UNIT 1: MEMORY AND

FORGETTING
What is memory?
Memory is an active system that receives information from
the senses, puts that information into useable form and
organizes it as it stores it away, and then retrieves the
information from storage
1.1 KEY CONCEPTS
Three processes of Memory
1. Encoding- set of mental operations that people perform on sensory
information to convert that information into a form that is useable
in the brain storage system
2. Storage- holding on to information for some period of time
3. Retrieval- getting information that is in storage into a form that is
useable
 Features of the memory
1. Input
The process through which information from the environment enters the memory system to be encoded and
stored.

2. Capacity
The total amount of information that a particular memory store can hold at any given time.
Sensory Register: Very large (but briefly held).
Short-Term Memory (STM): About 7±2 items (Miller, 1956).
Long-Term Memory (LTM): Unlimited (theoretically)

3. Duration
The length of time that information can be retained within a specific memory store before it decays or is
forgotten.
Sensory Register: 0.25 to 2 seconds.
STM: About 18–30 seconds.
LTM: Minutes to a lifetime.

[Link]
The ability to retrieve or recall information that has been stored in a memory system when it is needed.
Types of memory
Procedural memory is a type of implicit memory: it stores information about how to do things. It is the memory
for skilled actions, such as how to brush your teeth, how to drive a car, how to swim the crawl (freestyle) stroke.
If you are learning how to swim freestyle, you practice the stroke: how to move your arms, how to turn your
head to alternate breathing from side to side, and how to kick your legs.
Declarative memory has to do with the storage of facts and events we personally experienced. Explicit
(declarative) memory has two parts: semantic memory and episodic memory.
Semantic means having to do with language and knowledge about language. An example would be the question
“what does argumentative mean?” Stored in our semantic memory is knowledge about words, concepts, and
language- based knowledge and facts. For example, answers to the following questions are stored in your
semantic memory:
• Who was the first President of the United States?
• What is democracy?
• What is the longest river in the world?
Episodic memory is information about events we have personally experienced. Episodic memory is memory
about happenings in particular places at particular times, the what, where, and when of an event (Tulving,
2002). It involves recollection of visual imagery as well as the feeling of familiarity (Hassabis & Maguire, 2007).
 Forgetting
• Amnesia is the loss of long-term memory that occurs as the result of disease, physical trauma, or
psychological trauma.
There are two common types of amnesia: anterograde amnesia and retrograde amnesia

• In anterograde amnesia, you cannot remember new information, although you can remember
information and events that happened prior to your injury. The hippocampus is usually affected
(McLeod, 2011). Many people with this form of amnesia are unable to form new episodic or semantic
memories, but are still able to form new procedural memories.

• This was true of H. M., which was discussed earlier. The brain damage caused by his surgery resulted in
anterograde amnesia. H. M. would read the same magazine over and over, having no memory of ever
reading it—it was always new to him. He also could not remember people he had met after his surgery.
If you were introduced to H. M. and then you left the room for a few minutes, he would not know you
upon your return and would introduce himself to you again. However, when presented the same
puzzle several days in a row, although he did not remember having seen the puzzle before, his speed at
solving it became faster each day

• Retrograde amnesia is loss of memory for events that occurred prior to the trauma. People with
retrograde amnesia cannot remember some or even all of their past. They have difficulty remembering
episodic memories.
1.2 CRAIK AND LOCKHART’S
LEVELS OF PROCESSING (LOP)
MODEL OF MEMORY
Levels of processing model (Craik and Lockhart)
Levels of processing model is the model of memory that assumes information
that is more “deeply processed,” or processed according to its meaning rather
than just the sound or physical characteristics of the word or words, will be
remembered more efficiently and for a longer period of time.
Hence, in Craik and Lockhart’s Levels of Processing (LoP) model, memory is
understood not in terms of separate stores like in the modal model, but based
on the depth at which information is processed. This depth determines how
well the information is remembered.
A key idea in the LoP model is that storage is not a separate step—it is a by-
product of processing. This means that the more deeply we process
information, the more likely it is to be stored in memory and retrieved later.
There is no need for rehearsal or transfer between stores; rather, what matters
is the quality of engagement with the material during encoding.
Craik and Lockhart’s levels of processing (LoP)
model of forgetting
Craik and Lockhart’s Levels of Processing (LoP) model explains forgetting not as a result of
memory decay or displacement, but as a consequence of shallow processing during
encoding. According to this model, memory retention depends on the depth at which
information is processed—structural (appearance), phonemic (sound), or semantic
(meaning).

When information is processed only at a shallow level (e.g., focusing on the way a word
looks or sounds), it forms a weak memory trace, making it more likely to be forgotten. In
contrast, deep (semantic) processing creates stronger, more durable memory traces,
reducing the likelihood of forgetting.

Unlike traditional models that attribute forgetting to limited capacity or retrieval failure,
the LoP model views forgetting as a failure to encode information meaningfully in the first
place. Therefore, the model emphasizes that the quality of initial processing, not time or
rehearsal, is the key factor in whether information will be remembered or forgotten.
The model distinguishes between shallow processing and deep processing.
Shallow processing involves focusing on surface-level features such as
appearance or sound, which usually leads to weaker and short-lived
memories.
In contrast, deep processing involves focusing on the meaning of the
information, leading to stronger, more durable memory traces.
Craik and Lockhart proposed three levels of processing
1. Structured processing (appearance)
Processing based on appearance or physical features (e.g., "Is the word in
capital letters?").
2. Phonemic (Intermediate)
Processing based on sound (e.g., "Does this word rhyme with cat?").
3. Semantic processing (deep)
Processing based on meaning (e.g., "Is this word a type of animal?").
 Phonemic
Physical
Sound Meaning
appearance
Features of memory according to LOP model
Information can be processed at three levels:
Structural (shallow) – based on appearance (e.g., capital letters)
Phonemic (intermediate) – based on sound (e.g., rhymes)
Input
Semantic (deep) – based on meaning (e.g., category or use)
Deeper processing leads to better memory.
Hence, information input at the semantic level of processing is better remembered than structural or phonemic input.

Not limited by a fixed number.

Capacity Semantic (deep) processing results in better retention and meaningful storage. Structural and phonemic (shallow to
intermediate) processing leads to weaker memory traces and reduced capacity to retain useful information.

Semantic (deep) → Long-lasting memories

Duration Phonemic (intermediate) → Moderate duration
Structural (shallow) → Very short duration; information quickly forgotten

Semantic (deep) → Easier retrieval due to strong, meaningful associations

Phonemic → Moderate access
Access
Structural (shallow) → Poor retrieval; minimal associations make recall difficult
BETTER ENCODING- BETTER RETRIEVAL
 Named Studies- Levels of processing (Craik and Tulving, experiment 2 only)

Aim: To investigate the effect of depth of processing on memory.

Participants: 24 college students, including males and females

Procedure: Participants in the study were presented with a series of 60 common five-letter words. They read a ‘yes/no’ question
about each word to prompt processing at one of three levels: structural, phonemic, or semantic. The questions were presented in a
random order.
• Structural Processing: Participants were instructed to focus on the physical appearance of the word, e.g. “Is the word in capital
letters?”
• Phonemic Processing: Participants were instructed to focus on the sound of the word, e.g. “Does the word rhyme with ?”
• Semantic Processing: Participants were instructed to think about the meaning of the word, e.g. “Would the word fit in the
sentence ___________ ?”

After processing the words at one of these three levels, participants were given a typed list of 180 words: the 60 original words
and 120 distractor words. The participants identified the words in the typed list they had recognised from the first part of the
experiment. This recognition test was the measure of memory performance.
Results:
• Participants who had processed the words at a semantic level recognised more words than those who had processed the words
at phonemic or structural levels.
• Participants who had processed the words at a phonemic level recognised more words than those who had processed the words
at a structural level.
Conclusion: Deeper processing of information during encoding leads to better memory. This provides evidence for the influence of
different levels of processing on memory.
Argument for the LOP model
Craik and Lockhart’s Levels of Processing (LoP) model proposes that memory is influenced by the depth at which
information is processed, rather than by the existence of separate memory stores. According to this model, deep
(semantic) processing leads to stronger and longer-lasting memories, while shallow processing (structural or phonemic)
results in weaker memory traces.

Support for this model comes from a study conducted by Craik and Tulving (1975, Experiment 2), which aimed to
investigate the effect of different levels of processing on memory recall. In this experiment, 24 college students were
presented with a series of 60 five-letter words, each followed by a yes/no question that encouraged either structural,
phonemic, or semantic processing. For example, a structural question might ask if the word was in capital letters, a
phonemic question might ask if the word rhymed with another, and a semantic question might ask if the word would fit
in a particular sentence. The questions were presented in a random order to ensure varied processing. After the
processing phase, participants were given a recognition test containing 180 words (the original 60 and 120 new
distractor words), and were asked to identify the words they remembered from the first part.

The results showed that participants recognised significantly more words that had been processed semantically,
compared to those processed phonemically or structurally. Recognition was also better for phonemic processing than
for structural processing.

These findings support the LoP model by demonstrating that deeper, more meaningful processing leads to better
memory performance, regardless of repetition or time. The study provides strong evidence that it is not the type of
memory store or amount of rehearsal that determines recall, but rather the quality of cognitive engagement during
encoding.
Argument against the LOP Model
• One main criticism of Craik and Lockhart’s Levels of Processing (LoP) model is that it
doesn’t explain the limited capacity of memory.
• The model does a good job showing how deeper, more meaningful (semantic) processing
improves memory. However, it doesn’t offer a clear structure for how much information
we can handle at once. In contrast, models like the Multi-Store Model clearly describe
the limits of short-term and long-term memory. The LoP model doesn’t explain why we
might forget things due to overload or interference.
• This is a weakness because it means the model can’t fully explain why people forget
things even when they process them deeply. For instance, a student might study many
topics using deep understanding, but still forget some of them—not because the
processing was shallow, but because too much was processed too quickly. This can go
beyond the brain’s working memory capacity and cause interference.
• So, while the LoP model focuses on the quality of processing, it overlooks important
limits in how much we can actually process and remember—especially in real-life
situations.
Encoding Specificity Principle
Encoding specificity is the tendency for memory of information to be
improved if the retrieval conditions are similar to the conditions under
which the information was encoded.
Context-dependent memory means you're more likely to remember
something if you're in the same environment where you learned it.
Example: Studying in the library and recalling better in the library.
State-dependent memory means recall is better when your internal
state (mood or body) is the same as when you learned it.
Example: Learning something while you're calm and remembering it
better when you're calm again.
Application of knowledge of Craik and Lockhart’s levels of
processing (LoP) model of memory and forgetting to novel
scenarios
1. Studying for a History Exam
A student is preparing for a history exam and begins by skimming through the textbook, paying attention only to
the bolded words and headings. This involves structural processing, which is relatively shallow. Later, they read the
text aloud to themselves, focusing on the pronunciation and rhythm of the sentences—this moves into phonemic
processing. However, when they pause to understand the causes of a war, rephrase them in their own words, and
connect those ideas to recent global events or political patterns, they are engaging in semantic processing.
2. Remembering a Person’s Name
Imagine someone is introduced to a colleague named “Samantha.” If they focus only on how the name is spelled,
they are using structural processing, which is shallow and easily forgotten. If they repeat the name out loud or
notice that it rhymes with “Amanda,” that’s phonemic processing. But if they learn that Samantha works in
marketing—something they are also interested in—and link the name to a mental image of her giving a
presentation, they are engaging in semantic processing.
3. Learning a New Language
While learning French, a student comes across the word “château.” Initially, they notice its unusual spelling
(structural processing). Next, they repeat it several times to get the pronunciation right (phonemic processing).
However, when they learn that château means "castle" and start imagining French castles they’ve seen in films or
photos, and maybe even compare it to the English word “estate,” they begin to process the word semantically.
1.3 ATKINSON AND SHIFFRIN’S
MULTISTORE MODEL (MSM) OF
MEMORY
Multistore Model of Memory
The Multistore Model of Memory recognizes three types or stages of
memory: sensory memory, short- term memory and long- term memory
In the Atkinson-Shiffrin model, memory starts with a sensory input from the
environment. This input is held for a very brief time – several seconds at
most – in a sensory register. It can be stored as iconic (visual) memory
lasting for <1 s or as echoic (auditory) memory lasting for roughly for 2-4 s.
Information that is attended to and recognised in the sensory register may
be passed on to second stage of information processing, i.e. short-term
memory (STM) or working memory, where it is held for perhaps 20 or 30
seconds. Generally, 5 – 9 number of units can be processed at any given time
in STM.
Information that is rehearsed may then be passed along to long-term
memory (LTM); information not so processed is lost.
Unattended information Unrehearsed information
is forgotten is forgotten
Serial Position Effect
Serial position effect is a phenomenon in which people tend to
remember the first and last items in a list better than those in the
middle
Primacy effect refers to the tendency to remember items at the
beginning of the list more easily
Recency effect refers to the tendency to remember last items on the
list more easily
Forgetting
• According to the Multi-Store Model (MSM) proposed by Atkinson and Shiffrin (1968), the short-
term memory (STM) has a limited capacity, typically around 7±2 items. One reason for forgetting
in STM is displacement, which occurs when new information enters STM and pushes out
existing information due to lack of space.
• Decay refers to the fading of memory traces over time when information is not rehearsed or
attended to. In the MSM, decay occurs in both the sensory register (SR) and short-term memory
(STM). The sensory register holds information from the environment for a very brief period (a few
milliseconds to a couple of seconds). If attention is not paid, this information decays almost
instantly. Similarly, in STM, information typically lasts for about 15 to 30 seconds unless it is
actively rehearsed. Without rehearsal, the memory trace weakens and eventually disappears.
• In the Multi-Store Model (MSM), forgetting from long-term memory (LTM) is often explained by
retrieval failure, which means the information is stored but we are unable to access it when
needed. Long-term memory has a large capacity and can store information for a long time, but
sometimes we forget simply because we don’t have the right cues to help us remember. This is
known as an access problem. For example, during an exam, a student might struggle to recall an
answer even though they studied it well. Later, when relaxed, they may suddenly remember it.
This shows the information was still there in memory, but it couldn’t be retrieved at the moment.
Named studies- Serial position effect (Murdock)
Aim: To investigate the serial position effect of memory recall.
Participants: 103 psychology students, including males and females
Procedure: Participants were presented with a list of 10–40 common words. Each individual word
was read aloud to participants. Each word was presented for the same amount of time.
Participants were then tested on their memory recall. They had 90 seconds to write down as many of
the words as possible. The test was repeated with the same participants over a few days, using
different word lists each time.

Results:
• The results of Murdock’s study showed a primacy effect and a recency effect in word recall.
• The first three to four words at the beginning of the list (primacy effect) and the final eight words at
the end of the list (recency effect) were recalled the most.
• Words in the middle position on the list were the least recalled.

Conclusion: The primacy effect occurs because participants have more time to rehearse and transfer
words from short-term memory to long-term memory. The recency effect occurs when items are still
in short-term memory and readily available for recall. The findings support the theory that the short-
term and long-term memory stores are separate.
 Argument for MSM Model
• The Multi-Store Model of Memory, proposed by Atkinson and Shiffrin (1968), suggests that
memory consists of three distinct stores: the sensory register, short-term memory (STM), and
long-term memory (LTM). According to this model, information flows in a linear way from one store
to the next, and rehearsal is necessary for the transfer of information from STM to LTM.
• Support for this model comes from a study conducted by Murdock (1962), which aimed to
investigate the serial position effect on memory recall. In this experiment, 103 psychology students
were presented with lists of 10 to 40 common one-syllable words, with each word being read aloud
at a steady pace. After each list, participants were given 90 seconds to write down as many words
as they could remember. The test was repeated with the same participants over multiple days using
different word lists, ensuring reliability of the findings.
• The results showed that participants recalled words from the beginning and end of the list much
better than those in the middle. This demonstrated both a primacy effect (better recall of words at
the beginning of the list) and a recency effect (better recall of words at the end). Words in the
middle were recalled the least frequently.
• These findings support the Multi-Store Model of Memory by suggesting that the primacy effect
occurs because the first few words are rehearsed and transferred into long-term memory, while
the recency effect occurs because the last few words are still in short-term memory at the time of
recall. The clear difference in recall patterns provides strong evidence for the existence of separate
memory stores with different functions and capacities.
Argument against MSM model
• The Multi-Store Model (MSM) of memory, developed by Atkinson and Shiffrin
(1968), describes memory as a linear process involving three separate stores—
sensory register, short-term memory (STM), and long-term memory (LTM). It
highlights rehearsal as the main way information moves from STM to LTM.
• However, a major criticism of the MSM is that it doesn’t consider the depth of
processing. Craik and Lockhart’s Levels of Processing (LoP) model argues that how
deeply we process information—especially through meaning (semantic
processing)—has a bigger impact on memory than just repeating it. Their 1975
study showed that people remembered words better when they processed their
meaning, even if they didn’t rehearse them.
• This challenges the MSM’s idea that rehearsal alone creates long-term memories.
Instead, it suggests that how we think about information during learning (deep
processing) is more important than how often we repeat it. In this way, the MSM
oversimplifies memory by ignoring the importance of meaningful engagement
during encoding.
Application of MSM Model
1. Forgetting Someone’s Name at a Party
Imagine you are introduced to someone at a noisy party. You hear their name, but you're distracted by music
and other conversations. The name briefly enters your sensory memory, but because you don’t pay attention
or rehearse it, it doesn’t enter short-term memory. As a result, you quickly forget the name.
This shows that attention is needed to move information from the sensory register to STM.
2. Cramming Before an Exam
A student studies a list of definitions right before a test by repeating them several times. This repetition holds
the information in short-term memory long enough to perform well on the test. However, because the
material was not deeply processed or rehearsed enough, it is quickly forgotten afterward, indicating it did not
reach long-term memory.
This supports the idea that rehearsal is needed to move information from STM to LTM, and lack of it leads to
forgetting.
3. Memorizing a Phone Number
When someone tells you a phone number, you might repeat it a few times (maintenance rehearsal) to keep it
in your short-term memory long enough to dial it. But unless you save it in your contacts or associate it with
something meaningful, you’ll likely forget it soon.
This illustrates the limited duration and capacity of STM and the importance of encoding strategies to store
information in LTM.