1 / 41

Charisma Perception from Text and Speech

Charisma Perception from Text and Speech. Andrew Rosenberg NLP Group Meeting 11/03/05. Overview. Background Speech Study Transcript Study Conclusion & Future Work. Overview. Background What is charisma? Does charismatic speech exist? Why study charismatic speech? Speech Study

keefer
Download Presentation

Charisma Perception from Text and Speech

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Charisma Perception from Text and Speech Andrew Rosenberg NLP Group Meeting 11/03/05

  2. Overview • Background • Speech Study • Transcript Study • Conclusion & Future Work NLP Group Meeting

  3. Overview • Background • What is charisma? • Does charismatic speech exist? • Why study charismatic speech? • Speech Study • Transcript Study • Conclusion NLP Group Meeting

  4. Background - What is charisma? (What do I mean by charisma?) • Not “closed door”, face-to-face charisma. • Rather, political (or religious) charisma • The ability to attract, and retain followers by virtue of personality as opposed to tradition or laws. (Weber ‘47) • E.g. Ghandi, Hitler, Che Guevara. • Charismatic speech: Speech that encourages listeners to perceive the speaker as “charismatic”. NLP Group Meeting

  5. Background - Is there such a thing as charismatic speech? • Pro: • Potential charismatic leaders must communicate with would-be followers. • Charismatic leaders have historically had a particular gift at public speaking • Hitler, MLK Jr., Castro. • Con: • Charisma as a relationship between leader and followers. • The mythologizing of a charismatic leader extends beyond public address. NLP Group Meeting

  6. Background - Why study charismatic speech? • General scientific interest. • Feedback system for politicians and academic instructors. • Identification of potential charismatic leaders • Automatic generation of “charismatic-like” speech NLP Group Meeting

  7. Overview • Background • Speech Study • Questions Addressed • Experiment Design • Analyses of Responses • Transcript Study • Conclusion NLP Group Meeting

  8. Speech Study - Questions • Do subjects agree about what is charismatic? • What do subjects mean by charismatic? • What makes speech charismatic? NLP Group Meeting

  9. Speech Study - Experiment Design • Subjects: Friends and colleagues, not compensated monetarily • Interface: Presentation of 45 short speech segments (2-30secs) via a web form • Dependent variables: 5-point Likert scale ratings of agreement on 26 statements. • Duration: avg. 1.5 hrs, min 45m, max ~3hrs NLP Group Meeting

  10. Speech Study - Experiment Design • Interface • http://www1.cs.columbia.edu/~amaxwell/survey/ NLP Group Meeting

  11. Speech Study - Experiment Design • Materials: 45 tokens of American political speech • Speakers: 9 Candidates for Democratic Party’s nomination for President • Gen. Clark, Gov. Dean, Rep. Edwards, Rep. Gephardt, Sen. Kerry, Rep. Kucinich, Sen. Lieberman, Amb. Moseley Braun, Rev. Sharpton • Topics: Postwar Iraq, Healthcare, Bush’s Tax plan, Reason for Running, Content-Neutral NLP Group Meeting

  12. Speech Study - Analysis • How much do subjects agree? • Using the weighted kappa statistic with quadratic weighting, mean kappa was 0.213 across all subject responses. • Do subjects agree differently based on the stimuli? • No, there were no systematic differences across all tokens NLP Group Meeting

  13. Most consistent statements Charisma: 0.224 (8th) Least consistent statements Do subject agree differently on the 26 statements? NLP Group Meeting

  14. What do subjects mean by “charismatic”? • Using kappa we determined which pairs of statements were most closely and consistently correlated with the charismatic statement. NLP Group Meeting

  15. Are certain speakers more charismatic than others? • Yes, there is a significant difference between speakers (p=1.75e-2) • Most charismatic • Rep. Edwards (3.73) • Rev. Sharpton (3.40) • Gov. Dean (3.32) • Least charismatic • Sen. Lieberman (2.38) • Rep. Kucinich (2.73) • Rep. Gephardt (2.77) NLP Group Meeting

  16. Does the genre or topic of speech affect judgments of charisma? • The tokens were taken from debates, interviews, stump speeches, and a campaign ad • Stump speeches were the most charismatic. (3.28) • Interviews the least. (2.90) • Topic does not affect ratings of charisma. NLP Group Meeting

  17. Does recognizing a speaker affect judgments of charisma? • Subjects were asked to identify which, if any, speakers they recognized at the end of the study. • Subjects rated recognized speakers (3.28) significantly more charismatic than those they did not (2.99). NLP Group Meeting

  18. Duration (secs) Min, max, mean, stdev F0 Raw and normalized by speaker Min, max, mean, stdev intensity Speaking rate (syls/sec) Length (words, syls) 1st, 2nd, 3rd person pronoun density Function to content word ratio Mean syllables/word Number and ratio of disfluencies and repeated words Mean words per intermediate and intonational phrase Number of intonational, intermediate, and internal phrases Mean and stdev of normalized F0 and intensity across phrases Require manual labeling of phrase boundaries What makes speech charismatic?Acoustic/Prosodic and Lexical Properties Examined NLP Group Meeting

  19. What makes speech charismatic?Properties showing positive correlation with charisma • More Content • Length in secs, words, syllables, and phrases • Higher and more dynamic raw F0 • Min, max, mean, std. dev. of F0 over male speakers • Greater intensity • Mean intensity • Higher in a speaker’s pitch range • Mean normalized F0 • Faster speaking rate • Syllables per second NLP Group Meeting

  20. What makes speech charismatic?Properties showing positive correlation with charisma • Greater variation of F0 and intensity across phrases • Std. dev. of normalized phrase F0 and intensity • The use of more first person pronouns • First person pronoun density • The use of polysyllabic words • Lexical complexity (mean syllables per word) • Speaking fluidly • Number and ratio of disfluencies negatively correlate • Repeat yourself • Number and ratio of repeated words NLP Group Meeting

  21. Overview • Background • Previous Work • Speech Study • Transcript Study • Questions Addressed • Experiment Design • Analyses of Responses • Comparisons to Speech results • Conclusion NLP Group Meeting

  22. Transcript Survey - Questions • When reading a transcript of speech, do subjects rate charisma consistently? • What do subjects mean by charisma? • Do they mean the same thing when referring to text and speech? • How does what is said influence subject ratings of charisma, compared to how it is said? NLP Group Meeting

  23. Transcript Survey - Experiment Design • Subjects: 24 paid participants found • http://newyork.craigslist.org • “Talent gigs” section • Interface: Presentation of 60 short transcripts (words…) via a web form • Dependent variables: 5-point Likert scale ratings of agreement on 26 statements. • Duration: avg. 1.5 hrs, min 45m, max ~3hrs NLP Group Meeting

  24. Transcript Survey - Design • Interface: • http://www1.cs.columbia.edu/~amaxwell/textsurvey/A/ NLP Group Meeting

  25. Transcript Study - Design • Materials: 60 of 90 tokens of American political speech • The 90 transcripts were the 45 used in the speech study, and 45 longer paragraphs • Each subject was presented with all 45 short (mean ~28 words) and a semi-random set of 15 long transcripts (mean ~130 words) • Speakers: Identical to Speech Study • Topics: Identical to Speech Study NLP Group Meeting

  26. Transcript Study - Design • Examples: • Token 1: NLP Group Meeting

  27. Transcript Study - Design • Examples: • Token 2. NLP Group Meeting

  28. Transcript Study - Design • Examples • Token 3: NLP Group Meeting

  29. Transcript Study - Design • Examples • Token 4: NLP Group Meeting

  30. Transcript Study - Design • Some tokens are rated very similarly whether presented as speech or a transcript. • Example 1 always charismatic • Example 2 always uncharismatic • Others are rated very differently • Example 3 more charismatic in speech • Example 4 in text NLP Group Meeting

  31. Transcript Study - Analyses • How much do subjects agree? • Using the weighted kappa statistic with quadratic weighting, mean kappa was 0.149 • Do subjects agree differently based on different stimuli? • No significant differences across all tokens NLP Group Meeting

  32. Most consistent statements Charisma: 0.134 (18th) Least consistent statements Do subject agree differently on the 26 statements? NLP Group Meeting

  33. What do subjects mean by “charismatic”? • Using the kappa statistic determined which pairs of statements were most closely correlated with the charismatic statement. NLP Group Meeting

  34. What do subjects mean by “charismatic”? • Those statements that cooccur with the charismatic are in the speech and transcript study overlap greatly • From this we conjecture that subjects employ a consistent functional definition of “charismatic” • Namely “charming, enthusiastic, persuasive, convincing and passionate” NLP Group Meeting

  35. Does the identity of the speaker affect judgments of charisma? • There is a significant difference between speakers (p=1.67e-10) • Most Charismatic: • Gen. Clark (3.61) • Sen. Kerry (3.56) • Gov. Dean (3.54) • Least Charismatic: • Sen. Lieberman (3.03) • Rep. Kucinich (3.12) • Amb. Mosley-Braun (3.23) NLP Group Meeting

  36. Does the genre of a transcript affect judgments of charisma? • Genre demonstrates a significant influence on charisma (p=9.18e-14) • Stump (3.34) and debate (3.32) above mean (3.15) • Interview below mean (2.85) NLP Group Meeting

  37. Does the topic of a transcript affect judgments of charisma? • Topic was significantly influenced ratings of charisma (p=1.5e-10) • In speech study, topic had no impact. • Most charismatic topics: • Content-Neutral (3.64), Reason for running (3.53) mean:3.36 • Least charismatic: • Taxes (3.12), Iraq (3.22), Healthcare (3.28) NLP Group Meeting

  38. What makes a transcript charismatic? • More Content • Length in words, or syllables • Use of more function words • Density of function words • Use of fewer first person pronouns • First person pronoun density is negatively correlated • Speak fluidly • Number and ratio of disfluencies • Repeat yourself • Number and ratio of repetitions • Lexical complexity (syls/wd) doesn’t matter NLP Group Meeting

  39. Overview • Background • Previous Work • Speech Study • Transcript Study • Conclusion • Future Work NLP Group Meeting

  40. Conclusion • Despite not agreeing about what is “charismatic”, subjects employ a common definition of “charisma”. • “Enthusiasm, passion, charm, persuasion and being convincing” are consistently used to describe someone is “charismatic”. • In general, what is said is a dominant force in whether speech is perceived as “charismatic” or not, with how it is said modifying this. • Acoustic properties broadly reflect enthusiasm and passion NLP Group Meeting

  41. Conclusion - Future Work • Resynthesis Experiments • By modifying prosody of tokens can we make Lieberman charismatic? Sharpton uncharismatic? • Investigating Palestinian Arabic • What are the similarities and differences between American and Palestinian notions of charisma? • What lexical and acoustic/prosodic properties are displayed by charismatic Palestinian speech? NLP Group Meeting

More Related