[DSRP Evidence](https://dsrpevidence.org/)

# What is memory?

## The research claim

Memory is usually described as storage and retrieval. Something happened, a record of it was filed, and remembering fetches the record back. The metaphor is so settled that the failures get treated as defects in the filing: memories fade, get corrupted, become unreliable. Memory is the mental model at a time. Not a copy of what arrived, but the model as it stood — information organized, held. Remembering is running that model again. Everything the storage account calls a defect follows from this without needing a separate explanation. Models decay, so memories fade. Models are built from a point of view, so memories carry the bias of the point they were built from. Models mix with other models, so memories contaminate each other. None of these is damage to a record. They are the ordinary behaviour of a model, and a record would not behave this way at all. It also explains why remembering changes what is remembered. A record read twice is the same record. A model run twice has been run twice, and running it is one of the things that changes it.

## What would confirm it

Memory is the mental model as it stood at a time. Remembering is running that model again, not opening a file. Read in symbols: memory is the mental model at time t (𝕄ₜ), which is the information and the organization as they were at that moment ((I · O)ₜ). Both terms are subscripted because both change. What was known changes, and how it was arranged changes, and a memory is the pair of them held together at a point in the past.

## What would refute it

Memory is something other than a model held at a time — a stored record, retrieved rather than run. Read in symbols: memory is not identical to (≠) the mental model at time t. The clearest version of the rival is a faithful record: something laid down once, unchanged by being read, degraded only by damage. If memory behaves that way, this account is wrong.

## The evidence, and why this status

If a mental model is information times organization, then memory cannot be a stored record. It has to be a mental model held at a time, and that is a different construct with different properties — one that is rebuilt on each use rather than retrieved intact. Nothing here was designed as a memory study. What the studies establish is what a mental model does, and the claim is that memory is one, so evidence about the first is evidence about the second. Four of them converge on the same property from different materials. One shows a model supplying structure the stimulus did not contain: free description of an image built to mean nothing returned a small set of definite things. What came back was not what was there. One shows the same content held many ways at once. Six familiar objects, sorted freely by hundreds of people, produced hundreds of differently named wholes. A stored record of the objects would yield one arrangement. One shows the holding being altered by what came immediately before it. Priming one word of a pair moved what people produced next, in both directions the pair was tested. The information available never changed; the model doing the work did. And one bears on what a memory is made of: across standard language benchmarks, how information was structured predicted interpretation many times over while the amount of information predicted nothing. Different stimuli, different tasks, one property each time — what is held is constructed, not retrieved. That is the construct the form requires, and this is initial evidence for it.

## Supporting evidence (4 publications · 3 studies)

- [Cabrera D, Cabrera L (2025). The structure of thought and reality: empirical evidence for the three universal laws linking mind, matter, and meaning. Journal of Systems Thinking.](https://dsrpevidence.org/paper/504)

- [Cabrera D, Cabrera L (2025). From DSRP Theory to O-Theory: The Science of Organization. Journal of Systems Thinking 5(1):1-8.](https://dsrpevidence.org/paper/473)

- [Cabrera D, Cabrera L (2020). Defining Learning.: Defining Learning. Handbook of Systems Thinking.](https://dsrpevidence.org/paper/285)

- [Cabrera D, Cabrera L, Cabrera E (2022). Rorschach Inkblot. Systems 10(2):41.](https://dsrpevidence.org/experiment/339)

- [Cabrera D, Cabrera L, Cabrera E (2022). Dog–Lab–Coat. Systems 10(3):71.](https://dsrpevidence.org/experiment/319)

- [Cabrera D, Cabrera L, Cabrera E (2022). Sort Stuff. Systems 10(2):44.](https://dsrpevidence.org/experiment/343)

- [Cabrera D, Cabrera L (2026). Measuring Context as Structural Organization. Journal of Systems Thinking 6(1).](https://dsrpevidence.org/paper/604)

[All research questions](https://dsrpevidence.org/claims)