correlated with entrainment on jitter. For male-male pairs, trying to be liked is moderately correlated with entrainment on intensity mean and weakly cor-related with entrainment on jitter and NHR. Giving encouragement is moderately correlated with entrainment on intensity mean, intensity max, and NHR. For female-male pairs, trying to be liked is moderately correlated with entrainment on pitch mean. Giving encouragement is strongly correlated with entrainment on intensity mean and max and moderately correlated with entrainment on pitch mean and shimmer. However, it is negatively correlated with entrainment on jitter, although the correlation is weak. Conversation awkward is weakly correlated with entrainment on jitter.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Entrainment and social behavior", |
| "sec_num": "5" |
| }, |
| { |
| "text": "As we expected, giving encouragement is correlated with entrainment for all three gender groups, and trying to be liked is correlated with entrainment for male-male and female-male groups. However, trying to dominate is not correlated with entrainment on any feature, and conversation awkward is actually positively correlated with entrainment on jitter.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Entrainment and social behavior", |
| "sec_num": "5" |
| }, |
| { |
| "text": "Entrainment on jitter is a clear outlier here, with all of its correlations contrary to our hypotheses. In addition to being positively correlated with conversation awkward, it is the only feature to be negatively correlated with giving encouragement.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Entrainment and social behavior", |
| "sec_num": "5" |
| }, |
| { |
| "text": "Entrainment is correlated with the most social variables for female-male pairs; these correlations are also the strongest. We therefore conclude that acoustic entrainment is not only most prevalent for mixed-gender pairs, it is also more important to the perception of female-male social behavior than it is for same-gender pairs.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Entrainment and social behavior", |
| "sec_num": "5" |
| }, |
| { |
| "text": "We now examine acoustic/prosodic entrainment in our corpus according to four objective measures of dialogue success: the mean latency between turns, the percentage of turns that are interruptions, the percentage of turns that are overlaps, and the number of turns in a task. High latency between turns can be considered a sign of an unsuccessful conversation, with poor turntaking behavior indicating a possible lack of rapport and difficulty in communication between the partners. A high percentage of interruptions, another example of poor turn-taking behavior, may be a symptom of or a reason for hostility or awkwardness be- tween partners. We expect these measures to be negatively correlated with entrainment. Conversely, a high percentage of overlaps may be a symptom of a well-coordinated conversation that is flowing easily. In the guidelines for the turn-taking annotation of the Games Corpus (Gravano, 2009) , overlaps are defined as cases in which Speaker 2 takes the floor, overlapping with the completion of Speaker 1's utterance. Overlaps require the successful reading of turn-taking cues and by definition preclude awkward pauses. We expect a high percentage of overlaps to correlate positively with entrainment.", |
| "cite_spans": [ |
| { |
| "start": 903, |
| "end": 918, |
| "text": "(Gravano, 2009)", |
| "ref_id": "BIBREF5" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Entrainment and objective measures of dialogue success", |
| "sec_num": "6" |
| }, |
| { |
| "text": "The number of turns in a task can be interpreted either positively or negatively. A high number is negative in that it is the sign of an inefficient dialogue, one which takes many turn exchanges to accomplish the objective. However, it may also be the sign of easy, flowing dialogue between the partners. In our domain, it may also be a sign of a highachieving pair who are placing the object meticu- lously in order to secure every single point. We therefore expect the number of turns to be positively correlated with entrainment. As before, we consider a correlation strong if it is significant at the p < 0.00001 level, moderate at the p < 0.01 level, and weak at the p < 0.05 level. The significant correlations are presented in Table 4 .", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 734, |
| "end": 741, |
| "text": "Table 4", |
| "ref_id": "TABREF8" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Entrainment and objective measures of dialogue success", |
| "sec_num": "6" |
| }, |
| { |
| "text": "For female-female pairs, mean latency between turns is negatively correlated with entrainment on all variables except pitch max and NHR. The correlations are weak for intensity mean and pitch mean and moderate for intensity max, jitter, shimmer, and syllables per second. The number of turns is moderately correlated with entrainment on intensity max and shimmer and weakly correlated with entrainment on syllables per second. Contrary to our expectations, the percentage of interruptions is positively (though moderately) correlated with entrainment on shimmer; the percentage of overlaps is moderately correlated with entrainment on shimmer and weakly correlated with entrainment on intensity max. Male-male pairs show the most correlations between entrainment and objective measures of dialogue success. The latency between turns is negatively correlated with entrainment on all variables except shimmer and syllables per second; the correlations are moderate for intensity max and NHR and strong for the rest. The number of turns in a task is positively correlated with entrainment on every variable except intensity mean, jitter and shimmer: strongly for NHR; moderately for intensity mean, pitch mean, and pitch max; and weakly for syllables per second.. The percentage of overlaps is moderately correlated with entrainment on intensity mean and max. The percentage of interruptions is moderately correlated with entrainment on NHR.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Entrainment and objective measures of dialogue success", |
| "sec_num": "6" |
| }, |
| { |
| "text": "For female-male pairs, the number of turns is moderately correlated with entrainment on intensity mean and weakly correlated with entrainment on intensity max and shimmer. The percentage of overlaps is moderately correlated with entrainment on shimmer.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Entrainment and objective measures of dialogue success", |
| "sec_num": "6" |
| }, |
| { |
| "text": "For the most part, the directions of the correlations we have found are in accordance with our hypotheses. Latency is negatively correlated with entrainment and overlaps and the number of turns are positively correlated. A puzzling exception is the percentage of interruptions, which is positively correlated with entrainment on shimmer (for femalefemale pairs) and NHR (for male-male pairs).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Entrainment and objective measures of dialogue success", |
| "sec_num": "6" |
| }, |
| { |
| "text": "While the strongest correlations were for mixedgender pairs for the social variables, we find that the strongest correlations for objective variables are for male-male pairs, which also have the greatest number of correlations. It therefore seems that while entrainment is more important to the percep-tion of social behavior for mixed-gender pairs than it is for same-gender pairs, it is more important to the smoothness and flow of dialogue for male-male pairs than it is for female-female or female-male pairs.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Entrainment and objective measures of dialogue success", |
| "sec_num": "6" |
| }, |
| { |
| "text": "Since acoustic entrainment is generally considered an unconscious phenomenon, it is interesting to consider tasks in which a particular feature of a person's speech is particularly salient. This will occur when a feature differs significantly from the norm -for example, when a person's voice is unusually loud or soft. Chartrand and Bargh (1999) suggest that the psychological mechanism behind the entrainment is the perception-behavior link, the finding that the act of observing another's behavior increases the likelihood of the observer's engaging in that behavior. Based on this finding, we hypothesize that a partner pair containing one \"outlier\" speaker will exhibit more entrainment on the salient feature, since that feature is more likely to be observed and therefore imitated.", |
| "cite_spans": [ |
| { |
| "start": 320, |
| "end": 346, |
| "text": "Chartrand and Bargh (1999)", |
| "ref_id": "BIBREF2" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Entrainment in outliers", |
| "sec_num": "7" |
| }, |
| { |
| "text": "We consider values in the 10th or 90th percentile for a feature \"outliers.\" We can consider ENT x , the similarity between a speaker and the speakers of her partner's role and gender with whom she is never paired, the \"baseline\" value for the similarity between a speaker and her interlocutor when no entrainment occurs. ENT p \u2212 ENT x , the difference between the similarity existing between partners and the baseline similarity, is then a measure of how much entrainment exists relative to baseline.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Entrainment in outliers", |
| "sec_num": "7" |
| }, |
| { |
| "text": "We compare ENT p \u2212 ENT x for \"normal\" versus \"outlier\" speakers. ENT p should be smaller for outlier speakers, since their interlocutors are not likely to be similarly unusual. However, ENT x should also be lower for outlier speakers, since by definition they diverge from the norm, while the normal speakers by definition represent the norm. It is therefore reasonable to expect ENT p \u2212 ENT x to be the same for outlier speakers and normal speakers.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Entrainment in outliers", |
| "sec_num": "7" |
| }, |
| { |
| "text": "If ENT p \u2212 ENT x is higher for outlier speakers, that means that ENT p is higher than we expect, and entrainment is greater relative to baseline for pairs containing an outlier speaker. If ENT p \u2212 ENT x is lower for outlier speakers, that means that ENT p is lower than we expect, and pairs containing an outlier speaker entrain less than do pairs of normal speakers, even allowing for the fact that their usual values should be further apart to begin with. The results for t-tests comparing ENT p \u2212 ENT x for \"normal\" versus \"outlier\" speakers are shown in Table 5 . Outlier pairs have higher relative entrainment than do normal pairs for intensity mean and max, shimmer, and syllables per second. This means that speakers confronted with an interlocutor who diverges widely from the norm for those four features make a larger adjustment to their speech in order to converge to that interlocutor.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 558, |
| "end": 565, |
| "text": "Table 5", |
| "ref_id": "TABREF10" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Entrainment in outliers", |
| "sec_num": "7" |
| }, |
| { |
| "text": "An ANOVA shows that relative entrainment on intensity max is higher in outlier cases for malemale pairs than for female-female pairs and even higher for female-male pairs (F=11.33, p=5.3e-05). Relative entrainment on NHR in these cases is lower for male-male pairs than for female-female pairs and higher for female-male pairs (F=11.41, p=6.5e-05). Relative entrainment on syllables per second is lower for male-male pairs and higher for femalemale pairs (F=5.73, p=0.005). These results differ slightly from the results in Table 2 for differences in entrainment in the general case among gender pairs, reinforcing the idea that cases in which feature values diverge widely from the norm are unique in terms of entrainment behavior.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 524, |
| "end": 531, |
| "text": "Table 2", |
| "ref_id": "TABREF4" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Entrainment in outliers", |
| "sec_num": "7" |
| }, |
| { |
| "text": "Our study of entrainment on acoustic/prosodic variables yields new findings about entrainment behavior for female-female, male-male, and mixedgender dyads, as well as the association of entrain-ment with perceived social characteristics and objective measures of dialogue smoothness and efficiency. We find that entrainment is the most prevalent for mixed-gender pairs, followed by femalefemale pairs, with male-male pairs entraining the least. Entrainment is the most important to the perception of social behavior of mixed-gender pairs, and it is the most important to the efficiency and flow of male-male dialogues.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Conclusion", |
| "sec_num": "8" |
| }, |
| { |
| "text": "For the most part, the directions of the correlations of entrainment with success variables accord with hypotheses motivated by the relevant literature. Giving encouragement and trying to be liked are positively correlated with entrainment, as are percentage of overlaps and number of turns. Mean latency, a symptom of a poorly-run conversation, is negatively associated with entrainment. However, several exceptions suggest that the associations are not straightforward and further research must be done to fully understand the relationship between entrainment, social characteristics and dialogue success. In particular, the explanation behind the associations of entrainment on certain variables with certain social and objective measures is an interesting direction for future work.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Conclusion", |
| "sec_num": "8" |
| }, |
| { |
| "text": "Finally, we find that in \"outlier\" cases where a particular speaker diverges widely from the norm for intensity mean, intensity max, or syllables per second, entrainment is more pronounced. This supports the theory that the perception-behavior link is the mechanism behind entrainment and provides a possible direction for research into why speakers entrain on certain features and not others. In future work we will explore this direction and go more thoroughly into individual differences in entrainment behavior.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Conclusion", |
| "sec_num": "8" |
| }, |
| { |
| "text": "http://www.mturk.com", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| } |
| ], |
| "back_matter": [ |
| { |
| "text": "This material is based upon work supported in part by NSF IIS-0307905, NSF IIS-0803148, UBACYT 20020090300087, ANPCYT PICT-2009-0026, CONICET, VEGA No. 2/0202/11; and the EUSF (ITMS 26240220060).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acknowledgments", |
| "sec_num": null |
| } |
| ], |
| "bib_entries": { |
| "BIBREF0": { |
| "ref_id": "b0", |
| "title": "Dominance and accommodation in the conversational behaviours of same-and mixed-gender dyads", |
| "authors": [ |
| { |
| "first": "Amazon Mechanical", |
| "middle": [], |
| "last": "Turk ; Frances", |
| "suffix": "" |
| }, |
| { |
| "first": "R", |
| "middle": [], |
| "last": "Bilous", |
| "suffix": "" |
| }, |
| { |
| "first": "Robert", |
| "middle": [ |
| "M" |
| ], |
| "last": "Krauss", |
| "suffix": "" |
| } |
| ], |
| "year": 1988, |
| "venue": "Language and Communication", |
| "volume": "8", |
| "issue": "3/4", |
| "pages": "183--194", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Amazon Mechanical Turk, http://www.mturk.com. Frances R. Bilous and Robert M. Krauss 1988. Dom- inance and accommodation in the conversational be- haviours of same-and mixed-gender dyads. Language and Communication, 8(3/4):183-194.", |
| "links": null |
| }, |
| "BIBREF1": { |
| "ref_id": "b1", |
| "title": "Conceptual Pacts and Lexical Choice in Conversation. Journal of Experimental Psychology: Learning, Memory and Cognition", |
| "authors": [ |
| { |
| "first": "Susan", |
| "middle": [ |
| "E" |
| ], |
| "last": "Brennan", |
| "suffix": "" |
| }, |
| { |
| "first": "Herbert", |
| "middle": [ |
| "H" |
| ], |
| "last": "Clark", |
| "suffix": "" |
| } |
| ], |
| "year": 1996, |
| "venue": "", |
| "volume": "22", |
| "issue": "", |
| "pages": "1482--1493", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Susan E. Brennan and Herbert H. Clark. 1996. Concep- tual Pacts and Lexical Choice in Conversation. Jour- nal of Experimental Psychology: Learning, Memory and Cognition, 22(6):1482-1493.", |
| "links": null |
| }, |
| "BIBREF2": { |
| "ref_id": "b2", |
| "title": "The Chameleon Effect: The Perception-Behavior Link and Social Interaction", |
| "authors": [ |
| { |
| "first": "Tanya", |
| "middle": [ |
| "L" |
| ], |
| "last": "Chartrand", |
| "suffix": "" |
| }, |
| { |
| "first": "John", |
| "middle": [ |
| "A" |
| ], |
| "last": "Bargh", |
| "suffix": "" |
| } |
| ], |
| "year": 1999, |
| "venue": "Journal of Personality and Social Psychology", |
| "volume": "76", |
| "issue": "6", |
| "pages": "893--910", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Tanya L. Chartrand and John A. Bargh. 1999. The Chameleon Effect: The Perception-Behavior Link and Social Interaction. Journal of Personality and Social Psychology, 76(6):893-910.", |
| "links": null |
| }, |
| "BIBREF3": { |
| "ref_id": "b3", |
| "title": "Mark My Words! Linguistic Style Accommodation in Social Media", |
| "authors": [ |
| { |
| "first": "Cristian", |
| "middle": [], |
| "last": "Danescu-Niculescu-Mizil", |
| "suffix": "" |
| }, |
| { |
| "first": "Michael", |
| "middle": [], |
| "last": "Gamon", |
| "suffix": "" |
| }, |
| { |
| "first": "Susan", |
| "middle": [], |
| "last": "Dumais", |
| "suffix": "" |
| } |
| ], |
| "year": 2011, |
| "venue": "Proceedings of WWW", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Cristian Danescu-Niculescu-Mizil, Michael Gamon, and Susan Dumais. 2011. Mark My Words! Linguistic Style Accommodation in Social Media. Proceedings of WWW 2011.", |
| "links": null |
| }, |
| "BIBREF4": { |
| "ref_id": "b4", |
| "title": "Language: Contexts and Consequences", |
| "authors": [ |
| { |
| "first": "H", |
| "middle": [], |
| "last": "Giles", |
| "suffix": "" |
| }, |
| { |
| "first": "N", |
| "middle": [], |
| "last": "Coupland", |
| "suffix": "" |
| } |
| ], |
| "year": 1991, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "H. Giles and N. Coupland. 1991. Language: Contexts and Consequences. Pacific Grove, CA: Brooks/Cole.", |
| "links": null |
| }, |
| "BIBREF5": { |
| "ref_id": "b5", |
| "title": "Turn-Taking and Affirmative Cue Words in Task-Oriented Dialogue", |
| "authors": [ |
| { |
| "first": "Agust\u00edn", |
| "middle": [], |
| "last": "Gravano", |
| "suffix": "" |
| } |
| ], |
| "year": 2009, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Agust\u00edn Gravano. 2009. Turn-Taking and Affirmative Cue Words in Task-Oriented Dialogue. Ph.D. thesis, Columbia University, New York.", |
| "links": null |
| }, |
| "BIBREF6": { |
| "ref_id": "b6", |
| "title": "Turntaking cues in task-oriented dialogue", |
| "authors": [ |
| { |
| "first": "Agust\u00edn", |
| "middle": [], |
| "last": "Gravano", |
| "suffix": "" |
| }, |
| { |
| "first": "Julia", |
| "middle": [], |
| "last": "Hirschberg", |
| "suffix": "" |
| } |
| ], |
| "year": 2011, |
| "venue": "Computer Speech and Language", |
| "volume": "25", |
| "issue": "3", |
| "pages": "601--634", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Agust\u00edn Gravano and Julia Hirschberg. 2011. Turn- taking cues in task-oriented dialogue. Computer Speech and Language, 25(3):601-634.", |
| "links": null |
| }, |
| "BIBREF7": { |
| "ref_id": "b7", |
| "title": "Acoustic and Prosodic Correlates of Social Behavior. Interspeech", |
| "authors": [ |
| { |
| "first": "Agust\u00edn", |
| "middle": [], |
| "last": "Gravano", |
| "suffix": "" |
| }, |
| { |
| "first": "Rivka", |
| "middle": [], |
| "last": "Levitan", |
| "suffix": "" |
| }, |
| { |
| "first": "Laura", |
| "middle": [], |
| "last": "Willson", |
| "suffix": "" |
| }, |
| { |
| "first": "Stefan", |
| "middle": [], |
| "last": "Benus", |
| "suffix": "" |
| }, |
| { |
| "first": "Julia", |
| "middle": [], |
| "last": "Hirschberg", |
| "suffix": "" |
| }, |
| { |
| "first": "Ani", |
| "middle": [], |
| "last": "Nenkova", |
| "suffix": "" |
| } |
| ], |
| "year": 2011, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Agust\u00edn Gravano, Rivka Levitan, Laura Willson,Stefan Benus, Julia Hirschberg, Ani Nenkova. 2011. Acous- tic and Prosodic Correlates of Social Behavior. Inter- speech 2011.", |
| "links": null |
| }, |
| "BIBREF8": { |
| "ref_id": "b8", |
| "title": "Quantification of Prosodic Entrainment in Affective Spontaneous Spoken Interactions of Married Couples. Eleventh Annual Conference of the International Speech Communication Association", |
| "authors": [ |
| { |
| "first": "Chi-Chun", |
| "middle": [], |
| "last": "Lee", |
| "suffix": "" |
| }, |
| { |
| "first": "Matthew", |
| "middle": [], |
| "last": "Black", |
| "suffix": "" |
| }, |
| { |
| "first": "Athanasios", |
| "middle": [], |
| "last": "Katsamanis", |
| "suffix": "" |
| }, |
| { |
| "first": "Adam", |
| "middle": [], |
| "last": "Lammert", |
| "suffix": "" |
| }, |
| { |
| "first": "Brian", |
| "middle": [], |
| "last": "Baucom", |
| "suffix": "" |
| }, |
| { |
| "first": "Andrew", |
| "middle": [], |
| "last": "Christensen", |
| "suffix": "" |
| }, |
| { |
| "first": "Panayiotis", |
| "middle": [ |
| "G" |
| ], |
| "last": "Georgiou", |
| "suffix": "" |
| }, |
| { |
| "first": "Shrikanth", |
| "middle": [], |
| "last": "Narayanan", |
| "suffix": "" |
| } |
| ], |
| "year": 2010, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Chi-Chun Lee, Matthew Black, Athanasios Katsama- nis, Adam Lammert, Brian Baucom, Andrew Chris- tensen, Panayiotis G. Georgiou, Shrikanth Narayanan. 2010. Quantification of Prosodic Entrainment in Af- fective Spontaneous Spoken Interactions of Married Couples. Eleventh Annual Conference of the Interna- tional Speech Communication Association.", |
| "links": null |
| }, |
| "BIBREF9": { |
| "ref_id": "b9", |
| "title": "Entrainment in Speech Preceding Backchannels", |
| "authors": [ |
| { |
| "first": "Rivka", |
| "middle": [], |
| "last": "Levitan", |
| "suffix": "" |
| }, |
| { |
| "first": "Agust\u00edn", |
| "middle": [], |
| "last": "Gravano", |
| "suffix": "" |
| }, |
| { |
| "first": "Julia", |
| "middle": [], |
| "last": "Hirschberg", |
| "suffix": "" |
| } |
| ], |
| "year": 2011, |
| "venue": "Proceedings of ACL/HLT", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Rivka Levitan, Agust\u00edn Gravano, and Julia Hirschberg. 2011. Entrainment in Speech Preceding Backchan- nels. Proceedings of ACL/HLT 2011.", |
| "links": null |
| }, |
| "BIBREF10": { |
| "ref_id": "b10", |
| "title": "Measuring acoustic-prosodic entrainment with respect to multiple levels and dimensions", |
| "authors": [ |
| { |
| "first": "Rivka", |
| "middle": [], |
| "last": "Levitan", |
| "suffix": "" |
| }, |
| { |
| "first": "Julia", |
| "middle": [], |
| "last": "Hirschberg", |
| "suffix": "" |
| } |
| ], |
| "year": 2011, |
| "venue": "Proceedings of Interspeech", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Rivka Levitan and Julia Hirschberg. 2011. Measuring acoustic-prosodic entrainment with respect to multi- ple levels and dimensions. Proceedings of Interspeech 2011.", |
| "links": null |
| }, |
| "BIBREF11": { |
| "ref_id": "b11", |
| "title": "Gender differences in vocal accommodation: the role of perception", |
| "authors": [ |
| { |
| "first": "Laura", |
| "middle": [ |
| "L" |
| ], |
| "last": "Namy", |
| "suffix": "" |
| }, |
| { |
| "first": "Lynne", |
| "middle": [ |
| "C" |
| ], |
| "last": "Nygaard", |
| "suffix": "" |
| }, |
| { |
| "first": "Denise", |
| "middle": [], |
| "last": "Sauerteig", |
| "suffix": "" |
| } |
| ], |
| "year": 2002, |
| "venue": "Journal of Language and Social Psychology", |
| "volume": "21", |
| "issue": "4", |
| "pages": "422--432", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Laura L. Namy, Lynne C. Nygaard, Denise Sauerteig. 2002. Gender differences in vocal accommodation: the role of perception. Journal of Language and So- cial Psychology, 21(4):422-432.", |
| "links": null |
| }, |
| "BIBREF12": { |
| "ref_id": "b12", |
| "title": "Convergence of Mean Vocal Intensity in Dyadic Communication as a Function of Social Desirability", |
| "authors": [ |
| { |
| "first": "Michael", |
| "middle": [], |
| "last": "Natale", |
| "suffix": "" |
| } |
| ], |
| "year": 1975, |
| "venue": "Journal of Personality and Social Psychology", |
| "volume": "32", |
| "issue": "5", |
| "pages": "790--804", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Michael Natale. 1975. Convergence of Mean Vocal In- tensity in Dyadic Communication as a Function of So- cial Desirability. Journal of Personality and Social Psychology, 32(5):790-804.", |
| "links": null |
| }, |
| "BIBREF13": { |
| "ref_id": "b13", |
| "title": "High-frequency word entrainment in spoken dialogue", |
| "authors": [ |
| { |
| "first": "Ani", |
| "middle": [], |
| "last": "Nenkova", |
| "suffix": "" |
| }, |
| { |
| "first": "Agust\u00edn", |
| "middle": [], |
| "last": "Gravano", |
| "suffix": "" |
| }, |
| { |
| "first": "Julia", |
| "middle": [], |
| "last": "Hirschberg", |
| "suffix": "" |
| } |
| ], |
| "year": 2008, |
| "venue": "Proceedings of ACL/HLT", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ani Nenkova, Agust\u00edn Gravano, and Julia Hirschberg. 2008. High-frequency word entrainment in spoken di- alogue. Proceedings of ACL/HLT 2008.", |
| "links": null |
| }, |
| "BIBREF14": { |
| "ref_id": "b14", |
| "title": "Linguistic style matching in social interaction", |
| "authors": [ |
| { |
| "first": "Kate", |
| "middle": [ |
| "G" |
| ], |
| "last": "Niederhoffer", |
| "suffix": "" |
| }, |
| { |
| "first": "James", |
| "middle": [ |
| "W" |
| ], |
| "last": "Pennebaker", |
| "suffix": "" |
| } |
| ], |
| "year": 2002, |
| "venue": "Journal of Language and Social Psychology", |
| "volume": "21", |
| "issue": "4", |
| "pages": "337--360", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Kate G. Niederhoffer and James W. Pennebaker. 2002. Linguistic style matching in social interaction. Jour- nal of Language and Social Psychology, 21(4):337- 360.", |
| "links": null |
| }, |
| "BIBREF15": { |
| "ref_id": "b15", |
| "title": "On phonetic convergence during conversational interaction", |
| "authors": [ |
| { |
| "first": "S", |
| "middle": [], |
| "last": "Jennifer", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Pardo", |
| "suffix": "" |
| } |
| ], |
| "year": 2006, |
| "venue": "Journal of the Acoustical Society of America", |
| "volume": "119", |
| "issue": "4", |
| "pages": "2382--2393", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Jennifer S. Pardo. 2006. On phonetic convergence dur- ing conversational interaction. Journal of the Acousti- cal Society of America, 119(4):2382-2393.", |
| "links": null |
| }, |
| "BIBREF16": { |
| "ref_id": "b16", |
| "title": "Priming of Syntactic Rules in Task-Oriented Dialogue and Spontaneous Conversation", |
| "authors": [ |
| { |
| "first": "David", |
| "middle": [], |
| "last": "Reitter", |
| "suffix": "" |
| }, |
| { |
| "first": "Johanna", |
| "middle": [ |
| "D" |
| ], |
| "last": "Moore", |
| "suffix": "" |
| }, |
| { |
| "first": "Frank", |
| "middle": [], |
| "last": "Keller", |
| "suffix": "" |
| } |
| ], |
| "year": 1996, |
| "venue": "Proceedings of the 28th Annual Conference of the Cognitive Science Society", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "David Reitter, Johanna D. Moore, and Frank Keller. 1996. Priming of Syntactic Rules in Task-Oriented Di- alogue and Spontaneous Conversation. Proceedings of the 28th Annual Conference of the Cognitive Science Society.", |
| "links": null |
| }, |
| "BIBREF17": { |
| "ref_id": "b17", |
| "title": "Janet Pierrehumbert, Julia Hirschberg. 1992. TOBI: A Standard for Labeling English Prosody. ICSLP-1992", |
| "authors": [ |
| { |
| "first": "Kim", |
| "middle": [], |
| "last": "Silverman", |
| "suffix": "" |
| }, |
| { |
| "first": "Mary", |
| "middle": [], |
| "last": "Beckman", |
| "suffix": "" |
| }, |
| { |
| "first": "John", |
| "middle": [], |
| "last": "Pitrelli", |
| "suffix": "" |
| }, |
| { |
| "first": "Mori", |
| "middle": [], |
| "last": "Ostendorf", |
| "suffix": "" |
| }, |
| { |
| "first": "Colin", |
| "middle": [], |
| "last": "Wightman", |
| "suffix": "" |
| }, |
| { |
| "first": "Patti", |
| "middle": [], |
| "last": "Price", |
| "suffix": "" |
| } |
| ], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "867--870", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Kim Silverman, Mary Beckman, John Pitrelli, Mori Os- tendorf, Colin Wightman, Patti Price, Janet Pierrehum- bert, Julia Hirschberg. 1992. TOBI: A Standard for Labeling English Prosody. ICSLP-1992, 867-870.", |
| "links": null |
| }, |
| "BIBREF18": { |
| "ref_id": "b18", |
| "title": "The Nature of Rapport and its Nonverbal Correlates", |
| "authors": [ |
| { |
| "first": "Linda", |
| "middle": [], |
| "last": "Tickle-Degnen", |
| "suffix": "" |
| }, |
| { |
| "first": "Robert", |
| "middle": [], |
| "last": "Rosenthal", |
| "suffix": "" |
| } |
| ], |
| "year": 1990, |
| "venue": "Psychological Inquiry", |
| "volume": "1", |
| "issue": "4", |
| "pages": "285--293", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Linda Tickle-Degnen and Robert Rosenthal. 1990. The Nature of Rapport and its Nonverbal Correlates. Psy- chological Inquiry, 1(4):285-293.", |
| "links": null |
| }, |
| "BIBREF19": { |
| "ref_id": "b19", |
| "title": "Introducing Communication Theory: Analysis and Application", |
| "authors": [ |
| { |
| "first": "Richard", |
| "middle": [], |
| "last": "West", |
| "suffix": "" |
| }, |
| { |
| "first": "& Lynn", |
| "middle": [], |
| "last": "Turner", |
| "suffix": "" |
| } |
| ], |
| "year": 2009, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Richard West & Lynn Turner. 2009. Introducing Communication Theory: Analysis and Application.", |
| "links": null |
| }, |
| "BIBREF20": { |
| "ref_id": "b20", |
| "title": "Humanities/Social Sciences/Languages", |
| "authors": [ |
| { |
| "first": "", |
| "middle": [], |
| "last": "Mcgraw-Hill", |
| "suffix": "" |
| } |
| ], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "McGraw-Hill Humanities/Social Sciences/Languages, 4th edition.", |
| "links": null |
| } |
| }, |
| "ref_entries": { |
| "TABREF0": { |
| "text": "Trying to be liked? Trying to dominate the conversation? They were also asked questions about the dialogue as a whole: Does it flow naturally? Are the participants having trouble understanding each other? Which person do you like more? Who would you rather have as a partner?", |
| "num": null, |
| "content": "<table><tr><td>believe s/he is better than</td></tr><tr><td>his/her partner? Make it difficult for his/her partner</td></tr><tr><td>to speak? Seem engaged in the game? Seem to dis-</td></tr><tr><td>like his/her partner? Is s/he bored with the game?</td></tr><tr><td>Directing the conversation? Frustrated with his/her</td></tr><tr><td>partner? Encouraging his/her partner? Making</td></tr><tr><td>him/herself clear? Planning what s/he is going to</td></tr><tr><td>say? Polite?</td></tr></table>", |
| "type_str": "table", |
| "html": null |
| }, |
| "TABREF2": { |
| "text": "Evidence of entrainment for gender pairs. A tick indicates that the data shows evidence of entrainment on that row's feature for that column's gender pair.", |
| "num": null, |
| "content": "<table/>", |
| "type_str": "table", |
| "html": null |
| }, |
| "TABREF4": { |
| "text": "Effects of gender pair on entrainment. An arrow pointing up indicates that the group's normalized entrainment for that feature is greater than that of female-female pairs; an arrow pointing down indicates that it is smaller.", |
| "num": null, |
| "content": "<table/>", |
| "type_str": "table", |
| "html": null |
| }, |
| "TABREF6": { |
| "text": "Correlations between entrainment and social variables.", |
| "num": null, |
| "content": "<table/>", |
| "type_str": "table", |
| "html": null |
| }, |
| "TABREF8": { |
| "text": "Correlations between entrainment and objective variables.", |
| "num": null, |
| "content": "<table/>", |
| "type_str": "table", |
| "html": null |
| }, |
| "TABREF10": { |
| "text": "T-tests for relative entrainment for outlier vs. normal speakers.", |
| "num": null, |
| "content": "<table/>", |
| "type_str": "table", |
| "html": null |
| } |
| } |
| } |
| } |