Make your own free website on

Volta para História da Memória


Texto original encontra-se em

 Human Memory

Short Term Memory: A History

  • It is often described as a moment in time – but how long is that?
  • The capacity of an ‘immediate’ memory preoccupied a number of philosophers of the 19th century
  • Original proposal by William James (1890) under the name of ‘primary memory’
  • Compared to STM, the term primary memory places less emphasis on time (i.e. duration of memory storage), and more emphasis on the roles of attention, conscious processing and memory capacity.
  • The first systematic experimental work to be done on STM was by Joseph Jacobs (1887)
  • He devised a technique called ‘Digit Span’ which has played an important role in memory research
  • Most people can manage 6 or 7 digits, but there is a large range (4-10+)
  • This can be improved by speaking them aloud or by chunking
  • Intensive interest in STM developed in the late 1950’s
  • This came about as a result of studies by Brown in the England and the Petersons in the US
  • They showed that even sequences in such a short memory span could show clear forgetting, IF, the individual was prevented from thinking about it or rehearsing it.
  • Now known as the Brown Peterson Test
  • The Peterson result caused enormous interest for at least 2 reasons:
    1. It offered a neat and economical technique for studying short term forgetting
    2. The Petersons interpreted their result in terms of trace decay
  • Release from proactive inhibition/interference
    • developed by Delos Wickens
    • demonstrated nicely by Gardiner, Craik & Birtwisle (1972)

o       sequences of flower names that were separated into clusters of wild and cultivated flowers

o       after a number of clusters of cultivated flowers it switched to wild flowers. No subject noticed this and thus none showed release from PI

o       However, one group was warned of the change and suddenly showed release from PI

o       A third group received this info AFTER the presentation of the critical sequence but BEFORE recall. In fact, they showed release from PI


Are Short Term and Long Term Memory Separate?

  • The most compelling cognitive and truly scientific evidence against a unitary view of memory is seen in free recall and the recency effect!
  • The following affects LTM but not STM:

1.    Rate of presentation

2.    Familiarity of the words

3.    Level of distraction

4.    Age

  • Equally compelling evidence against a unitary view of memory comes from studies with brain damaged populations, in particular, amnesia
  • Patient H.M is the most noted patient for demonstrating differences between LTM and STM (Scoville & Milner, 1957)

Patient HM

  • Anterograde amnesia causing him to forget episodes of daily life as rapidly as they occur
  • Became amnesic as a result of a bilateral surgical excisions of the medial temporal region to relieve him of severe epilepsy
  • The removal was intended to include the amygdala, the hippocampal gyrus and the anterior two thirds of the hippocampus.


Summary of abilities since surgery:

  • Fewer seizures
  • Good vocabulary
  • Normal language skills
  • IQ in the normal to bright-normal range
  • Retains older memories


  • Lost memory for events a couple years pre-surgery
  • New learning is severely impaired


  • Normal digit span
  • Normal short term memory capacity unless distracted
  • Information is lost if rehearsal is prevented

Animal Research Supports Above

  • Same distinction between LTM and STM has been demonstrated in animal research
  • Using a radial maze with eight platforms extending from a center, rats were taught to visit the eight arms in a sequence determined by the experimenter
  • Rats were given a choice test involving arms 1 vs. 2, 4 vs. 5, or 7 vs. 8 and were rewarded for entering the arm that had been visited earlier in the sequence
  • Bilateral lesions of the hippocampus eliminated the primacy but not he recency effect
  • The introduction of a delay resulted in the loss of the recency effect as well

Where are we Now?

  • 19th century talk of ‘immediate memory’
  • William James (1890) proposal of ‘primary memory’
  • Joseph Jacobs (1887) experimental work on Digit Span
  • Brown and Peterson experiments in 1950’s
  • Free Recall work around the same time
  • Patient HM starting in 1953
  • Animal research in early 1960’s


  • Atkinson and Shiffrin Model (1968)
  • Craik and Lockart (1972) Levels of Processing

Atkinson and Shiffrin Model

  • By now there is a strong belief in separate STM and LTM systems
  • As we know – A&S thought there to be the three major components
  • STM plays a crucial role because without it info can’t get into LTM
  • According to A&S, STM not only stored info but it also was involved in control processes

o      Rehearsal was a control process - they believed the longer info was maintained in STM the more likely it was to go to LTM

Problems with the Modal Model

  1. How do we explain people with impaired STM but unimpaired LTM?
  2. Tulving has demonstrated simply repeating words does not enhance subsequent learning, rather active learning did.

These problems resulted in the loss of interest in the general area of STM. At the same time, Craik and Lockhart’s (1972) levels of processing was becoming a hit.

Levels of Processing

  • Emphasized the importance of focussing on the MODE of processing rather than hypothetical memory structures.
  • The more ‘deeply’ an item is processed the better it will be remembered
  • They still believe in a primary memory – but it’s role is to process incoming information
  • Longer storage results from deep processing, not from transfer from one store to another
  • This resulted in the distinction between maintenance rehearsal and elaborative rehearsal
  • Normal ‘block-tapping span’ but could not learn past his span
  • Same thing shown in studies of serial position curve for free recall
  • Normals and amnesics given lists of words to recall
  • Normals show serial position effects (primacy and recency)
  • Amnesics showed absent or reduced primacy but normal recency