Conversational-Linguistic Politeness for Chatbot Design: an Evaluation of Parameters and Card-Based Approaches
Open Access
- 1 January 2026
- journal article
- Published by Sociedade Brasileira de Computação in Journal on Interactive Systems
- Vol. 17 (1) , 30-47
Abstract
Chatbots often find it challenging to meet user expectations by providing appropriate responses and engaging in natural dialogues. Previous research suggests that incorporating linguistic and conversational strategies, particularly those centered on politeness, can significantly enhance user satisfaction and interaction quality. However, current conversation design practices often lack clear guidance to support these strategies, leaving designers to rely on personal preferences. To enhance the design of polite chatbots, previous work has investigated the use of politeness parameters, which help in understanding how politeness is expressed in chatbot dialogues through language function variety and indirect communication. In this work, our initial focus is on assessing participants' perceptions when exposed to different uses of these parameters. To achieve this, three fictional scenarios were presented to six participants to assess the combined impact of various politeness parameters. Although participants acknowledged different polite design strategies, discussing the parameters during the interactions proved challenging. To investigate the use of these parameters during the design process, we instantiated them in a card-based tool with four categories. A workshop was conducted for evaluation where seven participants first designed polite dialogues without the cards. Then, the participants were presented with the cards to reflect on whether they could have supported the design process. Some benefits mentioned by participants included supporting tailored dialogues, guiding design decisions, and maintaining a consistent tone. In summary, our contributions involve evaluating the effectiveness of politeness parameters as design tools, introducing a set of four cards to assist in formulating politeness strategies (currently used in educational settings), and exploring user perceptions when presented with these cards.Keywords
This publication has 29 references indexed in Scilit:
- A survey on near-human conversational agentsJournal of King Saud University Computer and Information Sciences, 2021
- Future directions for chatbot research: an interdisciplinary research agendaComputing, 2021
- Non-native speaker perception of Intelligent Virtual Agents in two languages: the impact of amount and type of grammatical mistakesJournal on Multimodal User Interfaces, 2021
- Co-designing Strategies to Provide Telecare Through an Intelligent Assistant for Caregivers of Elderly Individuals2020
- Card-based design tools: a review and analysis of 155 card decks for designers and designingDesign Studies, 2019
- Conversation Analysis2016
- Pedagogical Agents and Affect2016
- Cultural Issues in HCI: Challenges and Opportunities2015
- A politeness effect in learning with web-based intelligent tutorsInternational Journal of Human-Computer Studies, 2011
- Expression and Meaning1979