This paper describes how journalists, in the Cook Islands, use sentiment lexicon when reporting online news. To do so, we employ Sentiment Analysis (SA) in combination with sociolinguistic variationist theory and logistic regression analysis. SA relies on the Word-Emotion Association Lexicon source (Mohammad & Turney 2013), which comprises 10,170 lexical items. The bulk of research carried out on sentiment analysis only distinguishes between positive vs. negative emotions. By contrast, we provide a fine-grained coding by exploring how eight specific core emotions (i.e. ANGER, ANTICIPATION, FEAR, DISGUST, JOY, SADNESS, SURPRISE, and TRUST) are socially stratified in formal contexts. We built a small-scale corpus from web-based newspapers to find out (i) whether social factors (age and sex) condition the use of sentiment lexicon and (ii) to evaluate the socially acknowledged generalisations according to which females tend to use sentiment lexicon more than males. The data was quantitatively examined through mixed-effects Rbrul logistic regression analysis. The independent variables include: word class (i.e. nous, adjectives, verbs), sex, age, and word-frequency. Specifically, the latter is a variable involved in language processing and is commonly studied in psycholinguistics, sociolinguistics, and corpus linguistics (Mickiewicz 2019). To account for word-frequency we use the SUBTLEX-US corpus (Brysbaert & New 2009). Our findings suggest that sentiment lexicon is conditioned by age, with young and old speakers favouring the use of sentiment lexicon. Sex, word class, and word-frequency do not have a significant influence on sentiment lexicon in our data.
SOCIOLINGUISTIC VARIATIONIST ANALYSIS OF WORD-EMOTION LEXICON IN COOK ISLANDS ENGLISH ONLINE NEWS
Siria Guzzo
;Carmen Ciancia
2023
Abstract
This paper describes how journalists, in the Cook Islands, use sentiment lexicon when reporting online news. To do so, we employ Sentiment Analysis (SA) in combination with sociolinguistic variationist theory and logistic regression analysis. SA relies on the Word-Emotion Association Lexicon source (Mohammad & Turney 2013), which comprises 10,170 lexical items. The bulk of research carried out on sentiment analysis only distinguishes between positive vs. negative emotions. By contrast, we provide a fine-grained coding by exploring how eight specific core emotions (i.e. ANGER, ANTICIPATION, FEAR, DISGUST, JOY, SADNESS, SURPRISE, and TRUST) are socially stratified in formal contexts. We built a small-scale corpus from web-based newspapers to find out (i) whether social factors (age and sex) condition the use of sentiment lexicon and (ii) to evaluate the socially acknowledged generalisations according to which females tend to use sentiment lexicon more than males. The data was quantitatively examined through mixed-effects Rbrul logistic regression analysis. The independent variables include: word class (i.e. nous, adjectives, verbs), sex, age, and word-frequency. Specifically, the latter is a variable involved in language processing and is commonly studied in psycholinguistics, sociolinguistics, and corpus linguistics (Mickiewicz 2019). To account for word-frequency we use the SUBTLEX-US corpus (Brysbaert & New 2009). Our findings suggest that sentiment lexicon is conditioned by age, with young and old speakers favouring the use of sentiment lexicon. Sex, word class, and word-frequency do not have a significant influence on sentiment lexicon in our data.File | Dimensione | Formato | |
---|---|---|---|
Lingue e Linguaggi_Ciancia Guzzo_Giu23_Word-emoton lexicon in Cook Islands English online news.pdf
accesso aperto
Descrizione: This paper describes how journalists, in the Cook Islands, use sentiment lexicon when reporting online news. To do so, we employ Sentiment Analysis (SA) in combination with sociolinguistic variationist theory and logistic regression analysis. SA relies on the Word-Emotion Association Lexicon source (Mohammad & Turney 2013), which comprises 10,170 lexical items. The bulk of research carried out on sentiment analysis only distinguishes between positive vs. negative emotions. By contrast, we provide a fine-grained coding by exploring how eight specific core emotions (i.e. ANGER, ANTICIPATION, FEAR, DISGUST, JOY, SADNESS, SURPRISE, and TRUST) are socially stratified in formal contexts. We built a small-scale corpus from web-based newspapers to find out (i) whether social factors (age and sex) condition the use of sentiment lexicon and (ii) to evaluate the socially acknowledged generalisations according to which females tend to use sentiment lexicon more than males. The data was quantitatively examined through mixed-effects Rbrul logistic regression analysis. The independent variables include: word class (i.e. nous, adjectives, verbs), sex, age, and word-frequency. Specifically, the latter is a variable involved in language processing and is commonly studied in psycholinguistics, sociolinguistics, and corpus linguistics (Mickiewicz 2019). To account for word-frequency we use the SUBTLEX-US corpus (Brysbaert & New 2009). Our findings suggest that sentiment lexicon is conditioned by age, with young and old speakers favouring the use of sentiment lexicon. Sex, word class, and word-frequency do not have a significant influence on sentiment lexicon in our data.
Tipologia:
Versione editoriale (versione pubblicata con il layout dell'editore)
Licenza:
NON PUBBLICO - Accesso privato/ristretto
Dimensione
533.26 kB
Formato
Adobe PDF
|
533.26 kB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.