Psychological Representations and Inference Patterns of Facial Social Exclusion

Abstract

Social exclusion is a pervasive phenomenon in human social interaction, and the ability to accurately perceive and represent social exclusion from facial information is crucial for individual social adaptation. This paper explores the psychological representations and inference patterns associated with facial social exclusion. By integrating findings from social psychology and cognitive neuroscience, we analyze how individuals utilize facial cues to construct mental models of exclusion. We further examine the underlying cognitive mechanisms and the specific inference patterns—such as the over-generalization effect and the role of contextual cues—that govern how people judge whether an individual is being excluded or is an excluder. The findings suggest that psychological representations of social exclusion are not merely passive reflections of physical features but are active constructions influenced by the perceiver's internal states and social expectations.

1. Introduction

Social exclusion refers to the phenomenon where individuals are prevented from participating in social activities or are ignored and rejected by others. As social animals, humans have evolved a highly sensitive "social monitoring system" to detect potential threats of exclusion. Among various social signals, the face serves as a primary source of information. Understanding the psychological representation of facial social exclusion—how we mentally "visualize" an excluded versus an included person—is essential for understanding social cognition.

Based on Trait Attribution Theory, this study employed the reverse-correlation image classification technique to investigate the content of mental representations and the patterns of trait inference associated with faces that trigger social exclusion. Using Chinese university students as participants, the research induced a social exclusion context followed by a two-image forced-choice task.

The results indicate that the mental representations of excluded individuals contain specific trait information regarding trustworthiness and dominance. Specifically, the nose and mouth were identified as core diagnostic regions for trustworthiness, while the eyebrows served as the core diagnostic region for dominance. Both objective measurements and subjective evaluations consistently demonstrated that low trustworthiness and low dominance are typical trait characteristics of social exclusion, playing critical roles in the process of trait inference.

Furthermore, the study found that low trustworthiness carries a greater weight in this process, providing strong support for the "trustworthiness priority hypothesis." These findings reveal the functional role of facial mental representations in social exclusion and extend the trustworthiness priority mechanism from verbal materials to facial stimuli. This research provides a new theoretical basis for understanding the mechanisms of exclusion at the level of high-level social cognition.

2. The Psychological Representation of Social Exclusion

Psychological representation refers to the internal mental imagery or cognitive structures that individuals hold regarding specific social categories. In the context of social exclusion, research often employs the reverse correlation image reconstruction technique to visualize these internal templates.

2.1 Visualizing the "Excluded Face"

Studies have shown that the psychological representation of an excluded individual often involves specific morphological features. For instance, faces represented as "excluded" are frequently perceived as having lower levels of warmth, competence, and trust. These mental representations are not static; they are modulated by the perceiver's own experiences. Individuals with a high need to belong or those who have recently experienced exclusion themselves tend to form more vivid and sensitive representations of exclusion-related facial cues.

2.2 The Role of Emotional Expression

The psychological representation of social exclusion is inextricably linked to emotional perception. Excluded individuals are often represented as exhibiting "social hunger" or "social pain." This is reflected in mental templates that emphasize sadness, suppressed anger, or a lack of positive affect. These representations serve as diagnostic tools, allowing observers to quickly categorize the social status of others in a complex environment.

3. The Impact of Social Exclusion on Face Perception

The human face serves as a critical medium for conveying social information. For individuals experiencing social exclusion, the accurate identification and interpretation of facial expressions and characteristics are essential for re-establishing social connections. Research indicates that social exclusion significantly modulates the early stages of facial processing. Specifically, excluded individuals often exhibit heightened sensitivity to social cues, demonstrating a "social monitoring" mechanism. This mechanism directs attention toward faces that signal potential inclusion or acceptance, while simultaneously increasing vigilance toward threatening or further rejecting social signals.

3.1 Mental Representations of Social Targets

Mental representation refers to the internal cognitive structures that individuals use to encode, store, and interpret external information. In the context of social exclusion, the mental representations of others' faces are not objective reflections of physical reality but are instead reconstructed based on the observer's internal state and social needs.

[FIGURE:1]

Studies employing reverse correlation techniques have revealed that social exclusion can distort the mental templates individuals hold for specific social groups or individuals. For instance, when individuals feel excluded, their mental representations of "ingroup" members may become more idealized or positive as a compensatory mechanism to seek belonging. Conversely, the mental representations of those who perpetrated the exclusion may be biased toward more negative, untrustworthy, or hostile facial features.

3.2 The Role of the Need to Belong

The fundamental driver behind these representational shifts is the need to belong. According to the Social Monitoring System (SMS) framework, social exclusion activates a specialized cognitive system designed to track and process social information that could facilitate re-affiliation. This system influences how faces are mentally represented by prioritizing features that are relevant to social reintegration.

[TABLE:1]

As shown in [TABLE:1], the intensity of the exclusion experience correlates with the degree of representational bias. Individuals with a high chronic need to belong tend to form more distinct mental representations of smiling or "accepting" faces compared to those who are less sensitive to social rejection.

4. Theoretical Background and Trait Attribution

As social beings, humans rely on the establishment and maintenance of social relationships to ensure survival and reproductive success. Social exclusion is a pervasive phenomenon that encompasses a dual experience, involving not only the feelings of the excluded party but also the psychological state of the excluder. For the actor, such behavior often serves specific adaptive purposes, strengthening intimacy and cohesion within their own group \cite{Wyer & Schenke, 2016}.

According to Trait Attribution Theory, perceivers infer the personality characteristics of strangers based on two primary trait dimensions: trustworthiness and dominance \cite{Oosterhof & Todorov, 2008; Todorov et al., 2015}. These dimensions correspond to the warmth (trustworthiness) and competence (dominance) dimensions of the Stereotype Content Model (SCM; \cite{Fiske2007, Sutherland2016}).

4.1 Modeling Social Exclusion Probability

In the field of social perception, machine learning and deep learning models have increasingly been utilized to decode how humans process facial information. According to the valence-dominance model, individuals quickly evaluate faces along two primary dimensions: trustworthiness and dominance. The probability of exclusion $P_e$ can be expressed as a function of perceived facial traits:

$$P_e = f(\mathcal{T}, \mathcal{D}, \mathcal{A})$$

where $\mathcal{T}$ represents trustworthiness, $\mathcal{D}$ represents dominance, and $\mathcal{A}$ represents attractiveness. Previous research \cite{Rudert2017} suggests that the partial derivative $\frac{\partial P_e}{\partial \mathcal{T}} < 0$, indicating that as perceived trustworthiness increases, the likelihood of social exclusion decreases. Furthermore, the interaction between facial traits and situational pressure can be represented by the following relation:

$$\Omega = \int_{0}^{t} (\phi \cdot \mathcal{S} + \epsilon) dt$$

In this equation, $\Omega$ represents the cumulative social pressure, $\phi$ denotes the facial trait vector, $\mathcal{S}$ represents the situational impact factor, and $\epsilon$ accounts for stochastic noise in social perception.

5. Study 1: Visualization of Mental Representations

This study recruited 53 Chinese participants (mean age $19.60 \pm 0.83$ years) to generate visualized images of the mental representations of social exclusion using reverse correlation techniques based on random noise.

5.1 Methodology

A single-factor between-subjects experimental design was employed. The independent variable was the situation type (social exclusion vs. social acceptance). The base face was an average male face with a neutral expression. Random noise consisted of truncated sinusoidal fragments across six orientations and five spatial scales, forming 4,092 random contrast parameters. Participants performed a two-image forced-choice (2IFC) task over 300 trials.

5.2 Results

The Intraclass Correlation Coefficient (ICC) for pixel-wise intra-group consistency was analyzed. Results showed that the ICC among individuals in the social exclusion group was 0.972 (95% CI [0.972, 0.973]), and for the social acceptance group, it was 0.938 (95% CI [0.935, 0.942]), indicating excellent consistency. A significant negative correlation was found between the noise patterns of social exclusion and social acceptance mental representations ($r = -0.334, p < 0.001$).

[FIGURE:2]

Pixel test analysis revealed that the diagnostic information for social exclusion encompasses the forehead, eyes, nose, mouth, and peripheral contour regions. Brighter pixels in the mouth area were associated with lower perceived trustworthiness, while the eyebrow region served as a critical diagnostic feature for dominance.

6. Study 2: Trait Inference Patterns

This study investigated how trustworthiness and dominance predict the mental representation of social exclusion through objective measurement and subjective assessment.

6.1 Objective Measurement (Study 2a)

Using pixel-based regression, we predicted the mental representation of social exclusion based on fundamental facial trait dimensions. Low trustworthiness significantly and positively predicted the mental representation of social exclusion ($\beta = 0.78, 95\% \text{ CI} = [0.77, 0.78]$). Both high and low dominance positively predicted social exclusion ($\beta = 0.19$ and $\beta = 0.15$, respectively), but with smaller effect sizes than trustworthiness.

6.2 Subjective Assessment (Study 2b)

A total of 105 college students evaluated the generated mental representation images. Linear regression revealed that for the mental representation of social exclusion, low trustworthiness exerted the maximal positive predictive effect ($\beta = 0.21, p < 0.001$). Low dominance also showed a significant positive predictive effect ($\beta = 0.16, p < 0.001$). For social acceptance, only high trustworthiness exhibited a significant positive predictive effect ($\beta = 0.44, p < 0.001$).

7. Discussion and Conclusion

The results consistently indicate that the mental representations of faces associated with social exclusion contain integrated trait information regarding trustworthiness and dominance. Specifically, the nose and mouth serve as core diagnostic regions for trustworthiness, while the eyebrows are the primary diagnostic regions for dominance.

The findings provide empirical support for the "trustworthiness priority hypothesis." While both traits are encoded, trustworthiness carries significantly greater weight in the inference process. Mental representations characterized by "low trustworthiness and low dominance" (cold and incompetent) or "low trustworthiness and high dominance" (cold and competent) are typical features of social exclusion. This research extends the understanding of social monitoring systems and provides a visual basis for the psychological mechanisms underlying social rejection.