The Effect of Homophonic Puns on Green Purchase Intention: The Mediating Role of Affiliative Humor and the Moderation of Visual Salience

Abstract

Green brands are increasingly leveraging electronic commerce and interactive marketing to enhance persuasive efforts in new media environments. However, empirical research on the role of linguistic rhetoric in these interactive settings remains limited. This study examines whether the use of homophonic puns in green brand advertising increases green purchase intention on digital platforms, and how visual salience moderates this effect. Across four online experiments (total N = 1382), using stimuli modeled on social media feeds and mobile commerce interfaces, we manipulated advertising copy to include either homophonic puns or non-pun alternatives and varied the level of interface salience. The results indicate that homophonic puns significantly increase green purchase intention, primarily through the mechanism of affiliative humor, which proved stronger than alternative humor pathways. Moreover, visual salience moderates this effect: high-salience layouts amplify humor perception and purchase intention, while low-salience conditions yield no significant effects. This study advances the field of interactive marketing by proposing a language–emotion–behavior framework and introducing a cross-modal synergy mechanism that links linguistic rhetoric with visual attention in digital media contexts. The findings provide both theoretical insights and practical implications for e-commerce platforms, particularly in optimizing advertising copy, attention design, and message placement strategies to align pun-based messages with high-salience executions, thereby enhancing green brand engagement.

1. Introduction

With the rapid development of electronic commerce, interactive marketing has become the dominant paradigm in brand communication [1]. Green brands have leveraged the wide reach and interactivity of social media to establish new channels for consumer engagement [2,3]. This shift not only encourages more proactive adoption of sustainability but also enhances brand relatability, strengthens the product experience, and empowers consumers through greater participation and access to information [4,5,6]. Meanwhile, the communication landscape has migrated from traditional public media (e.g., television and billboards) to private digital devices such as smartphones and tablets, making e-commerce platforms key arenas for message delivery and conversion [7,8,9]. In this high-frequency, fragmented marketing environment, capturing attention and increasing ad engagement time has become a pressing challenge [10,11].

In response, many green brands have adopted homophonic puns as a core rhetorical strategy in interactive communication. A homophonic pun leverages the phonological similarity of homophones to elicit semantic association and inference, thereby producing humor and other rhetorical effects [12,13]. In green branding, such wordplay helps construct a compelling green vision, enhances brand likability, and further stimulates green purchase intention [14,15]. For instance, McDonald’s Norway used the phrase “take away your take away,” and Starbucks China used “Great job! Let the Earth have fewer carbon sighs (GOOD, GOOD, 让地球少‘碳’气)”, both effectively conveying green brand messages through wordplay. As an ancient rhetorical device, the homophonic pun has thus re-emerged as a central tool in green brand messaging. Revived within contemporary marketing, it endows copy with dual meanings via phonetic similarity despite semantic differences, providing brand messages with “dual engines” that boost both engagement and interpretive richness [16,17]. Functioning as a precisely calibrated key, this strategy ignites curiosity, infuses emotional nuance, and fosters tighter brand–consumer connections through interactive involvement; ultimately, it enables green brands to stand out and cultivate a novel model of sustainable marketing [18].

Although work on homophonic and homographic wordplay in green brand rhetoric is growing, three gaps remain. First, prior studies tend to treat humor as a unitary construct and seldom isolate the specific mechanism by which puns affect purchase-relevant outcomes; they rarely test the exclusivity of affiliative humor against non-affiliative styles [19,20,21,22]. Second, research often collapses pun subtypes, overlooking differences between homophonic and homographic puns in cognitive load and context dependence; while homophonic puns are phonetic, low-load, and well suited to attention-scarce interactive environments, evidence on whether they elicit affiliative humor that translates into green purchase intention remains limited [23,24]. Third, in digital commerce interfaces, scholarship has under-theorized boundary conditions, especially the role of visual salience that can gate when linguistic creativity helps rather than hurts; most studies consider visual and verbal cues in isolation and lack models of their synergy [25,26]. In sum, these theoretical gaps and practical misalignments highlight the need to investigate how, within interactive marketing contexts, green brands can strategically employ homophonic puns to elicit affiliative humor and optimize communication outcomes through visual salience. Addressing these limitations provides the conceptual foundation for this study.

To bridge these research voids, this study proposes and empirically validates a causal model whereby green brands’ deployment of homophonic pun rhetoric elevates consumers’ green purchase intentions through the activation of affiliative humor. Drawing on Benign Violation Theory, we explain how homophonic puns generate affiliative humor: when marketing messages create surprising yet harmless linguistic incongruities through phonetic play, consumers are more likely to experience positive humorous responses [27,28]. Additionally, we integrate visual salience as a moderating variable to examine how the prominence of visual cues shapes the efficacy of homophonic puns in eliciting affiliative humor, positing that greater visual salience amplifies the perceived humor effect [29]. Conducted within the interactive marketing context of an emerging market (China), this inquiry not only extends current literature but also offers new theoretical insights and empirical evidence for green brand communication and humor-based marketing.

This study addresses a single question: when, through what mechanism, and under what visual conditions do homophonic pun slogans used by green brands influence consumers’ green purchase intentions. We develop and test a theory driven process model that integrates Benign Violation Theory with visual attention. The model proposes that homophonic puns increase green purchase intention through affiliative humor and that visual salience moderates the link from pun to humor. We test three predictions: homophonic puns elicit affiliative humor that mediates purchase intention; affiliative humor is the unique mediator relative to non-affiliative styles; the indirect effect appears under high visual salience and not under low salience. Evidence from four experiments supports this account. Homophonic puns increase green purchase intention by eliciting affiliative humor, the mechanism is specific to affiliative humor, and the effect is stronger when visual salience is high. These results set a clear objective and a focused test of one mechanism and one boundary condition, and they offer practical guidance for combining linguistic creativity with strategic visual salience to strengthen emotional engagement, improve message clarity, and increase green purchase intentions.

This study advances digital commerce research in three ways. First, it isolates affiliative humor as the unique mediator linking homophonic-pun rhetoric by green brands to green purchase intentions, thereby moving beyond generic “humor effects” toward a precise language–emotion–behavior process model grounded in Benign Violation Theory. Second, it identifies visual salience as a theoretically bounded moderator that determines when linguistic creativity translates into purchase-relevant responses in visually crowded, feed-based interfaces. Third, it provides a reproducible experimental paradigm for creative optimization by jointly manipulating rhetorical strategy and visual design across categories, offering empirically grounded guidance for interactive marketing in digital channels.

The remainder of this article proceeds as follows: Section 2 develops hypotheses; Section 3 details the experimental design and measures; Section 4, Section 5, Section 6 and Section 7 report four experiments: Study 1 (main effect), Study 2 (mediation via affiliative humor), Study 3 (exclusivity of the mediator), and Study 4 (moderation by visual salience); Section 8 concludes with theoretical contributions, managerial implications, limitations, and directions for future research.

2. Theoretical Background and Hypothesis Development

2.1. Homophonic Pun Rhetorical Strategies and Green Purchase Intentions

In the rapidly evolving landscape of interactive marketing, green brands face the challenge of effectively communicating environmental values in high-information-density contexts where audiences have limited cognitive resources [30]. Studies have shown that emotionally engaging and entertaining linguistic styles are more effective than traditional rational appeals in capturing consumer attention and fostering identification [31]. In this contextual landscape, homophonic puns, which leverage phonetic similarity to create semantic shifts, have become a prevalent stylistic feature in green advertising and social media content owing to their playful tone, structural simplicity, and low processing cost [32]. By creating semantic tension between surface and implied meanings, homophonic puns establish a phonetic–semantic cognitive leap, allowing consumers to experience an “Aha” moment when deciphering dual meanings [33]. Such content enhances audience interest and memorability while reinforcing perceived brand creativity and approachability alongside environmental messaging [34].

Drawing on Benign Violation Theory (BVT), homophonic puns introduce a mild, easily resolvable incongruity that is appraised as harmless in green contexts, thereby increasing message engagement without eliciting negative affect [35,36]. Under attention-scarce, visually crowded interfaces, this manageable incongruity helps the message stand out, lowers inference burden, and supports favorable brand evaluations that translate into approach-oriented intentions toward environmentally friendly products. Within our integrated model, this implies a positive direct effect from homophonic-pun rhetoric to green purchase intention.

Therefore, this study focuses on homophonic pun rhetorical strategies in green branding and their influence on consumers’ green purchase intentions, proposing the following hypothesis:

H1. 

Green brands homophonic pun rhetorical strategies (vs. control group) positively affects consumers’ green purchase intentions.

2.2. Affiliative Humor Within the Framework of Benign Violation Theory

Humor functions as a widely used emotional appeal in green brand communication, enhancing message appeal, reducing resistance, and increasing perceived warmth [37]. Benign Violation Theory proposes that humor is more likely to arise when a message simultaneously involves a mild violation and a benign appraisal, and when audiences can reconcile the tension between these two elements [38]. Among humor styles, affiliative humor is characterized by lightness, inclusiveness, and the absence of hostility [39]. It aligns with the benign dimension of the theory and with the ethical tone and prosocial goals of green branding [40,41,42]. Given that green brands often communicate public morality, collective responsibility, and social participation, affiliative humor serves a distinctive regulatory function by softening moral pressure, fostering identification, and sustaining credibility in ethically sensitive topics [43,44].

As a linguistic strategy, homophonic puns naturally map onto BVT’s structure: phonetic manipulation briefly disrupts conventional linguistic logic, creating a mild semantic conflict that is resolved with an “Aha” moment once dual meanings are grasped [16,45,46]. The environmental context supports a benign reading, making the humorous experience compatible with brand ethics and tone [47]. Unlike aggressive, self-deprecating, or self-enhancing humor which can introduce threat or credibility concerns, homophonic puns are more likely to elicit affiliative humor without violating ethical boundaries [48]. In our model, affiliative humor thus functions as the psychological mechanism that transmits the persuasive impact of pun rhetoric to purchase intentions.

Therefore, this study proposes the following hypothesis:

H2. 

Green brands’ homophonic pun rhetorical strategies (vs. control group) are more likely to elicit consumers’ green purchase intentions, with affiliative humor serving as a mediating mechanism.

2.3. Visual Salience Within the Framework of Visual Attention Theory

In multimodal advertising, visual information is often processed prior to linguistic content, and its salience largely determines how attention is allocated. Visual attention theory posits that individuals selectively extract the most salient elements from complex displays through bottom-up mechanisms that prioritize brightness, color, spatial position, and motion [49]. The saliency-based model of Itti et al. integrates fundamental features and, via a center surround and winner-take-all process, generates saliency maps that identify prominent locations [50]. The Saliency Toolbox further provides an operational framework to quantify visual attention trajectories [51]. Although widely validated in cognitive science and computer vision, systematic applications in marketing communication remain nascent; many studies focus on single features and understate synergistic effects among visual properties [52]. In green advertising, visual salience not only affects attention to textual elements but also determines whether pun cues are effectively identified and decoded.

Within our integrated model, visual salience serves as a boundary condition that shapes when linguistic creativity yields humor and purchase-relevant responses. First, salient visual cues act as attention guides, increasing the probability that phonological rhetoric is noticed (e.g., focal placement, contrastive color, or iconic symbols). Second, they provide cognitive anchors that facilitate rapid comprehension, reduce inference load, and tilt appraisal toward a benign interpretation consistent with BVT. Evidence from neurocognitive research on co-activation of visual and semantic processing further supports this facilitation mechanism [53]. Consequently, when green advertising that incorporates homophonic puns is accompanied by high visual salience, the pun is more likely to be detected, appraised as harmless, and to elicit affiliative humor, thereby strengthening downstream purchase intentions. This yields the following moderation hypothesis.

H3. 

Visual salience level moderates the effect of green brands’ homophone pun rhetorical strategies (vs. control group) on affiliative humor.

To highlight the relationships discussed above, a conceptual model is presented in Figure 1.

3. Experimental Design

For testing our proposed hypotheses, we conducted four experimental studies. Study 1 employs a randomized between-subjects design using a laundry detergent advertisement to test the main effect of homophonic pun rhetoric on green purchase intention with a validated measure. Study 2 extends the paradigm with light bulb stimuli and demonstrates that affiliative humor functions as the underlying mechanism, which reveals the mediating effect of affiliative humor. Study 3 conducts an exclusivity test by evaluating other humor style accounts with established instruments, which enhances the value of the mediating effect proposed in this paper. Study 4 examines a theoretically specified boundary condition by manipulating visual salience within a 2 × 2 factorial design using bicycle advertisements, comparing high and low salience. Across studies, random assignment and manipulation checks are implemented to strengthen internal validity. The inferential sequence progresses from establishing the main effect, to identifying the mechanism, to verifying its exclusivity, and ultimately to testing the boundary condition.

4. Study 1: Testing the Direct Effect of Green Brands’ Homophonic Pun Rhetoric Strategy on Consumer Green Purchase Intentions

4.1. Experimental Procedure

The purpose of this experiment was to test Hypothesis 1 (H1), which posits that homophonic pun rhetoric by green brands is more effective in stimulating consumers’ green purchase intentions compared to the control group. Following the experimental procedure outlined by Jo [54], a between-subjects design with one factor and two levels (rhetorical strategy: homophonic pun group vs. control group) was employed, using laundry detergent, a common consumer product, as the experimental material. Data were collected via the Credamo online experimental platform, yielding 200 valid responses. The achieved sample size met a priori adequacy for a two-condition main-effect test. Data collection was conducted over three days. Of the participants, 89 were female (44.5%), and the average participant age was 25.40 years (SD = 3.73). Participants were randomly assigned to one of two groups and exposed to a banner advertisement for a fictitious brand named “Jiejin (洁净)”. The experimental group viewed an ad featuring a homophonic pun slogan, “‘Wash’ into a Better Life” (‘洗’迎美好生活”) (Figure 2), while the control group viewed with a non-pun slogan “One Pump to Clean, Fall in Love with Laundry” (一泵即洗，爱上洗衣) (Figure 2). Both slogans were adapted from a well-known brand advertisement. Participants were then required to complete a manipulation check to assess the effectiveness of the homophonic pun strategy (confirming whether the ad employed homophonic rhetoric). Afterwards, participants completed a questionnaire measuring their green purchase intentions. The green purchase intention scale was adapted from the study by Chen and Chang [55] (Cronbach’s α = 0.905, AVE = 0.763, CR = 0.906) that included three items: “I often pay attention to green products,” “I have purchased green products before,” and “I am willing to purchase this product”. Responses were measured using a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree).

4.2. Manipulation Check for Homophonic Pun Rhetoric

First, the group variable was dummy coded (homophonic pun rhetoric group = 1; control group = 0). An independent samples t-test was conducted with the manipulation check item (“This product advertisement used a homophonic pun strategy”) as the dependent variable. Results showed that participants in the homophonic pun group reported significantly higher agreement compared to those in the control group (M_{homophonic pun} = 5.92, SD = 0.78 vs. M_{control group} = 2.80, SD = 1.32, t (198) = 20.235, p < 0.001) (Figure 3). These results indicate that the manipulation of the homophonic pun rhetoric was successful.

4.3. Direct Effect of Homophonic Pun Rhetoric on Green Purchase Intentions

First, the group variable was dummy coded (group for green brands’ homophonic pun rhetoric strategies = 1; control group = 0). A one-way ANOVA was conducted with consumer green purchase intentions as the dependent variable. Results revealed a significant difference in consumer green purchase intentions based on the rhetorical strategy. Participants exposed to homophonic pun rhetoric reported significantly higher purchase intentions compared to those exposed to control group (M_{homophonic pun} = 4.22, SD = 0.92 vs. M_{control group} = 3.39, SD = 1.05; F (2.198) = 35.353, p < 0.001). These findings suggest that in H1, compared to the control group, homophonic pun rhetoric is more effective in stimulating consumers’ green purchase intentions.

5. Study 2: Testing the Mediating Role of Affiliative Humor

5.1. Experimental Procedure

The purpose of this experiment was to test H2, which proposes that homophonic pun rhetoric (vs. control group) by green brands is more likely to stimulate consumers’ green purchase intentions, with affiliative humor serving as a mediating mechanism. Following the experimental procedure outlined by Lin [56], a between-subjects design with one factor and two levels (rhetorical strategy: homophonic pun vs. control group) was employed, using light bulbs as the experimental material. Data were collected via the Credamo online platform, resulting in 400 valid responses. The achieved sample satisfied prespecified power targets for the mediation analysis, and data collection remained open for seven days. Among the participants, 195 were female (48.75%), and the average participant age was 26.75 years (SD = 4.47). Following a similar priming method as in Study 1, participants were randomly assigned to one of two groups and exposed to a banner advertisement for a fictitious light bulb brand called “Guangtou (光头).” The experimental group viewed an ad featuring a homophonic pun slogan, “Using ‘bald’ by discover that electricity bills are halved” (用上 “秃” 然发现电费减半) (Figure 4), while the control group viewed a non-pun slogan: Using suddenly discover that electricity bills are halved (用上突然发现电费减半) (Figure 4). Participants then completed a manipulation check to assess the effectiveness of the homophonic pun strategy (confirming whether the advertisement employed a homophonic pun). After the manipulation check, participants sequentially completed measures for affiliative humor responses and green purchase intentions. The affiliative humor scale was adapted from Martin’s [37] research (Cronbach’s α = 0.901, AVE = 0.647, CR = 0.902), with minor modifications based on his conceptualization of affiliative humor. The scale included five items: “I felt relaxed and happy while viewing this green advertisement,” “I could easily recognize the humorous intent of the advertisement,” “The advertisement triggered humorous associations in my mind,” “I found it difficult to perceive the humor in this advertisement” (reverse-coded), and “I found the humor in this advertisement to be creative.” Responses were measured using a 7-point Likert scale (1 = no sense of humor at all, 7 = strong perception of humor).

5.2. Manipulation Check and Direct-Effect Analysis for Homophonic Pun Rhetoric by Green Brands

Manipulation Check for Homophonic Pun Rhetoric: First, the group variable was dummy coded (homophonic pun rhetoric group = 1; control group = 0). An independent samples t-test was conducted with the manipulation check item (“This advertisement used a homophonic pun strategy”) as the dependent variable. Results showed that participants in the homophonic pun group reported significantly higher agreement compared to those in the control group (M_{homophonic pun} = 6.07, SD = 0.79 vs. M_{control group} = 2.96, SD = 1.33; t (398) = 28.397, p < 0.001). These results indicate that the manipulation of the homophonic pun rhetoric was successful.

Direct Effect of Green Brand Homophonic Pun Rhetoric on Green Purchase Intentions: First, the group variable was dummy coded (homophonic pun rhetoric group = 1; control group = 0). A one-way ANOVA was conducted with green purchase intentions as the dependent variable. Results revealed a significant difference in green purchase intentions based on the rhetorical strategy. Participants exposed to homophonic pun rhetoric reported significantly higher purchase intentions compared to those exposed to control group (M_{homophonic pun} = 4.20, SD = 1.31 vs. M_{control group} = 3.55, SD = 1.33; F (2, 398) = 24.323, p < 0.001). These findings suggest that, compared to non-pun rhetoric, homophonic pun rhetoric is more effective in stimulating consumers’ green purchase intentions.

5.3. Mediation Analysis

Study 2: Experimental data were collected via the Credamo platform using light bulbs (a common household product) as the stimulus to examine the effect of homophonic pun rhetoric by green brands on consumers’ green purchase intentions and the mediating role of affiliative humor. Results from an independent samples t-test showed that participants exposed to the homophonic pun rhetoric reported significantly stronger affiliative humor responses than those in the control group (M_{homophonic pun} = 4.045, SD = 1.269 vs. M_{control group} = 2.745, SD = 1.089; t (398) = 10.992, p < 0.001). Regression analysis further revealed that affiliative humor had a significant positive effect on green purchase intentions (β = 0.482, t = 7.210, p < 0.001).

Following the bootstrap mediation analysis procedure (Model 4), green purchase intention was set as the dependent variable, homophonic pun rhetoric (homophonic pun group = 1; control group = 0) as the independent variable, and affiliative humor as the mediator. The indirect effect through affiliative humor was 0.289, with a 95% confidence interval of (LLCI = 0.097, ULCI = 0.494), excluding zero, indicating a significant mediation effect.

Moreover, after controlling for affiliative humor, the direct effect of homophonic pun rhetoric on green purchase intentions remained significant (effect = 0.361, p = 0.017, LLCI = 0.064, ULCI = 0.658).

These results confirm the mediating role of affiliative humor and support H2; the mediation model is depicted in Figure 5.

6. Study 3: Testing the Exclusivity of Affiliative Humor as a Mediator

6.1. Purpose

This experiment aims to further verify the distinctiveness and exclusivity of affiliative humor as a mediating variable, as identified in Study 2. Although previous results demonstrated a significant mediating effect of affiliative humor in the relationship between homophonic pun rhetoric and consumers’ green purchase intentions, it remains unclear whether this mechanism is exclusive. In other words, it is unknown whether other types of humor (Aggressive Humor, Self-defeating Humor, Self-enhancing Humor) can produce similar effects. To address this issue, the study introduces experimental groups using alternative humor styles and systematically examines whether affiliative humor uniquely mediates the rhetorical to emotion to behavior pathway. This approach enhances the explanatory power and robustness of the theoretical model. Additionally, it clarifies the alignment between rhetorical strategies and emotional responses in green brand communication and provides more targeted theoretical support for selecting effective humor strategies in homophonic pun advertisements.

6.2. Methodology

This study employed a single-factor between-subjects experiment with two conditions. Participants were randomly assigned to view a green brand advertisement that either used a homophonic pun or a literal control headline. Drawing on the energy-saving light bulb materials from Study 2, we created two banner ad versions with identical product information; the headline was the only manipulated element. Affiliative, aggressive, self-defeating, and self-enhancing humor were then measured with validated short scales rather than manipulated in the stimuli. A pretest was conducted to confirm the clarity of humor style differentiation and to ensure consistency in language tone across conditions.

The experiment was distributed via the Credamo platform, and a total of 400 valid responses were collected. The achieved sample size was sufficient for the planned exclusivity test of alternative humor-style explanations. Data collection remained open for seven days. Participants were randomly assigned to one of the four groups. After viewing the advertisement, they completed questionnaires assessing humor type recognition and green purchase intention. Affiliative humor was measured using the same validated scale as in Study 2. For the other humor types, modified items based on the Humor Styles Questionnaire developed by Martin et al. [37] were used to capture aggressive, self-defeating, and self-enhancing humor styles. The green purchase intention scale remained consistent with the prior studies.

Data analysis was conducted using SPSS 26.0. Mediation effects were examined using PROCESS Model 4. The independent variable was the rhetorical strategy, coded as homophonic pun versus control. The dependent variable was green purchase intention. Each of the four humor types was entered as a parallel mediator. A bootstrapping procedure with 5000 samples and a 95 percent confidence interval was employed to test the significance of each mediation path. This approach allowed the comparison of indirect effects across humor types and provided evidence on whether affiliative humor serves as a unique and exclusive mediator in the pathway from rhetorical strategy to green purchase intention.

6.3. Integrated Results

Mediation analysis was conducted using PROCESS Macro (Model 4) in SPSS, with rhetorical strategy (homophonic pun vs. control group) as the independent variable and green purchase intention as the dependent variable. Affiliative humor and three types of non-affiliative humor (Aggressive, Self-defeating, and Self-enhancing) were tested as parallel mediators.

The analysis revealed that affiliative humor significantly mediated the effect of homophonic pun rhetoric on green purchase intention in the green brand condition. The indirect effect was 0.289, with a 95% confidence interval of (LLCI = 0.097, ULCI = 0.494). The indirect effect of self-enhancing humor was not significant, with an effect size of −0.0012 and a 95% confidence interval of (LLCI = −0.034, ULCI = 0.041), which included zero. Similarly, the indirect effect of self-defeating humor was not significant. The effect size was −0.0049 with a confidence interval of (LLCI = −0.040, ULCI = 0.033), also including zero. The indirect effect of aggressive humor was also non-significant, with a value of 0.0036 and a confidence interval of (LLCI = −0.037, ULCI = 0.035), also including zero. The mediating effect of affiliative humor was significantly greater than that of non-affiliative humor, indicating that affiliative humor plays a distinct and exclusive mediating role in the proposed mechanism. Control variables such as age and gender did not exert significant confounding effects, further validating the central role of affiliative humor as a mediator.

Taken together, only the affiliative humor pathway reached statistical significance and its mediating impact was stronger than those of the non-affiliative styles, indicating a distinct and exclusive mediating role that aligns with Benign Violation Theory: favorable attitudes and intentions are more likely when the linguistic deviation is appraised as harmless. In contrast, aggressive, self-defeating, and self-enhancing humor, although potentially entertaining, may evoke discomfort or credibility concerns and are less effective in fostering positive cognitive reappraisal or emotional transfer, which can attenuate a brand’s green image. These findings reinforce the causal progression from perceived semantic conflict to affiliative humor and then to purchase intention, offer clear guidance for selecting rhetorical and humor strategies in interactive marketing by green brands, and provide an empirical template for future examinations of mediator exclusivity.

7. Study 4: Testing the Moderating Effect of Visual Salience

7.1. Selection and Validation of Visual Salience Materials

To evaluate the visual salience characteristics of webpage banner advertisements, this study adopted the STB (Saliency Toolbox) visual attention computational model developed by Walther [51], and constructed an algorithmic simulation system using the MATLAB R2021a platform. This model utilizes a center-surround contrast mechanism, setting the radius of the focus circle to 1/16 of the image width. It calculates multidimensional feature differences, including chromaticity, luminance, and orientation gradients, between target pixel groups and their neighboring regions, thereby extracting potential areas of high visual contrast (Figure 6). Through iterative computations until all salient regions are identified, the model ultimately generates a visual salience map, in which regions are ranked in descending order of salience intensity (Figure 7). Experimental validation confirmed that the top five regions ranked by salience in the scan map demonstrated high recognition validity. Based on this validation, the present study confined the measurement of visual salience to the top five visual elements identified in the salience map.

For each poster image, if the target banner advertisement is detected among the top five candidate regions in the scan sequence, its ranking position is standardized and converted into a visual salience score within the (0,1) range. Specifically, the salience score is calculated by dividing its ranking order by the total number of scanned objects. The calculation formula is as follows:

VS = 1 − (p − 1)/x

Visual salience (VS) quantifies the perceptual prominence of the headline region in each banner. For each creative, we computed $\mathrm{V}\mathrm{S}=1-(p-1)/x$, where $p$ is the rank of the headline region among the top $x$ peaks on the saliency map; we set $x=5$. If the headline region was not detected among these candidates, VS was set to 0. When a headline contained multiple glyphs, the ad level VS equaled the maximum of the glyph level scores. Following the thresholding approach proposed by Wilson [52], we operationalized visual salience as a binary construct. Consistent with this approach, we applied fixed thresholds rather than a mean split: VS ≥ 0.50 was coded as high salience and VS < 0.50 as low salience; values between 0.40 and 0.60 were not used for the manipulation.

7.2. Experimental Design and Procedure

This experiment used bicycles as the stimulus material and adopted a 2 (rhetorical strategy: homophonic pun vs. control) × 2 (visual salience: high vs. low) between-subjects design. A total of 382 valid responses were collected. The achieved sample met a priori power for the 2 × 2 interaction test. Data collection remained open for seven days. Among the participants, 157 were female (40.99%), and the average age was 26.53 years (SD = 4.33). Following the priming method used in Study 1 and Study 2, participants were randomly assigned to one of four groups. They were exposed to a banner advertisement for a fictitious bicycle brand called “Feichi.” The experimental group viewed an ad featuring a homophonic pun slogan, “Refuse to be a ‘carbon’ pioneer, and ride into endless joy” (拒做”碳”路先锋，畅享”骑”乐无穷) (Figure 8), while the control group saw a non-pun slogan: Be a trailblazer and enjoy endless fun (勇做探路先锋，畅享其乐无穷) (Figure 9). Participants then completed a manipulation check to assess whether the advertisement employed a homophonic pun strategy. Afterwards, participants completed measures for humor responses and green purchase intentions. In Figure 8, in the homophonic pun advertisement, the glyph “碳” with an exhaust pipe icon ranked third among the top five peaks, yielding VS = 0.60, and the glyph “骑” with a bicycle icon ranked second, yielding VS = 0.80. The ad level VS therefore equaled 0.80 and the Visual Salience was classified as high. In the control advertisement, the glyph “驰” with a wheel icon ranked second, yielding VS = 0.80 and a high classification. In a low-salience ad, the headline region ranked fifth, yielding VS = 0.20 and a low classification.

7.3. Manipulation Check and Direct Effect Analysis

Manipulation Check for Homophonic Pun Rhetoric: First, the group variable was dummy coded (homophonic pun rhetoric group = 1; control group = 0). An independent samples t-test was conducted with the manipulation check item (“This advertisement used a homophonic pun strategy”) as the dependent variable. Results showed that participants in the homophonic pun group reported significantly higher agreement than those in the non-pun group (M_{homophonic pun} = 5.46, SD = 1.14 vs. M_{control group}= 3.03, SD = 1.22; t (380) = 20.183, p < 0.001). These results indicate that the manipulation of the homophonic pun rhetoric was successful.

Direct Effect of Homophonic Pun Rhetoric on Green Purchase Intention: First, the group variable was dummy coded (homophonic pun rhetoric group = 1; control group = 0). A one-way ANOVA was conducted with green purchase intentions as the dependent variable.

Results revealed a significant difference between the groups: participants exposed to homophonic pun rhetoric reported significantly higher green purchase intentions than those exposed to non-pun rhetoric (M_{homophonic pun} = 4.51, SD = 1.05 vs. M_{control group} = 3.82, SD = 0.92; F (2, 380) =46.014, p < 0.001). These findings suggest that, compared to the control group, homophonic pun rhetoric is more effective in stimulating consumers’ green purchase intentions.

7.4. Experimental Results

We tested the moderating role of visual salience in the relationship between homophonic pun rhetoric by green brands and consumers’ affiliative humor responses and estimated a moderated mediation model using the PROCESS macro (Model 7) with 5000 bootstrap samples and a 95% confidence interval. An analysis was conducted with green purchase intention as the dependent variable, homophonic pun rhetoric as the independent variable (homophonic pun group = 1; control group = 0), affiliative humor as the mediator, and visual salience as the moderator (high visual salience = 1; low visual salience = 0). The mediating effect of affiliative humor on green purchase intentions varied across levels of visual salience. Specifically, under conditions of high visual salience, the mediating effect of affiliative humor was significant, with an indirect effect size of 0.490 and a confidence interval of (LLCI = 0.134, ULCI = 0.626), whereas under low salience, it was not significant. These findings indicate that higher visual salience strengthens humor elicitation and produces a conditional indirect effect on green purchase intentions, thereby supporting H4. The model and conditional indirect effects are shown in Figure 10.

8. Discussion

This study tested a theory-driven process model for how homophonic pun rhetoric by green brands shapes consumer response and identified both a mechanism and a boundary condition within one framework. Across four experiments, a clear pattern emerged. Homophonic puns increased green purchase intention, and the effect operated through affiliative humor. Visual salience functioned as a boundary condition that links language to attention. When salience was high, the pun cue was more likely to be noticed and quickly interpreted, a benign appraisal followed, affiliative humor was elicited, and downstream effects strengthened. When salience was low, affiliative humor was not reliably elicited and the indirect effect did not appear. Two results refine common assumptions. Non-affiliative humor styles did not transmit the effect, which challenges undifferentiated views of humor and supports the core claim of Benign Violation Theory that a mild and harmless violation is most naturally expressed as affiliative humor in sustainable branding. Study two also showed partial mediation, suggesting that novelty or processing fluency may contribute alongside affiliative humor. Taken together, these regularities and exceptions confirm the positive main effect, specify a language to emotion to behavior pathway centered on affiliative humor, and show that the pathway depends on visual conditions that bring the cue to the foreground. The pattern integrates Benign Violation Theory with visual attention, explains variation in prior findings, and clarifies when language-based persuasion emerges in multimodal advertising (

Table 1. Summary of Key Findings and Aligned Theoretical and Managerial Implications.

Study	Key Finding	Theoretical Implication	Managerial Implication
Study 1	Homophonic pun slogans increased green purchase intention relative to the control condition.	Confirms that linguistically engaging rhetoric can move intentions in interactive contexts and motivates a language–emotion–behavior	Use concise homophonic puns to seed favorable intent when attention is scarce, especially in feed environments.
Study 2	The effect operates through affiliative humor.	Supports Benign Violation Theory: a mild, harmless incongruity elicits affiliative humor that links language to intention.	Homophonic pun copies that signal warmth and inclusivity to elicit affiliative humor, not merely wit.
Study 3	Affiliative humor is the exclusive mediator among tested styles.	Qualifies “humor effects” by isolating the benign appraisal path as affiliative humor in sustainable branding.	Avoid non-affiliative humor styles for green messages since they do not transmit purchase relevant responses.
Study 4	Visual salience conditions the pathway; effects strengthen when the pun cue is visually salient.	Integrates visual attention with the language pathway; salience acts as a gating factor for decoding and appraisal.	Pair homophonic puns with high-salience design of the headline region to improve noticeability and downstream impact.

).

8.1. Theoretical Implications

This study advances the theory on green brand interactive marketing by linking the four studies into an integrated storyline and using this storyline to present our theoretical contributions. In line with recent calls in interactive marketing to demonstrate contributions through storytelling [57], we connect linguistic design, emotional experience, and behavioral response to explain how homophonic puns operate in green brand communication.

First, this research supports and extends rhetorical theory in green branding by articulating a micro-level pathway from language to emotion to behavior. Prior work that links rhetorical style to persuasion [33,58] has largely remained at the macro-level and has treated rhetorical figures as global message features. Our evidence goes beyond these accounts and shows that, in interactive green brand communication, homophonic pun rhetoric exerts a positive main effect on green purchase intention. Study 1 documents this main effect of homophonic pun rhetoric on green purchase intention. Subsequent studies refine this picture by showing that the effect is carried by specific emotional processes rather than operating as a purely cognitive shortcut. The focus on homophonic rather than homographic wordplay clarifies that forms with lower processing demands and weaker context dependence are especially suitable for interactive environments in which consumers scroll, click, and respond rapidly. This extension provides a refined theoretical basis for the design of rhetorical strategies in green branding.

Second, this research clarifies and qualifies Benign Violation Theory and related work on humor in marketing, and it connects these ideas to emerging interactive marketing research on the boundaries of brand playfulness. Introducing Benign Violation Theory into brand rhetoric research [38], Study 2 demonstrates that the effect of homophonic puns on purchase intention operates through affiliative humor. This finding supports the core proposition of the theory that mild and harmless violations generate positive affect and approach tendencies. Study 3 further qualifies broad claims about humor in marketing. Prior research often treats humor as a unitary driver [37,59]. Our results contradict that undifferentiated assumption by showing that non affiliative styles do not carry the effect in the green branding context. In doing so, we extend the application scope of Benign Violation Theory and specify affiliative humor as the key process variable that links linguistic deviation to consumer response in sustainable communication. At the same time, our findings resonate with interactive marketing research showing that creative brand tactics can generate either resonance or resistance in customer engagement and that dark or aggressive humor, such as roasting, is appreciated only within certain boundaries and among specific segments [60,61]. By isolating affiliative humor as the effective style for homophonic puns in green interactive campaigns, our study adds a linguistic and emotionally differentiated layer to this stream of work.

Third, this research extends visual attention theory to a boundary condition for language-based persuasion in multimodal interactive settings. Integrating visual attention theory [50] with language processing, Study 4 shows that visual salience conditions the success of homophonic puns in green brand messages. High salience improves access to the pun cue and accelerates inference, which increases the likelihood of a benign appraisal and the experience of affiliative humor. This moderated pathway extends visual attention theory into multimodal persuasion by identifying salience as a gating mechanism that shapes decoding efficiency and humor appraisal, and it responds to calls to consider linguistic and visual elements jointly rather than in isolation [52]. The null pattern under low salience qualifies simple main effect expectations by showing that linguistic creativity alone is insufficient when attention is not effectively guided. The findings also introduce a quantitative paradigm based on salience modeling that links visual stimuli to language-based cognition in consumer research.

Taken together, the four studies support a unified conceptual model in which homophonic pun rhetoric influences green purchase intention through affiliative humor and in which visual salience operates as a boundary condition. The results both support key propositions in existing theories and qualify or extend them by identifying the unique role of affiliative humor and by specifying when the mechanism is likely to emerge in interactive green brand communication.

8.2. Managerial Implications

The findings of this study offer important managerial implications for green brand management in interactive new media contexts. Most notably, this study proposes a framework for multimodal synergistic communication, combining homophonic pun rhetoric with visual salience. In green brand marketing, companies should prioritize the use of homophonic pun rhetoric, transforming environmental messages into affiliative humor expressions through the creative use of phonological similarity, thereby reducing consumers’ psychological distance toward green information. At the same time, firms should enhance the synergistic effect between visual salience and pun rhetoric by employing high-contrast colors, dynamic symbols, and metaphorical icons (e.g., green leaves, images of the Earth) to increase the perceptual prominence of the pun elements. By constructing an integrated decoding environment that combines “semantic conflict in language” with “reinforcement through visual cues,” brands can accelerate consumers’ comprehension of environmental intentions via cross-modal information synergy, thereby stimulating affiliative humor responses and enhancing green purchase intentions.

Our study further suggests that green brands should dynamically adjust their communication strategies based on application contexts and audience characteristics. Specifically, brands should flexibly adjust the level of visual salience and the complexity of pun rhetoric according to different media characteristics and consumers’ cognitive contexts: In high-visual-load environments (e.g., outdoor advertisements, short videos), brands should employ strong contrast visual elements and straightforward homophonic puns to quickly capture consumer attention. In low-visual-load environments (e.g., text-based social media), brands should enhance the recognizability of puns through localized symbolic markers such as quotation marks or color highlights, thereby minimizing decoding difficulties. Meanwhile, communication strategies should be tailored based on the target audience’s level of environmental knowledge: For audiences with strong environmental awareness, brands can design homophonic puns that combine depth and creativity, supplemented by abstract visual metaphors. For the mass market, preference should be given to catchy and easily understood homophonic puns paired with intuitive visual symbols, aiming to achieve a balance between low processing effort and high cognitive impact.

Finally, green brands can employ empirical research to quantitatively assess the synergistic effects of “pun rhetoric strategies and visual salience”, analyzing consumer attention duration, semantic decoding efficiency, and emotional response intensity under different advertising conditions, thereby enabling targeted optimization of ad design. The core objective is to transform brand environmental appeals into perceptible humorous experiences for consumers through cross-modal synergy, thereby stimulating green purchase intentions. Companies must recognize that the key to green marketing lies not merely in information transmission but in the precise coordination of language and visual elements, embedding environmental values into consumer cognition through a “harmless conflict—benign humor” approach. This ultimately facilitates a sustainable marketing loop of “emotional arousal—value identification—behavioral transformation”, helping brands build a differentiated competitive advantage in the green consumer market.

8.3. Limitations and Future Research Directions

Although this study reveals the underlying mechanisms by which homophonic pun rhetoric combined with visual salience influences green brand communication, several limitations remain. Issues of external validity and generalizability should be acknowledged, and these limitations suggest several directions for future research [62].

At the theoretical and linguistic level, the study focuses solely on homophonic puns and does not examine how other linguistic strategies, such as semantic puns, metaphors or irony, may interact with visual elements in green brand communication. A central boundary condition concerns linguistic dependency. Our stimuli draw on Chinese homophony within a character-based writing system, where dense phonological overlap and morpheme-to-character mapping facilitate pun formation and may shape the path from phonetic ambiguity to benign appraisal. Moreover, the measurement of humor responses is limited to affiliative humor and does not consider the potential influence of other humor types. This restriction may narrow the scope of the proposed mechanism and leaves open the question of how different humor styles shift the balance between benign and non-benign appraisals in green branding.

At the methodological level, all experimental studies are scenario-based and conducted in a Chinese context. Participants responded to hypothetical green brand campaigns with controlled stimuli and evaluated their intentions in scripted scenarios rather than in live interactive environments. This design strengthens internal validity but constrains external and ecological validity, because it may not fully capture spontaneous reactions, social interactions or behavioral choices that occur on real digital platforms. The sample is largely drawn from a single cultural context with specific norms regarding green consumption, brand humor and homophonic wordplay, so the generalizability of the findings across cultural settings remains to be validated. In addition, the operationalization of visual salience primarily relies on basic visual features such as color and motion, with limited attention to more complex visual elements such as typography, spatial layout, interface clutter and platform-specific design cues that are typical of real interactive feeds. The research program also relies exclusively on laboratory-style experiments with controlled stimuli and online panels. We did not conduct large-scale natural or field experiments embedded in live digital platforms, which limits ecological validity and may constrain inference about performance under marketplace dynamics such as algorithmic ranking, social interactions and competing content. Finally, the present studies measure key constructs using self-report questionnaires only. The absence of behavioral and physiological indicators raises the risk of common method variance and may understate or overstate effect sizes.

Future research could be expanded in several directions. First, scholars can incorporate cross-linguistic designs that explicitly compare character-based and alphabetic languages by creating meaning-equivalent homophonic slogans, minimal pair manipulations or orthographic versus phonological cues, and thereby test whether the affiliative humor pathway holds outside the Chinese context. Second, future studies can include a wider range of linguistic rhetorical strategies and humor types to build a more comprehensive theoretical framework for green communication and to identify when different forms of wordplay and humor are beneficial or detrimental. Third, researchers can refine the dimensional categorization of visual salience and use technologies such as eye tracking and fMRI to investigate the underlying cognitive and neural mechanisms that link visual guidance, pun resolution and humor appraisal. Fourth, cross-cultural and cross-linguistic comparative studies can examine how language structure and cultural norms jointly condition the effectiveness of visual–linguistic synergy in interactive green branding. Fifth, longitudinal research designs can be employed to track the long-term impact of humor responses on green consumer behavior, thereby providing more dynamic theoretical support for sustainable brand communication. Combining these extensions with richer visual manipulations, behavioral and physiological indicators and field-based evidence will offer a more precise and externally valid test of the proposed model.