Notes from p226-230 of Complete Psychology (2004) Edited by Graham Davey published by Hodder & Stoughton.

It's the core textbook for psychology at London Met and a very very easy read, so I usually start there when I'm trying to learn about anything. The other good thing is that the book usually suggests rather good further reading. I skim read then pick out the things which catch my eye for further investigation.

Modal model of memory - Atkinson & Shiffrin (1968)

Based on observing double disocciation of function in  neuropsych patients (one patient with impaired LTM but functioning STM and another with the exact opposite).

Consists of three connected modules; the sensory register, the STM and the LTM.

STM has control systems: coding, retrieval and rehearsal (seen as key to getting info from STM into LTM).

Criticisms of this model:

1. The suggestion that info can only get into the LTM via the STM appears not to hold true - one of the aforementioned neuropsych patients had a knackered STM but could still store things afresh in the LTM.
2. Rehearsal doesn't seem to be essential to getting things into the LTM. As was pointed out by one of my clever collaborators, lots of things are stored without effort.
3. Not much was suggested about the purpose of the STM other than remembering phone numbers (not sure I really get this point).

Enter Baddeley & Hitch (1974), who wondered further about the role of the STM. The specifically examined the role of the STM in understanding spoken senstences and in verbal reasoning.

Experiment:

Prose comprehension and simple reasoning tasks given to participants while they were simultaneously trying to memorise a set of digits.

The Atkinson & Shiffrin model would predict that the memorised digits would bung up the small capacity of the STM, leaving no processing capacity spare for the cognitive task. Therefore, performance on the cognitive task should be badly impaired. But it wasn't. (but maybe the participants found the cognitive task more interesting and actually didn't pay much attention to memorising the numbers)

This was the end of the modal model and the dawn of the working memory model (Baddeley & Hitch 1974; Baddeley 1986; 1998).

The working memory model

Sees the notion of STM as a passive receptacle supplanted by the idea of a dynamic system geared towards cognitive processing.

3 components to working memory (WM):

1. Central executive - supervises cognitive functioning, allocating resources & directing attention to the relevant task.
2. Phonological loop - a slave system of the central executive - speech-based, short-term store with limited capacity and limited duration. It's a bit like STM in the modal model and it accounts for the experimental findings on the nature of STM.
3. Visuospatial scratchpad - another slave of the central exec - a short term store for visual information. When someone asks you to say what colour your front door is, you invoke your visuospatial scratchpad in order to answwer.

Criticisms of this model:

1. Slave systems suspected not to play as significant a role in cognition as was expected.
2. The central executive is a black box and divers aspects of cognition are too often attributed to it.

Sounds potentially contentious. Goody!

Further reading:

Evaluation of WM model - Cognitive Psychology: A Student's Handbook - Eysenck & Keane

Baddeley, AD (1997) Human memory theory & practice (revised edn) Hove: Psychology Press