TY - GEN

T1 - Lexical Entrainment and Intra-Speaker Variability in Cooperative Dialogues

AU - Menshikova, Alla

AU - Kocharov, Daniil

AU - Kachkovskaia , Tatiana

PY - 2021

Y1 - 2021

N2 - In dialogues, intra-speaker variability is often explained by the relationship between interlocutors. A person may speak differently with a friend and a stranger or depending on the interlocutor’s gender or age—in all these cases we expect speech entrainment, but the degree of entrainment may vary. In this research, we measured lexical entrainment in a series of dialogues, where each one of 20 “core” speakers talked to five different interlocutors: a sibling, a close friend, an unfamiliar person of the same gender and similar age, an unfamiliar person of the other gender and similar age, and an unfamiliar person of the same gender, greater age and higher job position. We hypothesized that the degree of speech entrainment systematically varies according to the type of interlocutor, across all the “core” speakers. The following measures of entrainment were used: parts of speech statistics, verb forms statistics, language style matching, and lexical density. Our data have shown that a person speaks very similarly to his/her sibling; dialogues with a friend or a samegender stranger of similar age show fewer similarities; the least “common language” is observed in dialogues with a stranger of the opposite gender and with a stranger of greater age and higher job position.

AB - In dialogues, intra-speaker variability is often explained by the relationship between interlocutors. A person may speak differently with a friend and a stranger or depending on the interlocutor’s gender or age—in all these cases we expect speech entrainment, but the degree of entrainment may vary. In this research, we measured lexical entrainment in a series of dialogues, where each one of 20 “core” speakers talked to five different interlocutors: a sibling, a close friend, an unfamiliar person of the same gender and similar age, an unfamiliar person of the other gender and similar age, and an unfamiliar person of the same gender, greater age and higher job position. We hypothesized that the degree of speech entrainment systematically varies according to the type of interlocutor, across all the “core” speakers. The following measures of entrainment were used: parts of speech statistics, verb forms statistics, language style matching, and lexical density. Our data have shown that a person speaks very similarly to his/her sibling; dialogues with a friend or a samegender stranger of similar age show fewer similarities; the least “common language” is observed in dialogues with a stranger of the opposite gender and with a stranger of greater age and higher job position.

KW - speech entrainment

KW - intra-speaker variability

KW - lexical entrainment

KW - lexical density

KW - language style matching

UR - https://www.interspeech2021.org/proceedings

M3 - Conference contribution

SP - 1957

EP - 1961

BT - Proc. Interspeech 2021

PB - International Speech Communication Association

ER -