Using Linguistic Cues for the Automatic Recognition of Personality in Conversation and Text
By François Mairesse, Marilyn A. Walker, Matthias R. Mehl, Roger K. Moore. 2007
1,079 citations recorded by the source corpus.
It is well known that utterances convey a great deal of information about the speaker in addition to their semantic content. One such type of information consists of cues to the speaker's personality traits, the most fundamental dimension of variation between humans. Recent work explores the automatic detection of other types of pragmatic variation in text and conversation, such as emotion, deception, speaker charisma, dominance, point of view, subjectivity, opinion and sentiment. Personality affects these other aspects of linguistic production, and thus personality recognition may be useful for these tasks, in addition to many other potential applications. However, to date, there is little work on the automatic recognition of personality traits. This article reports experimental results for recognition of all Big Five personality traits, in both conversation and text, utilising both self and observer ratings of personality. While other work reports classification results, we experiment with classification, regression and ranking models. For each model, we analyse the effect of different feature sets on accuracy. Results show that for some traits, any type of statistical model performs significantly better than the baseline, but ranking models perform best overall. We also present an experiment suggesting that ranking models are more accurate than multi-class classifiers for modelling personality. In addition, recognition models trained on observed personality perform better than models trained using self-reports, and the optimal feature set depends on the personality trait. A qualitative analysis of the learned models confirms previous findings linking language and personality, while revealing many new linguistic markers.
Subfields: Artificial Intelligence, Social Psychology, Natural Language Processing, Cognitive Science & Psychology
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Erik Cambria
- Soujanya Poria
- Yejin Choi
- Percy Liang
- Mohit Bansal
- Saif M. Mohammad
- Luke S. Zettlemoyer
- Björn Schuller
In Social Psychology
- Saif M. Mohammad
- Lyle Ungar
- Lucie Flekova
- Margaret L. Kern
- Erik Cambria
- Michal Kosinski
- Martin E. P. Seligman
- Soujanya Poria
In Natural Language Processing
- Yejin Choi
- Mohit Bansal
- Soujanya Poria
- Erik Cambria
- Ronan Le Bras
- Luke S. Zettlemoyer
- Saif M. Mohammad
- Myle Ott
In Cognitive Science & Psychology
- Michal Kosinski
- Jennifer Golbeck
- Karen Turner
- Cristina Robles
- Saif M. Mohammad
- Svetlana Kiritchenko
- François Mairesse
- Marilyn A. Walker
Citations in the corpus
- Neural Theory-of-Mind? On the Limits of Social Intelligence in Large LMs (2022, background): “…(Dhelim et al., 2021), knowing to call for help when an elderly person falls (Pollack, 2005), inferring personality and intentions in dialogues (Mairesse et al., 2007; Wang et al., 2019), reasoning about public commitments (Asher and Lascarides, 2013), predicting emotional and affective states…”
- Chameleons in Imagined Conversations: A New Approach to Understanding Coordination of Linguistic Style in Dialogs (2011, background): “…authorship attribution (Mosteller and Wallace, 1984; Juola, 2008), personality-type classification (Argamon et al., 2005; Oberlander and Gill, 2006; Mairesse et al., 2007), gender categorization (Koppel et al., 2002; Mukherjee and Liu, 2010; Herring and Paolillo, 2006), identification of…”
- Controlling User Perceptions of Linguistic Style: Trainable Generation of Personality Traits (2011, methodology): “…this work focuses on controlling the perceptions of the output of a dialogue system, but an important next step is to use these techniques together with personality-based user modeling methods (Mairesse et al. 2007) to simultaneously model the personality of the user and the system in dialogue.”
- Automatic personality assessment through social media language. (2015, methodology): “…rich source of relevant trait cues (Tausczik & Pennebaker, 2010); it has been used to accurately predict personality by both human (Mehl et al., 2006) and automated judges (e.g., Iacobelli et al., 2011; Mairesse, Walker, Mehl, & Moore, 2007; Schwartz et al., 2013b; Sumner et al., 2012).”
- Mining large-scale smartphone data for personality studies (2013, methodology): “matic procedure using NLP and audio features to detect the Big-Five traits from conversation extracts [19, 20].”
- Executive Extraversion: Career and Firm Outcomes (2018, methodology): “…the Q&A portion because it is less scripted than the presentation section (Hollander, Pronk, and Roelofsen 2010).11 The trained personality models (Mairesse et al. 2007) take as input the dialogue for each executive from the call and generate personality ratings based on linguistic feature counts.”
- Sentiment Analysis Is a Big Suitcase (2017, background): “13 Recently, Navonil Majumder and his colleagues presented a deep learning method to determine personality type from stream-of-consciousness essays by detecting the presence or absence of the Big Five traits in the author’s psychological profile.”
- Paralinguistics in speech and language - State-of-the-art and the challenge (2013, background): “A more recent account of the state-of-the-art, especially on the automatic recognition of personality with the help of speech and linguistic information, and experimental results can be found in Mairesse et al. (2007).”
- ASCERTAIN: Emotion and Personality Recognition Using Commercial Sensors (2018, methodology): “[35] use acoustic and lexical features, while Staiano et al.”
- Bridging the Gap between Social Animal and Unsocial Machine: A Survey of Social Signal Processing (2012, neutral): “In contrast, the works in [109], [110] predict the way prosodic features influence the perception of personality, namely, the way traits are perceived by others.”
- A Survey of Personality Computing (2014, neutral): “The features were statistics (minimum, maximum, mean and relative entropy of differences between consecutive samples) of pitch, energy, first two formants and 96 96 conversation prosody C(2) 73:0 61:3 67:7 73:9 57:0 transcripts LIWC, MRC ACC ACC ACC ACC ACC R 79:9 96:7 82:7 86:7 101:6 ER ER ER ER ER”
- You are what you tweet: Personality expression and perception on Twitter (2012, result): “Neuroticism was correlated with first-person singular pronouns, replicating a previous finding in blogs (Yarkoni, 2010), and with negation words, a relationship which has previously been documented in stream of consciousness essays (Mairesse et al., 2007).”
- Predicting personality with social media (2011, methodology): “Following the methods used in [26, 36] as well as other studies of Facebook behavior, such as [13], we utilized the Linguistic Inquiry and Word Count (LIWC) tool [35] to analyze the text.”
- Finding Deceptive Opinion Spam by Any Stretch of the Imagination (2011, methodology): “traits (Mairesse et al., 2007), to study tutoring dynamics (Cade et al.”
- Returning the N to NLP: Towards Contextually Personalized Classification Models (2020, background): “While most of the above-listed experiments have been conducted on Twitter, a variety of other datasets have been used, including phone conversations (Mairesse et al., 2007; Ivanov et al., 2011), blogs (Mukherjee and Liu, 2010; Schler et al., 2006), Facebook (Markovikj et al., 2013), or YouTube…”
- Lab And/Or Field? Measuring Personality Processes and Their Social Consequences (2015, background): “Also, personality studies so far have had a tendency to be interested in broad sets of potential behavioural manifestations (Mairesse et al., 2007; Mehl et al., 2006) rather than a narrow set of target behaviours for which computer scientists are conducting first validations studies (e.g. social…”