📄 social_science

The Persistent Self: Social Presence Alters Memory Without Disrupting Self-Organization

This study demonstrates that while first-person encoding yields higher recognition accuracy than third-person encoding, the stability of recollective quality and subjective self-presence across perspectives suggests that the first-person view primarily serves as an uncontested organizer of the memory episode rather than a disruptor of self-organization.

Original authors: Zofia Farkowska, Jie Sui

Published 2026-06-28

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Zofia Farkowska, Jie Sui

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Question: Who is Taking the Photo?

Imagine you are walking down a street and you see a beautiful cake on a stall. You want to remember it later. The researchers wanted to know: Does it matter how you "see" yourself in that memory?

Usually, we remember things from our own eyes (First-Person). But sometimes, we remember things as if we are watching a movie of ourselves (Third-Person). Previous studies suggested that watching yourself from the outside makes your memory weaker and less vivid.

The big mystery this paper tried to solve was: Why does the memory get weaker?

Theory A: Maybe seeing yourself from the outside messes up the "filing system" in your brain when you first learn the event (Encoding).
Theory B: Maybe the filing system is fine, but when you try to pull the memory out later, you just can't "feel" like yourself anymore (Retrieval).

The Experiment: A Virtual City Walk

To test this, the researchers built a virtual city using a VR headset (like a video game you wear on your head). They had 78 people walk down a street and look at items like fruit, clocks, and towels.

They split the people into three groups, changing only the "camera angle" of their experience:

First-Person: You see the street through your own eyes. No body is visible. (Like playing a game where you are the character).
Third-Person: You see a robot avatar walking in front of you, and you follow it. (Like watching a movie of yourself walking).
Second-Person: You see a robot avatar walking right next to you, like a friend. (Like walking side-by-side with a companion).

After the walk, they tested the participants' memory. They asked two things:

Did you recognize this item? (Accuracy)
Do you "Remember" it with a specific feeling of reliving the moment, or do you just "Know" it was there? (This is a special test to measure how vivid the memory feels).

The Results: The "Filing" vs. The "Feeling"

The results were surprising and revealed a split between how well we remember and how we feel about the memory.

1. The Accuracy Test (The "Photo Quality")

First-Person: People remembered the most items.
Third-Person: People remembered the fewest items.
Second-Person: People were in the middle.
The Takeaway: Seeing yourself from the outside (Third-Person) definitely made the memory "blurrier" or less accurate. It's like taking a photo from a distance; you miss some details.

2. The "Reliving" Test (The "Feeling")

Here is the twist: Even though the Third-Person group had worse accuracy, they still felt just as much like themselves when they remembered the items.
The "Remember" feeling (the sense of "I was there") was exactly the same across all three groups.
The Takeaway: Even if the memory is less accurate, the feeling that "this happened to me" didn't change. The "self" wasn't lost; the memory just had fewer details.

3. The "Who is Walking With Me?" Test

The researchers asked people how it felt to be in the virtual world.
In the Third-Person group, people felt a strange mix: they felt like themselves, but they also felt like a "stranger" was walking right next to them. The "stranger" presence was so strong it felt equal to their own presence.
In the Second-Person group (walking next to the robot), people felt more like they were with a "friend."
The Takeaway: The Third-Person view introduced a "stranger" into the mental picture, which confused the brain's filing system (making accuracy drop), but it didn't stop the person from feeling like the main character of the story.

The Conclusion: The Librarian vs. The Reader

The paper concludes that the "Self" plays two different roles, and these roles can be separated:

The Librarian (Encoding): When you first experience an event, your "Self" acts like a librarian organizing books. If you view yourself from the outside (Third-Person), the librarian gets distracted by a "stranger" standing nearby. The books get filed less neatly, so you can't find them as easily later (lower accuracy).
The Reader (Retrieval): When you go back to read the memory later, you are still the same person reading the book. Even if the librarian filed it poorly, the feeling of reading the story remains yours. You still feel like the protagonist.

In short: Changing your visual perspective changes how clearly you remember the details, but it doesn't change the fact that the memory belongs to you. The "Self" that organizes the memory is different from the "Self" that experiences it later.

Technical Summary: The Persistent Self: Social Presence Alters Memory Without Disrupting Self-Organization

Problem and Motivation
Episodic memory is defined by autonoetic consciousness—the subjective sense of personally re-experiencing a past event. While the self is central to both the encoding and retrieval of these memories, it remains unclear whether the self functions as a unified agent across both stages or if its roles are separable. Previous research indicates that encoding from an observer (third-person) perspective typically yields less accurate memories than field (first-person) encoding. The prevailing interpretation suggests this accuracy cost stems from weakened self-referential processing during encoding, which subsequently impairs the richness of the recollective experience (autonoetic consciousness) at retrieval.

However, this causal link has not been empirically separated. It is unknown whether reduced accuracy in observer memories reflects a genuine degradation of the recollective quality or merely a reduction in mnemonic strength while the subjective sense of "reliving" remains intact. Furthermore, inconsistencies exist in the Head-Mounted Display (HMD) Virtual Reality (VR) literature regarding perspective effects; some studies find no accuracy cost when controlling for field of view, while others find reliable deficits. Additionally, the mechanism behind observer-perspective costs is debated: does the visible avatar in third-person VR impair memory by introducing a competing social representation (a "stranger" or "other") that disrupts self-organization?

Methodology
The study employed a between-subjects design with 78 healthy young adults (ages 18–48) randomly assigned to one of three encoding perspective conditions within a VR environment:

First-Person: Camera at eye height with no body avatar rendered.
Third-Person: A robot-like avatar placed ahead of the camera, facing forward, moving synchronously with the participant.
Second-Person (Novel): The same robot-like avatar placed laterally at the participant's right shoulder (approx. 0.5m), simulating walking alongside a companion.

Participants navigated a virtual city street, viewing 32 items (organic and non-organic) on market stalls. The environment was controlled for field of view, item exposure, luminance, and locomotion to address prior inconsistencies in VR literature.

Following the encoding phase, participants completed:

Presence Ratings: Retrospective self-reports on the degree to which they felt a "self," "friend," or "stranger" was present during the experience.
Recognition Memory Task: A 2D desktop test involving 64 items (32 old, 32 new). Participants indicated "Old/New" and, for "Old" items, provided a "Remember" (recollective) or "Know" (familiarity) judgment.

Data were analyzed using Signal Detection Theory to calculate recognition sensitivity ( $d'$ ) and separate sensitivities for Remember and Know responses. Generalized Linear Mixed Models (GLMM) were used to test perspective effects, controlling for age, presence scores, item visibility, and response time.

Key Results

Recognition Sensitivity: First-person encoding produced significantly higher recognition sensitivity ( $d' = 1.99$ ) compared to third-person encoding ( $d' = 1.49$ ). The second-person condition ( $d' = 1.84$ ) did not differ significantly from either the first or third-person conditions. This confirms a reliable accuracy cost for third-person encoding even when visual field and other confounds are controlled.
Recollective Quality (Remember Sensitivity): Contrary to the hypothesis that accuracy and recollective quality are linked, Remember sensitivity did not vary significantly across perspectives. Notably, the second-person condition showed numerically higher Remember sensitivity than the first-person condition, a pattern inconsistent with a simple directional reduction in recollective quality based on perspective.
Felt Presence:
- Self-Presence: Ratings of felt self-presence were stable and equivalent across all three conditions.
- Stranger-Presence: Both avatar conditions (second and third-person) significantly elevated felt stranger-presence compared to the first-person condition. Crucially, in the third-person condition, felt stranger-presence reached statistical parity with felt self-presence.
- Friend-Presence: The second-person (lateral) perspective selectively elevated felt friend-presence compared to the first-person condition, distinguishing it from the stranger-like quality of the third-person view.

Key Contributions

Dissociation of Encoding and Retrieval Roles: The study demonstrates that visual perspective affects the self as an organizer of the memory episode (encoding) more than the self as the experiencer of the memory (retrieval). While third-person encoding reduced memory accuracy, it did not diminish the subjective quality of autonoetic consciousness (Remember sensitivity).
Resolution of VR Inconsistencies: By rigorously controlling field of view and item visibility, the study establishes a reliable accuracy cost for third-person encoding in VR, resolving discrepancies in prior literature where such costs were sometimes absent.
Social Mechanism of Perspective: The introduction of the second-person perspective revealed that the accuracy cost in third-person encoding is likely driven by the introduction of a competing social representation (stranger-self parity) rather than the mere absence of a first-person view. The lateral positioning (second-person) altered the social context (increasing friend-presence) without restoring the accuracy deficit seen in the third-person condition, suggesting specific social dynamics drive the encoding disruption.

Significance and Claims
The authors claim that these findings support a model where the self serves as an uncontested organizer of an episode during encoding in first-person views, whereas third-person views introduce a competing social agent that disrupts item-context binding. However, this disruption affects the strength of the memory trace (accuracy) without altering the phenomenological nature of the memory once retrieved (recollective quality).

The paper concludes that the self's role in organizing a memory at formation is distinct from its role in re-experiencing that memory at retrieval. The study modestly notes that while the data support this dissociation, the causal link between the "stranger-self parity" and memory accuracy requires formal mediation analysis in future work. Furthermore, the authors acknowledge that the lack of a body avatar in the first-person condition prevents a definitive distinction between agentic self-salience and the absence of a competing "other," and that spontaneous perspective shifts at retrieval could influence the null results for recollection. The primary significance lies in the empirical separation of accuracy and recollective quality, suggesting that observer-perspective memories are less accurate but not necessarily less "one's own" in terms of subjective reliving.

The Big Question: Who is Taking the Photo?

The Experiment: A Virtual City Walk

The Results: The "Filing" vs. The "Feeling"

The Conclusion: The Librarian vs. The Reader

Technical Summary: The Persistent Self: Social Presence Alters Memory Without Disrupting Self-Organization

More like this