Metaphors in Business Applications: Modelling Subjectivity Through Emotions for Metaphor Comprehension

Metaphors in Business Applications: Modelling Subjectivity Through Emotions for Metaphor Comprehension

Sunny Rai, Shampa Chakraverty, Devendra Kumar Tayal
Copyright: © 2021 |Pages: 20
DOI: 10.4018/978-1-7998-4240-8.ch006
OnDemand:
(Individual Chapters)
Available
$37.50
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

Commercial advertisements, social campaigns, and ubiquitous online reviews are a few non-literary domains where creative text is profusely embedded to capture a viewer's imagination. Recent AI business applications such as chatbots and interactive digital campaigns emphasise the need to process creative text for a seamless and fulfilling user experience. Figurative text in human communication conveys implicit perceptions and unspoken emotions. Metaphor is one such figure of speech that maps a latent idea in a target domain to an evocative concept from a source domain. This chapter explores the problem of computational metaphor interpretation through the glass of subjectivity. The world wide web is mined to learn about the source domain concept. Ekman emotion categories and pretrained word embeddings are used to model the subjectivity. The performance evaluation is performed to determine the reader's preference for emotive vs non emotive meanings. This chapter establishes the role of subjectivity and user inclination towards the meaning that fits in their existing cognitive schema.
Chapter Preview
Top

Introduction

The growing significance of creative text in today’s digital world is indisputable. From interactive digital advertisements to ubiquitous product reviews, it has become imperative for business organizations to automatically process as well as generate novel creative constructs to engage users actively. A text is said to be creative or figurative if its intended meaning is different from its literal meaning. Metaphor, sarcasm, and irony are few such figurative speeches, used frequently in one’s daily communications on social media.

Today’s digital market, including Zomato, AirBnB and Amazon, has user’s ratings and product reviews at its core for product promotion. In fact, the popularity of a business model is often gauged by its mentions on social media platforms. A significant chunk of research work strives to accurately mine user’s opinion and brand perception in the market (Asur and Huberman, 2010; Zadeh and Sharda, 2014; Jurafsky et al., 2014). One of the major hurdles in processing online reviews is creative texts. Consider the following product reviews on Amazon to understand this point:

  • (1)

    “At first she was just there with no support. She sat in that hole with no protection. Shifting, jiggling not knowing what can happen. There has to be something I can do. I looked and search and what I found is the “Fabric Case” for Alexa. This is it! A support she will be happy to wear. No more accidental popping out, no more shifting and readjusting, she is finally safe.”

  • (2)

    “I must add that this jacket’s repellent powers are infinitely multiplied when coupled with the included Medal of Yavin. Without it, the untrained female eye may confuse this ceremonial jacket with a Justin Timberlake style biker coat.”

As in utterance (1), metaphorical references such as ‘A support she will be happy to wear’ and ‘she is finally safe’ convey strong opinion in reviews or feedback. Similarly, in utterance (2), the buyer is not pleased about his purchase, a visually-unappealing jacket and thus sarcastically expresses his displeasure towards the product. Traditional opinion mining techniques struggle to accurately capture the intended intent or sentiment for figurative texts. Nevertheless, these creative constructs convey strong reactions whether positive or negative in comparison with a literal feedback and thus, act as a significant factor in determining the true feelings of the user.

Figure 1.

Metaphors in Advertisements: (a) Nokia (Source: Nokia Connecting People Campaign1) (b) Audi Wake up Campaign (Source: Audi Wake up Campaign2)

978-1-7998-4240-8.ch006.f01

Once widely popular digital campaign Connecting People of NOKIA, with its vivid imagery forges mental associations leaving an unforgettable imprint on the viewer's mind. In Fig. 1-(a), the phrase connecting people has an emotional connotation which subtly associates the human’s basic instinct to reach out and build bonds with NOKIA. Various studies demonstrate that the human mind enjoys a bit of an enigma and often, metaphorical advertisements enforce viewer’s attention along with providing a sense of pleasure (Danesi, 2002). Another such campaign ‘Wake up!’ by Audi is provided in Fig. 1-(b). At first glance, one merely sees white lane markings on a tar road. However, the quest to make sense of it, takes us to the fine print on this ad. This provides the context of Audi automobiles and fatigue. Here, this campaign makes a clever yet intriguing use of lane markings to convey shut eyes. Through this visual metaphor, Audi automobiles promote the feature fatigue detector in their cars while simultaneously, providing an ‘Aha!’ moment for its audience on deciphering this puzzling ad.

Figure 2.

Google Translate: Translating creative text literally

978-1-7998-4240-8.ch006.f02

Key Terms in this Chapter

Cosine Similarity: It is defined as cosine of the angle between two non-zero word vectors, used to measure word or text similarity.

Figurative Text: A text is said to be figurative text if the phrase is not intended to be understood literally.

Nominal Metaphor: An explicit mapping between a concept from Target Domain and a concept from Source Domain where a copular verb is used as an adhesive.

Emotion: An immediate yet brief reaction or perception of a feeling towards an event such as meeting your long-lost friend or delicious meal.

Perception: A psychological process to interpret sensory stimuli such as visual, auditory, or gustatory on the basis of one’s past experiences.

Localization: The transformation of a product such as a mobile app or website to incorporate the linguistic and cultural aspects of a local market.

Word Embedding: A learned numerical representation for a word where words having similar meaning have a similar representation.

Complete Chapter List

Search this Book:
Reset