The Concept of Love From the Perspectives of Big Data and Corpus Semantics: Meaning, Reference, and Metaphor
Engin Arık, Beril Tezeller ArıkBüyük Veri ve Derlem Anlambilim Açısından Aşk Kavramı: Anlam, Gönderim ve Metafor
Engin Arık, Beril Tezeller ArıkThis study investigates the meaning of love in Turkish, using Big Data (National Institute of Standards and Technology [NIST], 2015) obtained from a corpus and social media. Love is defined as one’s commitment and/or attachment to somebody else, often involving a sexual relationship. However, debate exists over whether or not love is an emotion (Felmlee & Sprecher, 2006). Previous research has shown that love can be categorized in many ways, from romantic love to love for family members (Lee,1988; Murstein, 1977; 1988; Regan, 2011, 2016; Regan, Kocan, & Whitlock, 1998; Shaver, Schwartz, Kirson, & O’Connor, 1987; Sternberg, 2006). Nevertheless, a layperson’s definition of love can differ from the definitions used in literature/art and science. To contribute to this field of inquiry, following two theoretical approaches, the prototype theory (Goldstone, Kersten, & Carvalho, 2017; Rosch, 1978, 1999; Rosch & Mervis, 1975) and corpus semantics (Sinclair, 1991, 2004; Stubbs, 2001, 2009), we provided a definition of love in Turkish, on the basis of corpus and social media data.
The prototype theory suggests that people categorize things based on those things’ resemblance to the most typical example in a category. Corpus semantics argues that a word’s meaning is not restricted to the traditional definitions that appear in dictionaries or to its denotations and connotations as argued in traditional linguistics. In addition to these, meaning is based on the use of the word and the (linguistic) context in which the word appears, such that the frequency of the targeted word, its linguistic forms, the words that co-occur with the targeted word, phrasal structures, and genre all contribute to the targeted word’s meaning. Following these theories, we aim to find the most central definition of love in Turkish on the basis of its use in (in)formal contexts.
Method
The data consisted of formal and colloquial dictionaries published by the Turkish Language Association; 2,500 lines, 10,671 types, and 27,855 tokens from the Turkish National Corpus; and, 85,980 short messages (tweets), consisted of 130,016 types, and 1,280,736 tokens from Twitter. The tweets were obtained automatically using Rich Site Summary (RSS) and Application Program Interface (API) between December 2017 and December 2018. Each line or tweet contained the word aşk (Eng. love) at least once. The data were analyzed by the type / token frequencies, logged frequencies, Mutual Information (MI), entropy and relative entropy, lexical gravity (G), probabilistic uncertainty (Δp), and Log-Likelihood (LL or G2 ). For frequencies, MI, and LL, the software programs AntConc and Lancsbox were used.
Results
The results showed differences between the definitions of love, including idiomatic expressions in the formal dictionaries and the language used in daily life, because love was used not only in relation to intimate relationships or attachment but also in the context of pain, hate, illness, etc. The results also showed that love in Turkish was not only related to positive emotions such as compassion and fun but also negative emotions such as pain and hate. Therefore, love was associated with both positive and negative emotion concepts. A closer examination of the data indicated that the concepts were related to prototypes such as QUALITY, QUANTITY, TIME, PLACE, REALITY, MIND, BODY, LANGUAGE, BELIEF, THING, PERSON, PLANT, and VEHICLE. Love was related to close relationships including both real and imaginary attachment. The objects of attachment included not only persons but also collective things such as a soccer club. The duration of love could be very temporary or prolonged, even eternal. The meaning of love also differed depending on the context and with regard to (in)formal use of the language. For example, the Turkish National Corpus contained frequent uses of love stories, love poems, and love in families, which were significantly greater in number than the uses that appeared on Twitter. By contrast, there were frequent uses of love pains and extramarital love, which were significantly greater in number than the uses that appeared in the Turkish National Corpus.
Discussion
These findings contribute to the study of abstract concepts such as love and their meanings in daily life. This study highlights the importance of using new methods and perspectives such as Big Data and corpus semantics to examine the meaning of love. Because differences were found in the formal definitions of love and its meanings in formal use, as in the corpus and in daily use in the social media, the study calls for additional studies in this domain, especially studies of lexicography and theoretical and experimental studies of intimate relationships. Our future studies will investigate the meaning of love by considering love to be a lemma and by also focusing on its derivations and inflected forms. Researchers should take many meanings of love into account depending on the language usage when they develop tools and inventories to measure love and intimate relations.