Spatial Memories

Ege Selin Islekel (Texas A&M University)

On February 7, 2023, Southern regions of Turkey and the Northern border of Syria were hit by two back-to-back earthquakes. According to the reports, 680,000 homes collapsed or were majorly damaged in Turkey, together with another 106,000 homes in Syria. Official numbers suggest that around 50,000 people died in Turkey; and unofficial reports point to more than double that number, estimating around 130,000, including those who were not found and unrecorded refugees in the region. As entire cities that were destroyed left their place to mountains of rubble, in the days following the earthquake, rebuilding efforts started even before rescue efforts were concluded, plans being released all the while those under and above the rubble were still screaming:

Recently, on the third anniversary of the earthquake on February 7, 2026, survivors gathered in the streets, repeating this phrase: even though the rubble is gone, and now most of what is surrounding are new buildings or empty lots, the question still echoes through these streets. Can memories be erased when the surroundings are destroyed, demolished, or even entirely rebuilt from the ground up?

Memory and Space

Memories are indeed spatial. They not only exist ‘inside’ our heads, but also outside. Sense data contributes not only to the formation of memory, but also, its affective character. This character of memory is familiar from Proust’s famous example of eating a madeleine. Here, Proust’s narrator realizes that memory has an intricate relation to sensible data: memories are formed by data, for sure, so one remembers moments together with colors, smells and the like. But also, sensations can bring forth memories. Thus, for the narrator, the smell of a madeleine can bring joy as the joy he felt when his grandmother would make him madeleines, even in most unexpected circumstances.

Much of this character of memory is due to the distinction between voluntary memory and involuntary memory. According to Henry Bergson, voluntary memory, where what we now call episodic memory largely resides, operates through deliberate recall. It is mobilized by the needs of the present and retrieves the past in response to a practical demand: we try to remember, and memory answers that call. For example, when you try to recall the layout of a childhood home in order to describe it to someone, you mentally reconstruct the hallway, the kitchen, etc. Voluntary memories work as snapshots, drawing on images of the past in the context of the present. The connection between past and present may not be faithful, and nevertheless, we remember as if the connection is faithful.

Thus, in the case of a space of destruction for example, the remaining residents of earthquake zones, in each recall, would remember as if they are there once again, as if they have access to the full set of memories. The replacement of the external sources of sense data in places of destruction would not replace the voluntary memory or the memory images at all: in each recall, the residents thus would still remember the streets of their childhood, their kitchen tables, their living rooms, as if they are there once again, however different the place currently looks.

Involuntary memory, on the other hand, is not only triggered by sense data, but it is connected to sensations and affect more generally. Involuntary memories do not have programmable correlations to emotions: what we remember and what we feel does not correlate linearly to a particular sense experience.

Imagine walking down the street where you grew up. You might intentionally recall a specific moment, say, playing soccer with a friend, and feel the joy or fun attached to that memory. But at the same time, other sensations may surface without being actively summoned: perhaps a tightening in the chest, an inexplicable anxiety, a trace of fear that you can’t quite pinpoint why. These are not separate, fully articulated recollections. They are affective residues of other experiences, perhaps moments of vulnerability, survival, or loss, that were not consciously recalled but nevertheless become present. In this case, replacement would especially not be possible, because there is no erasure that is involved in memory in any case: there is a multiplication of memories in involuntary memory.

Traumatic Memories

One might say, however, that forgetting would precisely be at stake in the case of memories such as the survivors of earthquake may have: these are after all not ordinary memories, but rather, traumatic memories, often described as functioning according to different rules. In the case of traumatic experiences, such as earthquakes and destruction, details may be blurry, fragmentary, or entirely inaccessible. A survivor returning to the site of destruction might even feel surprised at encountering the rebuilt place, as though the event was strangely unmoored from its physical location.

Psychoanalysts, such as Cathy Caruth, talk about traumatic experiences as ‘non-locatable’ experiences: experiences that do not become fixed to a linear succession of events in memory, and would have a different place than what happened right before or right after. So, a person who survived a violent event, would not necessarily situate the event in regular temporal succession. If this is so, one might think rebuilding completes the forgetting. When a collapsed neighborhood is replaced by glass towers and cafés, when rubble gives way to polished sidewalks, perhaps the sensory world itself overwrites what was there. If the memory is already fragmentary, perhaps the new environment finishes the work of erasure.

However, this would be to move too quickly: even though traumatic memories might be incomplete, this does not mean that they are erased, or they can completely be missing from the field of memory. Instead, survivors of trauma report feeling the sensations of the event, even though the details of the event may be incomplete in their details. Because involuntary memory works beyond programmable associations, the survivors feel the sensations even in the absence of complete memories.

Indeed, forgetting the details of the event can increase the sensations: in other words, forgetting can preserve the event and its effects, rather than erasing them or making their replacement possible.

The site may look unfamiliar, but the sensation of unease persists. The event survives not as a clear image but as atmosphere. One may not remember the details or the totality of the event that has taken place, and nevertheless, the affective responses to traumatic event would fester even when the surroundings have been altered.

Memory Rubble

How to think, then, the relation between space and memory, if memory can persist even spatial surroundings have been altered? One way to think about this would be to insist in spatial terms: memories sediment, they become like the rubble that remains in the aftermath of the disaster. They are formed in relation to environmental cues. However, instead of becoming altered, they become layered on top of each other.

Walter Benjamin once wrote that one must approach memory “like a person digging:” but memory is not what is hidden under the rubble, but rather, the rubble itself, all its layers, gaps and fillings. Voluntary memories are what we pull out of the rubble: at times complete and often times filled in and added on with different pieces. Involuntary memory forms that ground upon which we walk, layers after layers, sedimented on top of each other.

Thus, earthquake survivors neither forget what their living room looked like, the street that they grew up on and the games they played; nor the days of waiting, silence, and the question “Can anybody hear me?” Building new living centers and new housing adds new layers and new sediments, but neither erases the memories nor resolve the trauma. The question thus how to engage with those sediments, without assuming that they will ever be fixed and complete.

If this is true, it has larger implications for the study of memory than the rebuilding of disaster sites. It suggests that a strong dichotomy between memory and forgetting is quite insufficient, especially when it comes to involuntary memory.

Memories, even when they are incomplete, are nevertheless not erased: sediments remain, and become filled with other parts. So, our episodic memory of who we played with on our childhood street might be incomplete from the “actual occurrence” of events. And yet those fragments do not stand alone. They are thickened by other impressions, other sensations, other events that settle into them over time.

Memory does not preserve the past like a recording, rather it composes it. It also means that, to a large extent, our memory takes place not through the things that we actively recall, but the myriad other bits and pieces, added on top of each other, broken and filled in simultaneously. To think of memory as sediment rather than as a binary of remembering or forgetting is to shift the question. The issue is not whether something has been lost or preserved. It is how layers persist, how they press upon one another, and how we live within their composition.

Ege Selin Islekel is Assistant Professor of Philosophy and Affiliated Faculty in Race and Ethnic Studies at Texas A&M University. Her work explores the politics of death and mourning, with particular attention to memory, trauma, and collective remembrance. Her recent book, Nightmare Remains: The Politics of Mourning and Epistemologies of Disappearance (Northwestern University Press, 2024), analyzes the epistemic and political stakes of collective mourning.