Construct · 6 fields · 18 citations
One name. Different field. Different construct. One map.
● jingle · one name, six constructs Nearest construct by meaningcollective memorycompare the meaningsEvery construct on the map is compared with every other by the meaning of its definitions, and the closest one is surfaced here. The nearest construct to memory is collective memory.
Hold this as a prompt, not a verdict. Two constructs sitting close in meaning is a reason to read them side by side and decide for yourself whether they are one idea under two names. It is not proof that they are. Compare the meanings in the Collider.
Spokes are fields; dashed arcs are shared lineage. Tap a field for its definition, unit, and measures.
Drag to watch the word get claimed. Each field appears at the publication date of the earliest source cited for it. Tap a year to see which.
A component that lets a model carry information across time steps or across calls, distinct from the weights learned in training. Hochreiter and Schmidhuber's 1997 long short-term memory fixed the recurrent version: a cell whose contents persist unless gates decide to write, forget, or read, solving the vanishing-gradient problem that had kept recurrent networks from learning long dependencies. Weston, Chopra, and Bordes's 2015 memory networks separated the store from the network entirely, an addressable array the model reads by attention. In large language models the word now covers the context window, retrieval over an external store, and summaries written back by an agent, which Zhang and colleagues' 2024 survey sorts by what is stored, how it is written, and how it is read. In every case memory is a buffer with an addressing scheme, and forgetting is a design choice.[1][2][3]
The addressable storage a processor reads and writes, organised as a hierarchy (registers, caches, main memory, disk) so that the small fast levels hide the latency of the large slow ones. Denning's 1970 survey fixed virtual memory: the addresses a program uses are a fiction mapped onto physical storage by the system, which pages data in on demand and relies on locality of reference to make the illusion cheap. Hennessy and Patterson's textbook made the hierarchy the organising principle of the field, with hit rate and miss penalty as its two numbers. Nothing here is remembered in any psychological meaning; memory is a resource measured in bytes and nanoseconds, and its content is whatever was last written to the address.[4][5]
The set of processes by which people encode, hold, and retrieve information, carved into systems by the patterns of what is remembered and forgotten. Atkinson and Shiffrin's 1968 model fixed the multi-store account: a sensory register, a short-term store with control processes such as rehearsal, and a long-term store. Baddeley and Hitch replaced the short-term store in 1974 with working memory, a central executive and two slave systems, argued from dual-task experiments in which holding digits barely hurt reasoning. Tulving split long-term memory in 1972 into episodic, for personally experienced events with a time and place, and semantic, for facts without one. The units are stores and systems; they are inferred, not seen, from dissociations in performance.[6][7][8]
A lasting change in the strength of connections between neurons, produced by activity and read out as altered firing. Hebb proposed the rule in 1949: when one cell repeatedly helps fire another, the connection between them grows. Scoville and Milner's 1957 report on patient H.M., who lost the ability to form new long-term memories after bilateral removal of the medial temporal lobes while keeping old memories and motor learning, tied a psychological system to a structure and showed that storage and formation come apart. Kandel's 2001 Nobel lecture set out the molecular version from Aplysia: short-term memory is covalent modification of existing synaptic proteins, long-term memory requires gene expression and the growth of new synaptic connections. The unit is a synapse, and remembering is a cell that fires differently than it used to.[9][10][11]
A witness's account of what they perceived, offered as evidence, and the question of how far a fact-finder may rely on it. Loftus and Palmer showed in 1974 that the wording of a question changed the speed people reported for a filmed collision and whether they later remembered broken glass that was not there; Loftus's 2005 review summarised thirty years of the misinformation effect, post-event information rewriting the report. Wells and Olson's 2003 review sorted the variables into those the justice system controls, such as lineup procedure, and those it does not. The law's own test is the Supreme Court's 1972 decision in Neil v. Biggers, which lists five factors for whether an identification is reliable despite a suggestive procedure. Memory here is testimony, its owner is a witness, and the criterion is whether a court may admit and a jury may believe it.[12][13][14]
The shared representations of the past that a group maintains, and the practices (commemoration, monuments, archives, ritual) by which it does so. Halbwachs argued in 1925 that even individual recollection depends on social frameworks: we remember as members of families, classes, and religions, and a memory with no group to sustain it fades. Nora's 1989 essay separated memory, lived and carried by communities, from history, the reconstruction of a past no longer lived, and named the lieux de mémoire, sites where memory crystallises once its milieu is gone. Assmann's 1995 distinction between communicative memory, which lasts about three generations of oral transmission, and cultural memory, fixed in texts and rites, gave the construct a time scale. No brain is involved; the carrier is a group, and the memory is a property of its institutions.[16][17][18]
Next
Combine items across the memory scales to see whether the meanings separate empirically, or export every reference on this page in one click.
Open in ColliderExport BibTeXReport this collisionCited on this page?
If your work is cited behind one of these entries, you know better than we do whether we captured the meaning right. Confirm it, suggest an edit, or dispute it, and your name goes on the entry as an author who has checked it, linked to your ORCID.
Claim your construct