• Attributes
      • Duration – Long (can be a lifetime)
      • Capacity – very large (functionally infinite)
      • Coding – almost certainly a variety of codes
      • Retrieval: not like STM (serial, exhaustive scanning would be too slow)


    • Information loss
      • 3 suspects from STM research:
        • displacement (Atkinson & Schifrin)
        • decay (Peterson and Peterson) (like bananas)
        • Interference (Waugh and Norman)
      • Do any of these apply to LTM?


    • Theories for Forgetting
      Recall probability over number of intervening ...
      Image via Wikipedia
        • Memories deteriorate over time (like bananas); it’s intuitively plausible, seems reasonable, the idea of “using it or losing it”
        • Problems with decay theory:  These problems come from different lines of evidence. Thus, problems are thought to be more severe.
          • Physiological evidence
            • We lose neurons every day but does it really matter. Probably not since we start with so many. Some estimate we lose less than 2% over a 70 year period. Also recent research suggests that new neurons are created (neurogenesis).
          • Experimental evidence – (Jenkins and Dallenbach) 1925
            • Two groups of subjects learned a 10 item list of nonsense syllables. One group learns them early in morning (day subjects) and a second group learning the list at night (night group). The day subjects returned to the lab 1, 2, 4, and 8 hours later and attempted recall. The night subjects were awoken 1, 2, 4, and 8 hours after retiring and attempted recall. Night subjects did better on recall. If decay theory were the case, the two would do equally as well. So something else is going on…… (interference is my guess)
          • Anecdotal Evidence
            • Memory in the elderly – they can often recall long-ago events better than recent events; exactly the opposite of the prediction of decay theory


      • Interference Theory
        • information in LTM can interfere with the ability to recall other information
          graph chart about proactive and retroactive me...
          Image via Wikipedia
            • Old learning interferes with new learning, shown in list learning experiments using various lists. The two lists are similar semantically. They mix up which list a certain word was part of because of the similarities. Example: trying to learn a new phone number
          • Retroactive interference
            • New learning interferes with the old learning. Two similar lists of animals shown to subjects. Asked to recall the first list. The new stuff interferes with old stuff. Example: trying to remember an old phone number. Which explains Jenkins and Dallenbach’s results? Day subjects had more retroactive interferences.


Encoding Specificity

  • State Dependence (also called contextual theory)
    • A memory is available in (LTM) but not accessible, however given enough cues, the memory could be retrieved. A specific example: encoding specificity (Tulving and Thompson, 1973).
      • It’s easier to remember something if the same cues present at learning are also present at the time of the retrieval.
      • Some memory cues:
        • Locations : where something was learned, revisiting elementary school
        • Time: when something was learned, “anniversary depression”
        • Mental state: how you felt at the time (emotional state, state of awareness)
      • Experiment:
        • Subjects learn things either sober or drunk. Then asked to recall later on either drunk or sober.
        • We would expect environmental support (how the cue is exhorting its affect) : people who were sober or drunk in both learning and recall should perform better than the others.
      • Implications
        • Nothing may ever be really “forgotten”
        • It may simply become hard to retrieve
        • like a mis-shelved book in the library but it’s available (in the library) but not accessible (easily found)
        • What varies, from moment to moment, is the amount of environmental support.


      • An example: Tip-Of-the-Tongue State (TOT) State
        • Temporary inaccessibility of a memory. You know that you know something but can’t retrieve it (like an actor’s name)
        • Studied in the lab by Brown and McNeil (1966) – definitions of uncommon words were used to trigger TOT states
        • Brown and McNeil’s results
          • Subjects were above chance at recalling something about the word: length of the word, number of syllables and first letter.
          • Subjects errors were often similar in sound, but not in meaning: sextant =>sexton
          • Shows evidence for multiple types of codes in LTM.
        • Implications
          • LTM codes memories in a variety of ways: sound, meaning, length
          • At any point in time, some of this information may be more available than at other times.
          • Words that sound alike may be stored near each other in LTM. People ion TOT states go through all similar words in their heads and create a lot of interference making it harder for them.
Related articles