A Social-Interactionist Approach to Emergent Phenomena

Cognition in a Social Context: A Social-Interactionist Approach to Emergent Phenomena

Introduction to the Social-Interactionist Approach

• Seventy Years of Cognitive Investigations: For $70$ years, research has focused on how the cognitive system transforms sensory input into behavioral output, understanding perception, memory, reasoning, and behavior. This focus has traditionally been on information processing within the individual.

• Beyond Individual Behavior: The effects of cognition extend beyond individual behavior, influencing and being influenced by social contexts. There is a growing body of literature on the interaction between cognitive and social processes.

• Emergent Phenomena: Recent research highlights large-scale phenomena emerging from individual interactions, such as:

  • Shared attention (Risko, Richardson, & Kingstone, $2016$; Shteynberg, $2010$).

  • Collective memory (Hirst, Yamashiro, & Coman, $2018$).

  • Collective emotions (Mackie, Smith, & Ray, $2008$; Yzerbyt, Kuppens, & Mathieu, $2016$).

  • Collective action (Sebanz, Bekkering, & Knoblich, $2006$).

• Proposed Framework: This paper advocates for an empirical framework to explore how the cognitive system's capacity constraints, biases, and operations have emergent properties at a social level. The goal is to investigate how interactions among individuals lead to collective-level outcomes.

• Addressing Limitations of Previous Approaches: Previous attempts to link microlevel psychological processes to macrolevel social outcomes (e.g., symbolic interactionism, methodological individualism, social constructivism) have been criticized for:

  • (a) Overreliance on qualitative methods.

  • (b) Lack of theoretical development.

  • (c) Low predictive power (Van Lange, $2006$).

• Goal: To overcome these limitations by focusing on the bidirectional influences between cognitive mechanisms and social contexts.

Structure of the Proposed Social-Interactionist Framework

This framework aims to bridge cognitive and social psychology to explain how microlevel cognitive phenomena lead to large-scale social outcomes. It involves three phases:

1. Phase 1: Establishing Boundary Conditions: Identifying the relevant cognitive phenomena and understanding their inherent limitations and characteristics at the individual level.

2. Phase 2: Investigating Social Context Influence: Exploring how individual cognition is influenced by the social context in which it manifests.

3. Phase 3: Studying Propagation in Social Networks: Examining how dyadic-level influences propagate throughout social networks, leading to community-level effects.

• Benefits of this Approach:

  • (a) Illuminates large-scale consequences of well-established cognitive phenomena.

  • (b) Fosters interdisciplinary dialogues between psychology and other social sciences.

  • (c) Offers greater relevance for public policy compared to existing approaches.

Emergent Cognition: Collective Memory as a Case Study

The proposed approach is illustrated using Retrieval-Induced Forgetting (RIF) to demonstrate its emergent properties at a collective level, specifically in the formation of collective memories.

Defining Collective Memory

• Shared Individual Memories: Collective memories are defined as shared individual memories that bear on people's identities. They characterize groups ranging from small families to large nations (Hirst et al., $2018$).

• Mechanism of Emergence: While collective memory has received attention in psychology (Roediger & Abel, $2015$; Roediger & DeSoto, $2014$), a mechanistic framework for its emergence has been lacking. This approach proposes that conversations are the primary engine.

• Conversational Remembering: When people discuss an event, they reciprocally influence each other's memories. These interactions, when occurring across a community, lead to convergence toward a similar memory of the past.

• Rigorous Understanding Requires Three Steps:

  1. Establishing how retrieval alters the memory of the encoded event.

  2. Exploring the attenuation and facilitation of these mnemonic changes during conversational remembering.

  3. Conducting social network research to understand how mnemonic influences spread across the community.

Phase 1: Finding the Relevant Cognitive Phenomena (Individual Level)

• Conversations as the Engine: The study of collective memory formation begins with identifying conversations as the main mechanism for synchronizing individual memories. Conversational remembering involves selective retrieval.

• Retrieval-Induced Forgetting (RIF): Drawing on cognitive psychology literature, selective retrieval influences previously encoded memories in two ways (Anderson, Bjork, & Bjork, $1994$; Fig. $1$ in original article):

  • (a) Strengthening: Increasing the accessibility of the retrieved information.

  • (b) Forgetting (RIF Effect): Reducing the accessibility of information related to the retrieved memories.

• Example (September 11th Attacks):

  • If one remembers waking up at $7:30 ext{ a.m.}$ and learning about attacks at $9:25 ext{ a.m.}$ on September $11$. Rehearsing the $7:30 ext{ a.m.}$ detail would:

    • Strengthen the memory of waking up at $7:30 ext{ a.m.}$.

    • Induce RIF for the $9:25 ext{ a.m.}$ detail (reducing its accessibility), relative to an unretrieved category like one's location during the attacks.

• Empirical Support: Coman, Manier, and Hirst ($2009$) found support for both strengthening and RIF in September $11 ext{th}$ memories among NYC residents.

• Boundary Conditions: An extensive literature on RIF's boundary conditions (Murayama et al., $2014$) helps predict contexts that facilitate RIF at a social level.

Phase 2: Attenuation and Facilitation of Individual Cognition in Social Settings (Dyadic Level)

• Collaborative Remembering: Retrieval often occurs collaboratively in social contexts, not in isolation (Rajaram & Maswood, $2017$).

• Socially Shared RIF (SSRIF): Joint conversational remembering can trigger strengthening and RIF effects in both speakers and listeners.

  • Listeners concurrently and covertly retrieving information with a speaker also experience strengthening of discussed memories and SSRIF of unretrieved but related information (Cuc, Koppel, & Hirst, $2007$).

• Example (September 11th Attacks continued):

  • Listening to a speaker recall waking up at $7:30 ext{ a.m.}$ on September $11$ reduces the listener's memory accessibility for when they learned about the attacks. This occurs because the listener's concurrent retrieval of the $7:30 ext{ a.m.}$ detail suppresses related memories.

• Mnemonic Similarity: If strengthening and RIF occur simultaneously in both speaker and listener, their memories become more similar after conversational remembering, as found by Coman and Hirst ($2012$).

• Social Modulation: Listeners' concurrent retrieval can be socially modulated:

  • Facilitated: When listeners are motivated to relate to the speaker (Coman & Hirst, $2015$).

  • Attenuated: When listeners perceive speakers as dissimilar (Barber & Mather, $2012$).

• These factors influence individual cognition and affect outcomes in larger communities.

Phase 3: Cognition in Social Networks (Network Level)

• Propagation of Mnemonic Influence: To achieve mnemonic convergence across a community, the influence of a speaker on a listener's memories must propagate beyond the initial dyadic interaction (Drost-Lopez & Coman, $2018$).

  • Example: If the influence between two New Yorkers regarding September $11 ext{th}$ memories spreads through subsequent conversations, local synchronization leads to community-level mnemonic convergence. This happens because individuals strengthen and forget similar memories post-conversation.

• Complexity of Real-World Networks: Real-world communities are not simple chains but involve time-dependent, often redundant social influences. Extrapolating from dyads to networks can be complex due to nonlinear dynamics.

  • Nonlinear Dynamics: More dyadic synchronization might paradoxically lead to less collective convergence (e.g., in politically polarized communities where subgroups synchronize internally but diverge collectively; Hastorf & Cantril, $1954$).

• Novel Approaches to Address Complexity:

  • a) Lab-Created Social Networks: Experimental studies with sequential dyadic interactions in social networks show mnemonic convergence depends on strengthening and SSRIF (Coman, Momennejad, Drach, & Geana, $2016$; Fig. $1$ in original article).

    • Outcome: Discussed items become widely shared, while undiscussed but related items are forgotten more extensively than undiscussed unrelated items (Fig. $2 ext{a}$ in original article).

    • Impact: Social influences on individual memory directly affect the community's collective memory content.

  • b) Agent-Based Simulations: These simulations involve artificial agents with psychological characteristics (e.g., identity, memory) in communities (Epstein, $2006$). They assess whether microlevel cognitive mechanisms cause large-scale social phenomena and the conditions for attenuation or facilitation.

    • Findings: Luhmann and Rajaram ($2015$) and Coman, Kolling, Lewis, and Hirst ($2012$) showed that the magnitude of strengthening and SSRIF affects community convergence. Network connectivity also matters: smaller, densely connected communities (e.g., $30$ agents) converge faster than larger, equally dense communities (e.g., $100$ agents).

• Model for Psychological Research: This social-interactionist perspective on collective memory formation provides a methodological and theoretical refinement that can model investigations of other emergent phenomena (Fig. $2 ext{b}$ in original article).

Bridging Microlevel Cognitive Phenomena and Macrolevel Social Outcomes

The proposed framework offers a mechanistic exploration of cognition's emergent properties at a large social scale, applicable beyond memory effects.

1. Generalizability of the Social-Interactionist Approach

• Applicable to Any Cognitive Mechanism: This approach applies to any cognitive mechanism whose effect is manifested in a social context.

• Cognitive-Capacity Constraints: These influence collective-level phenomena.

  • Attention Example: If individuals can only attend to a limited number of objects, these attentional constraints limit encoded and communicated information, thereby limiting collective reconstruction of visual displays (Miller, $2003$). This relates to collective attention (Fig. $3$ in original article).

  • Emotional Processing Bias: Preferential processing of emotionally arousing/negative information (Rozin & Royzman, $2001$) could lead to more similar information encoding across a community, accelerating convergence in communicative exchanges (Fig. $2 ext{b}$ for collective emotions).

  • Risk Preferences: The cognitive