34
Chapter i A. Introduction And Historical Resume The simplest assumption one could make about the contents of permanent memory is that they are simply copies of reality. People's faces in permanent memory might be like photographs; words like words; smells like smells, and the like. This implies that there are no coding operations - operations that create a representa- tion of the input - or codes in permanent memory. The coding processes will use some or all of the information available in the input to form a representa- tion of the input. For example a visual image of the person's face and an audit'oing image of the person's voice would be components of the representation. Visual Permanent Memory: Perhaps our most impressive memory feats are associated with pictures and human faces for example, in one study by Standing et. al. (1970), subjects were able to recognize color slides of pictures at a success rate in excess of 9CSi several days after they had been shown 2,56 0 stimuli.

Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

  • Upload
    others

  • View
    2

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

Chapter i

A. Introduction And Historical Resume

The simplest assumption one could make about the

contents of permanent memory is that they are simply

copies of reality. People's faces in permanent memory

might be like photographs; words like words; smells like

smells, and the like. This implies that there are no

coding operations - operations that create a representa-

tion of the input - or codes in permanent memory.

The coding processes will use some or all of the

information available in the input to form a representa-

tion of the input. For example a visual image of the

person's face and an audit'oing image of the person's

voice would be components of the representation.

Visual Permanent Memory:

Perhaps our most impressive memory feats are

associated with pictures and human faces for example, in

one study by Standing et. al. (1970), subjects were able

to recognize color slides of pictures at a success rate

in excess of 9CSi several days after they had been shown

2,56 0 stimuli.

Page 2: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

Memory for faces may involve a unique memory system,

Yin (1970) studied brain damaged patients who were suff-

ering from an inability to recognize familiar faces in

a recognition - memory task. Based on the observation

that inverted faces are more difficult to recognize than

are inverted photographs of buildings, Yin used both sets

of stimuli with patients who were suffering brain damage

in the right or left cerebral hemisphere. For the normal

controls, inversion of faces led to substantially poorer

facial recognition, as it does for patients with lesions

in the left parietal lobe. However, facial inversion did

not appreciably after the ability of patients with right

lesions to identify faces. This result seems to indicate

that particular visual cues are associated with faces

that serve as a reliable base for recognition, and that

the ability to use these cues is lost among patients

suffering certain lesions in the right hemisphere. There

is reason to believe that the encoding operations perf-

ormed on faces are normally based on semantic features

such as the so called "pleasantness" of a face. However,

such features are less discernible in inverted faces,

thus making the faces less memorable. The fact that right

damaged patients were relatively unaffected by inversion

suggests, that they do not encode faces along semantic

dimensions. We shall return to the notion of semantic

encoding momentarily.

Page 3: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

Auditory permanent memory:

The experience of identifying a friend's voice

which has not been heard for many years is common to

us all. Surprisingly, very little systematic informa-

tion is available concerning the recaliability of

auditory information. One exception to tfciE generali-

zation is memory for music and musical notes. Although

relatively permanent memory for tones is usually poor,

individuals with perfect pitch are able to remember tones

for long periods of time. Acc. to Siegel (1974) those

with perfect pitch are able to label a note and later

use the label in remembering the tone, rather than

committing the tone itself to memory.

Generally, there is good evidence that stimuli are

coded on a number of auditory dimensions stored in perma-

nent memory, and used as a basis for recognition in

subsequent memory tasks.

Coding of Context:

Episodes or events obviously do not occur in a

temporal or spatial vacuum. They always have a context.

Contextual information includes specific information

about when and where an event occurred, how often it is

known to have occurred and about other circumstances

surrounding the event. Each of these elements depending

on the goals of the central processor, may be encoded

Page 4: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

and may become a component in the representation of

the episode.

Temporal, Locative & Frequency Coding;

In addition to knowing something about the

characteristics of the event itself, we also know the

time and location of its occurence and how many times

it occurred. That we code information with respect

to time and location and frequency.

Elaborative Coding of Context:

An episode is embedded in a context. We evidently

do not have as good a memory for an episode when the

context cues present at coding and storage time are

not present at retrieval time.

Semantic Encoding:

Memory representations usually include more

than information about a specific episode and its

context. Most of all, we remember the meaning of the

input. While we are capable of retaining modality -

specific information, and will do so for certain

information forms such as pictures and faces especially

when the task demand it, we seem more inclined to base

the coding of input on more abstract,- modality - free

dimensions referred to as semantic encoding.

A convenient way of illustrating a subjects'

Page 5: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

i:;5

preference for semantic coding of information is to

present the subject with a series of words that are to

be tested later for recognition. After a lengthy reten-

tion interval, the person is shown a series of words some

of which are new & others which he or she has previously

seen. The new words may be semantically similar to one new

of the old words (e.g. The old word is bat, the/word

is ball) or phonetically similar (e.g., the new word is

bat). Semantically similar new Words are more likely to

be falsely recognized as the old words than are phoneti-

cally similar words (Anisfeld and Khapp 1968). Thus, in

this type of experiment, the words were coded semantically

(e.g. a ball is an object to be used in a base ball game).

Even though the old words could have been encoded

phonologically, subjects opted for a meaning code.

It is interesting to note that some alternatives

to the three-store approach have been willing to accept

the existence of sensory registers but have instead

presented argudnents that Short and Long term stores may

better beviewed as a single store (e.g., Murdock, 1974;

Wickelgran 1970) .

The second basi^ component of our system is the

ST store. This store may be regarded as the subject's

"working memory". Information entering the STS is

assumed to decay fljnd disappear completely but the time

required for the information to be lost is considerably

Page 6: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

h';6

longer than for the sensory register. The character of

the information in the ST store does not depend necessa-

rily upon the form of the sensory input.

The last major component of our system is the

long term store. This store differs from the preceding ones

in that information stored here does not decay and become

lost in the same manner. All information eventually is

completed and lost from the sensory register and the

STS, where as, in the LTS is relatively permanent.

Coltheart (1972) attempted to study the role of

STS in concept formation by means of the acoustic simila-

rity effect, the tendency for STM to be disrupted when

the material to be ronembered comprises items that are

phonemicaiiy similar to each other (Baddelay, 1966;

Conrad 1962). She contrasted the effect of acoustic

similarity on concept formation with that of semantic

similarity, which typically effects LTM rather than

STM (Baddeley 1966a). Unfortunately for the working

memory hypothesis, here results showed clear evidence

of Semantic rather than acoustic coding, suggesting that

the LTS rather than STS was playing a major role in

concept - formation task.

In a recognition experiment items are presented

as in free-recall experiments, but different testing

procedures are employed. Basically, there are three

Page 7: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

types of recognition tests; all three involve the use of

distractor items. The distractor items are selected

from the same set as the learning items, and new distra-

ctor items are used on each trial in multitrial recogn-

ition experiments. In a single item tests, learning

items and distractof items are presented one at a time

in random order, and the subject responds with either

"old" or "new" to each item. In a multipe - choice test

each learning item is canbined with several distractor

items and the subject is asked to pick out the old item

from each set of alternatives.

In all recognition tests the similarity between

the learning items and the distractor items is a very

powerful variables. Underwood & Freund (1968) have shown

that if subjects are given a multiple - choice recogn-

ition test consisting of the correct word, a high assoc-

iate of the correct word, a formally similar word, and

a neutral word, the high associate words provide the

major source of errors. Similarly, Anisfeld & Knapp (1968)

have demonstrated that in a Yes - no recognition task in

which subjects were shown words that had been presented

before, words that were common associates or synonyms

of preceding words, and neutral control words, subjects

made more false - recognition responses to ccmmon associ-

ates and synonyms than to control words.

Page 8: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

In recognition tests it is often possible to

recognize an item correctly on the basis of some remem-

bered detail; in recall, on the other hand* memory for

an isolated detail is usually less helpful. The nature

of the learning material determines how important

part - recognition will be if the learning material

is poorly integrated, for instance if consonant trigrams

are to be studied, part - recognition may be quite

effective, depending upon the confusability of the

distractor items used in the experiment.

In the discussion of face recognition studies,

we come to know that the encoding of faces encompasses

several dimensions. Faces need not be coded only in

term of their physical and structural features. Some

facial expressions are seen as pleasant, and others

as threatening. Perhaps if faces were evaluated along

such dimensions, rather than physical features, they

would be more memorable. Indeed, a study by V/arrington

5iAckroyd (1975) confirms the prediction: recognition

memory of faces was enhanced by semantically encoding

faces in terms of their pleasantness. This finding

suggests that the crucial cues which are distorted by

inverting faces, and produce low recognition scores, >

are ones the enable social semantic encoding. The fact

that right damaged patients in Yin's study (1970) were

able to identify inverted faces about as well as those

Page 9: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

I'O.Q

In an upright position indicates that they were not

using semantic encodings, but rather were basing

their representations on physical and structural features

of the faces.

Recognition Versus Recall;

It is not uncommon to find that people who have

learned a list of words in a psychological experiment

are able to recall for fewer of the words than they

will recognize as having occurred on the list. This

is certainly consistent with our every day experi-

ences. We may not recall the name of a popular song,

but we can often recognize whether a name provided is

correct or not. Pew of us could recall all of the

presidents of our country, but we would probably

recognize the names of most of them if we were shown

a list containing names of presidents and nonpresidents.

How can we account for such an observation? And is

the superiority of recall over recognition uniformly

the case?

The models of memory have received a good deal

of attention from Psychologists in recent years. One

model takes the view that recognition is superior to

recall because recall requires word generation and

recognition of the generated word as one that had

appeared on the list. Recognition, by passing the

generation phase, requires only the latter. This is

Page 10: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

referred to as a generation recognition model of

retrieval.

Stimulus Recognition:

Before a subject can retrieve the appropriate

response in a paired associate test he must be able

to recognize the stimulus term. Several experiments

have shown that recognition of the stimulus term is a

necessary condition for the establishment of a stimulus

response connection (Bernbach 1967, Martin 1967).

As an inllustration we may take a study by Martin (1967)

in which subjects learned eight trigram - number pairs.

Study trials and test trials were alternated, and on

each test trial the subject had to make two responses

a yes - no recognition response and recall of the

response digit, guessing if necessity. On test trials

the eight stimulus items were shown together with 16

distractor items inrandom order. When the recognition

response was correct, the probability of recall incre-

ased over trials as a learning curve should; when the

recognition response was incorrect, no recall learning

occurred, and the probability of recall oscillated

around the value of 1/8 expected by chance. Even more

impressive is the finding that it does not matter how

often the si±iject has already given a correct recall

response to a particular item on previous trials: If he

Page 11: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

11

fails to recognize it recall performance on that trial

is at the chance level. Similarly Bernbach (1967) has

reported that it makes no difference how often a subject

has already correctly recognized an item previously;

as long as he does not recognize i"^n a given trial,

recall performance is no better than chance.

An interesting observation with respect to

stimulus recognition has been made by Wicker (1970)

who compared learning with pictorial and verbal stimuli.

It is well known that subjects acquire a paired associ-

ate list with pictorial stimuli much faster than a

verbal list. Wicker was able to show that this difference

lies entirely in the greater recognizability of

pictures.

If stimulus recognition is a necessary pre-

requisite of associative recall. It follows that

anything that makes stimulus recognition difficult

should interfere with learning. Thus items that are

well integrated and easily recognized, are encoded in

the same way on different trials and indifferent contexts

and should be easier to learn than items that the subject

responses to differently every time he encounters them.

This appears to be one reason why low-meaningful non-

sense syllables give the learner so much trouble. On

one occasion TLQ may be encoded as "something tht starts

with a T", on the next as "encoding with a Q", and

Page 12: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

12

whatever is learned about one will be unrelated and

independent of what is learned about the other.

I magery:

Frost (1972) gave subjects simple line drawings

of 16 common objects to study. The objects could be

classified in two ways; in terms of semantic categories

(e.g. there were four pictures of animals, four pictures

of vehicles etc.) or in visual terms (the main axis of

the picture was horizontal, vertical or slanted either

right or left). Half of Frost's subjects expected a

recognition test and half expected a recall test. M l ,

however, were asked to recall the pictures. The

subjects who prepared themselves for a recall test

clustered items in their recall by semantic categories.

Subjects who were led to expect a recognition test,

however, recall the pictures either in semantic or shape

clusters. That is, these subjects ted stored in their

long term memory information about the visual shape

of the stimulus items and had grouped these items in

terms of their shapes.

This study demonstrates that visual representations

may be retained in memory, which should of course,

surprise no one. We hBe mentioned recency, frequency and

modality information in memory, as well as acoustic and

semantic. It is, however, true that until recently the

study of how visual characteristics are represented in

memory has been relatively neglected given the

Page 13: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

preoccupation of psychologists with verbal materials,

A change in this stete of affairs was brough about

largely through the work of Paivio, whose dual coding

theory of memory has been a major stimulus for work

in this area (e.g. Paivio, 1971, 1975).

Paivio maintains that there are two independent

but interconnected memory systems; an imagery system that

contains holistic analogues of things, and a verbal

memory that bears no resemblance to things.

Linguistic memory and processing appear to be

localized in the dominant hemisphere of the brain

(usually the left one). Nonlinguistic memory and process-

ing (e.g. spatial relations and music) are localized

in the right or nondominant hemisphere. The two systems

appear to be quite independent in that a person with

lesions in one half of the brain may show\ery specific

linguistic deficits, while his nonlinguistic perform-

ance remains unimpaired, and vice versa (e.g. Gazzaniga,

1970). It is even claimed that normal subjects process

verbal inputs better with their right hemisphere, while

the opposite is the case for nonlinguistic processes

(Kimura, 1973).

The neuropsychological data that suggest separate

storage and processing for words and imagery are in

good agreement with a large body of test data that show

Page 14: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

people's verbal and nonverbal abilities to be relat-

ively independent ( e.g./ Guilford 1967). Behaviourally^

the imagery and verbal systems are also independent to

some extent. We have already reviewed the considerable

evidence that the case with which subjects can form

an image or a word (i.e. its imagery value) is one

of the most powerful determinants of learning words

that are easy to imagery are easy to learn. If learning

is incidental, pictures are generally remembered better

than words - except when the words are imaged during

learning, in which case they are retained as well as

pictures, as was denvDnstrated by Paivio and Csapo(l973),

The same authors also showed that when an item is

repeated first as a word and then as a picture the

effect of this repetition is significantly greater

then when it is repeated twice as a word or twice as

a picture what is involved here is of course another

instance of encoding variability.

Paivio's dual coding theory postulates not only

the existence of independent verbal and imagery systems,

but also makes a qualitative distinctions.

A study that suggested that memory for the

meaning of a sentence must be distinguished from memory

for the linguistic expression itself was reported in

1967 by Sachs. Sachs read subject a paragraph of normal

English test. One of the sentences in the paragraph

Page 15: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

;i5

was selected as the test sentence. Subjects were

given a recognition test after listening to the

paragraph with either the sentence in its original

form^ a formally changed sentence. The subject was

asked to tell whether the test sentence was the

same as one in the original passage or not. When the

test sentence was embedded somewhere in themiddle of

the paragraph/ subject's memory for both syntactic

changes and formal changes was minimal; at the same

time* they were still very good at detecting changes

in meaning.

The Dual Coding Theory of Memory:

The courts accept the hypothesis that vivid

images are better retrained than memories deviod of

interest. Research support for this common belief

has been provided by the theoretical and experimental

studies of Allan Paivio & others.

Paivio 1969, 1971 assumes that memory is based

on two distinct systems on imagery system and a verbal

system. The two systems are said to be functionally

independent but at the same time partially inter-

connected/the encoding storage, and retrieval of

information. They are differentiated by the quality

of symbolic information that they process. Imagery

is said to be specalized for the pictorial represen-

Page 16: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

16

tations of concrete objects, situations and scences,

while the verbal system is characterized by its ability

to deal with more abstract concepts such as language.

The functional distinctions between the two systems

include their manner of organizing information.

Imagery is distinguished for its spatial organization

of memory contents much as visual perception handles

objects in space, whereas the verbal system is specia-

lized for processing verbal information sequentially

or in a linear way.

For an illustration of this theory imagine that

a witness is asked to report his observations of being

held up. According to the dual coding approach the

witness may have had one or more of the following kinds

of sensory information hitting his recepterss Visual,

the physical appearance of the robber including facial

impressions: auditory, the verbal Commands.

It is assumed that these sensory inputs are

represented and processed by the two independent but

inter connected imaginal & verbal symbolic Codes.

The importance of such codings for remembering

is that multiple encoded information is likely to be

more available for retrieval (recall or recognition)

than information stored only in one code or the other.

Page 17: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

17

However, it is possible also to retrieve information

from one system without involving the other.

Although the availability of one or both Codes

often will facilitate recall, there is also the chance

of error when one symbolic code is transformed into

the other. This is particularly Asthana & Bedi (1982)

conducted an experiment with 24 undergraduaters to

investigate whether the amount of recognition failure

is significantly affected by the creteria for recogn-

ition judgement. The theoretical explanation of the

phenonemon of recognition failure is presented in

terms of the dual access theory of recognition, which

distinguishes between presentation code and

conceptual (retrieval) code.

Boles and David (1983) in the validation of

Paivio dual coding theory did experiment and has

observed that Paivio asserted that the representational

memory of words is indexed by familiarity and that

associated imagery is indexed by image ability and

concreteness, predicting that word recognition near

threshold will be influenced only by the former and

not by the latter atributes.

The Semantic Theory of Memory;

Accordingly, in this approach, man's memory

does not consist of word and images, but instead

Page 18: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

consists of codes connected to each other through

the mediation of propositions, or abstract descriptions.

This process is assumed to operate in the same way

that a computer uses abstract mathematical languages

to connect symbolic information. Memory is theorized

to consist of two parts, a data base representing a

person's skills & knowledge, and processes that operate

on this data base, such as encoding, storage & retrieval.

The data base consists of a net work of nodes or

concepts which are linked to each other by their relative ,

association. To illustrate this structure, presents

part of the semantic network proposed by Collin &

Quillian (1969).

Collin Sc Quillian* s semantic memory model is only

one of several models now available which describe the

meaning of memory through prepositional networks. These

models, in general, have received experimental support

confirming the hypothesis that semantic information is

organized in the form of a memory network.

Semantic networks of memory may also influence

memory for what is pragmatically implied or suggested as

if it were a direct assertion.

The finding we have just discussed led Alan Paivio

(1969, 1971, 1974) to propose a dual-coding theory of

memory. According to this view, long term memory contains

two qualitatively different but interconnected systems

Page 19: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

!>' 0

of representation, and is especially adapted for

handling seriallyordered information. The other is

an imagery system that is specialized for representing

spatial information. The two systems are connected

in that the individual can produce a label given a

picture & vice versa.

If it is assumed that information which is

represented in both Codes will be more likely to be

recalled than information which is contained in only

one of these codes, then the dual-coding theory can

account for both the finding that imagery instructions

aid recall & the finding that memorability is correlated

with imagery value. It is argued that imagery instru-

ctions lead the person to encode the word in both the

verbal and the imagery systems. Further more, words

rated high in imagery are likely to be encoded in

both systems, where as words low in imagery, value

are likely to be encoded in only one.

Facial Recognition:

The most powerful means to recognitfioaa us by qve

perceiving, storing and retrieving^aspects of facial

configuration. This may be due to the fact that the

face is the most salient feature of a person's being,

and indeed, quantification of various aspects of

social interactions have confirmed the notion the

Page 20: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

people look mainly at each others faces (Argyle, 1969).

While Harmon (1973) believes that determining how one

recognises a face is probably an intactable problem at

present and although Knapp (1972), suggests that

"the face is the researcher's nightmare", none the

less we will attempt to set out how the problem of face

identification has been approached and what progress

has been achieved.

Goldstein & Machenberg (1966) & Nash (1969) and

others found that the upper features of human faces

convey appreciably more information for recognition than

lower features, but is at oddswith Howells 1938. Fisher & mis)

Cox found that the eyes contributed most to identification,

but they were careful to stress that revealed features

below the level of the eyes were sufficiently informative

to allow recognition of famous people in 64% of cases.

Over and above specific feature saliency. Fisher &

Cox stressed features in combination because they found

that lower part of the face contribute approximately

15% more information when presented within the context

of upper features than when presented alone (i.e. with

no upper features of the face). A cogfigurational

aspect is also suggested by the fact that the addition

of eyes produces the greatest increase in recognition

suggesting that eyes are the most important facial

feature for recognition, but only 12% of subjects

Page 21: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

recognized the face when eyes were presented in

isolation. This suggests that the importance of eye

detail is enhanced considerably by the iimiediately

surrounding context.

Perceptual Gestalt for Faces:

Arnheim (1949) argued that the whole structure

of a face rather than the sum of its parts determines

expression. Yin (1969) noted that in facial recogni-

tion tasks his subjects attempted to get a general

impression of the whole picture rather than search

for some distinguishing feature and Tversky [I969j while

finding a suggestion of additivity of various features,

still maintained"that Gestalt effects are important.

Direct evidence for perceptual gestalt effects

has been given by Home et. al. (1976) . Their initial

orientating interest stemmed from work in verbal memory

and perception where it has been shown that the

perceptibility of any letter in a word is greater than

the perceptibility of that same letter shown in isola-

tion or the same letters scrambed to form unpronounce-

able words (Baron & Thurston, 1973; Reicher 1969).

Home et, al. wished to see if this word superiority

effect (W,S.E.) would generalize to the perception of

visual forms, particularly faces. Home et, al, found

that eyes & mouths were much better perceived than

Page 22: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

!' o 3

were noses when either normal or scrambled faces were

used, but not when single feature displays were

used.

Cross et. al. (1971) examined the role of

beauty in recognition and found that faces initially

rated high in beauty were subsequently better recog-

nized. However, a subjective impression on the part

of Cross et. al. was that the good-looking faces

(which eventually were better recalled) were actually

processed for a shorter time than the other faces.

Sorce & Campos (1974) served to consolidate the

importance of expression change for recognition by

showing that the greater the difference between a face's

expression on initial viewing and later recognition

testing the poorer was recognition performance.

Ellis et. al. (1975) who used two photographs

suggested that there were individual differences among

faces in the extent to which eyes or mouth were

important for processing and remembering. McKelvie

showed th^t masking the eyes always caused greater

errors in recognition than masking the mouth.

Individual & Group Differences in Person Recognition.

EUis, et. al. (1973) tested 12 & 17 years old

subjects and found the latter group to be much better

Page 23: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

recognizing faces of young adults. One study,

which at first sight appears to deny this relation-

ship, is one by Cross, et. al, (1971) where they

looked for an improvement in face recognition with

age but failed to find it in subjects aged 7, 12 or

17 years adults and found out the results recognition

performance from worst to best was 7, 12, 17 years and

then adults.

Sex Differences In Facial Memory:

The general finding is that females are better

than males at recognizing faces they have previously

Seen. McKelvie (1972) found no sex differences in

recognition of Schemantic faces, and Howells (1938)

also found no difference with photographs. These

two studies however are somewhat against the tide

because Goldstein & Chance (1971) did find a differ-

ence between men and women observes when they presented

photographs of faces snowflakes and inkblots for

later recognition. The females were better than

males, that only for faces. Cross et. al. (1971)

failed to show an overall female superiority with both

male and female photographs as stimuli. However, the

female observes performed significantly better than

male observers on female faces, with the male observers

being equally good on the male and female faces.

Page 24: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

Witryol and Kaess (1957) did show women to be superior

overall compared to men but also found that male

observers were better with male faces and female observers

were better with female faces. In similar a way Ellis

et. al, (1973) found that while there was no significant

difference between male and female observers for male

faces* there was a significant difference for female

faces, with women performing better. Studies by both

McCall, et. al. (1974) and Mazanec and McCall (1975)

show female superiority in observational accuracy on

a number of tests, but again they are concerned to

stress the "Same Sex Effect".

As regards sex, women generally show higher

accuracy scores on female than male faces whereas these

trends are often reverse in male subjects.

While the research literature on sex differences

in recognition memory of faces is not entirely consis-

tent, it suggests that females are superior to males,

particularly on female faces. Most studies reporting

a complete analysis of their sex data (Borges and Vaughn^,

1977 Cross et. al. 1971, Shephest et. al. 1974,

Feinman & Entwistle, 1976; Going and Read 1974,

Stefanatos, et al. 1978, Yarmey, 1975) have shown no

overall significant difference between male and female

recognition accuracy, although some (Ellis, et. al.l973.

Page 25: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

Yarmey, 1974, Yarmey and Pasharuk, 1975) favour females

over males. However, six of these 10 studies yielded

a significant interaction between sex of subject and

sex in face. In four cases (Gross et. al. 1971;

Yarmey and Pasharuk 1975) females recognized more female

than male faces, whereas males performed equally well

on both types, and in the remaining two (Ellis et. a^,

1973; Stefanatos et. al., 1978) males recognized fewer

female than male faces, whereas female performance was

not significantly different.

Although other investigators (e.g. Ellis 1975)

have also reached the conclusion that females, relative

to males, recognize female better than male faces, the

effect is by no means clearly established (four out of

the 10 studies cited above found no significance sex

interaction).

Personality Differences: Introverts it. Extroverts.

These terms were first used by Jung but have been

given major explanator value by H.J. Eysenck. Broadly

speaking, extroversion refers to the kind of behaviour

which is out going, or outwardly oriented. The person

is highly aware of what is going on around him and he

relates to outside objects or people. The introverted

personality on the other hand is inward looking, orienta-

tion is towards self, and the understanding of one's own

experiences.

Page 26: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

•?6

Gilliland and Burke (1926) presented faces for

immediate recall and found that extroverts were signi-

ficantly better than introverts. Hunt (1928) found a

correlation of +O.Ss between extroversion and memory

for names and faces (which was a subtest of a much

larger test. Huckabee (1974), who showed that it was

introverts who reported greater imagery to concrete

words rather than extroverts the possibility of

differential imagery ability in extroverts and intro-

verts and its role in memory is clearly in need of

much greater study. Gale et. al. (1972) reported a

significant correlation between performance on a version

of the Bettee vividness of imagery scale and extroverts.

Extroverts reported more vivid imagery than did intro-

verts,,. and this relationship was supported by Morris

and Gale (1974). As we have argued above, individual

differences in vividness of imagery correlate with

performance in memory tasks (e.g. Marks 1972) and thus

it is possible that the short term superiority of extro-

verts is mediated by imagery coding. The problem here

is that the rated amount of imagery evoked by separate

words does not correlate with extroversion.

Intelligence & Education:

Except for an early study by Howells 1938 and two

more recent investigations by Peinman & Entussle (1976)

Page 27: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

and by Kaess & Witryol (1955), little empirical infor-

mation is avilable on the relation between intelligence

and/or educational background and facial identification.

Howells found that face recognition correlated with

intelligence score ( r = .27, n = 91) and with grades

( r = .33, n = 112). Obviously, these correlations

are low and account for very little of the variance

in recognition performance. Similarly, both Kaess &

Witryol (1955), in an investigation using adults and

subjects, and Feinman & Entwisle (1976), who used

children as subjects found negligible correlations

between 1 Q and facial recognition ability.

Memory for Faces and Names:

Few people have gone to a party or class reunion

and not heart at least one person say, "I can remember

your faces but I can not remember your name". This

experience illustrates one again the common finding

that recognition memory is usually superior to recall

memory (Mac Dougall 1904, Postman (1950) .

LTM for names is related to our knowledge about

that person (Yarmey 1973) e.g. We know that celebrities

are associated with specific contexts ~ the arts, poli-

tics, sports etc, - and when we try to remember their

names, we search over memory for certain situational

cues which help to characterize these persons, A

subject trying to remember the name of a particular

Page 28: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

ioO

celebrity may remember first that he is a movie star,

may then remember the name of the movie he was last

seen in imagine scenes from the film and so on. Imagery

is a useful mechanism for remembering names and count-

less other information in our everyday world (see

lorayne and Lucas 1976; Jones and Hampson 1978,

Yarmery 1970).

In an investigation of the differences between

recognition and recall of faces and names^ Clarke(1934)

found that recognition was superior to recall for both

faces and names but names were slightly easier to been

recognize (97%) than were faces (91%). I-tha;̂ ® repli-

cated the greater accuracy in recognizing names than

faces (Yarmey 1970). This result is interesting since

it tradicts the general conclusion that recognition

memory for picture of objects is superior to the recog-

nition of their word labels (Jenkins, et. al. 1967).

Whether or not this result is another indication that

memory for faces is unique and different from other

types of visual memory is debatable (Yin 1969).

More likely evidence showing that names are easier to

identify than faces occurs because names used in

research studies are common and familiar stimuli.

Carm-icheal, et. a. (1932) showed that word

labels influence the perception and reproduction of

stimulus objects. If an object is presented to a

Page 29: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

subject and is called by a particular name, the

subject tends to reproduce (draw that object in a

manner that is consistent with all objects given that

particular label.

Recognition memory for photographs of faces is

reliably better twhen position than in an inverted

orientation (Brooks and Goldstein, 1963; Hochberg

and Galper 1967; Yin 1969). This finding is not

surprising and is not new. Kohler (1940), several

decates ago, accounted for this difference by attri-

buting it to the loss of facial expression in the

inverted photograph normally, all of these stimuli

including faces are viewed in an upright orientation,

Howells (1938) pointed out that subjects who

were superior at details of facial photographs

were not superior at recognize them. Malpass et. al.

(1973) found that training in giving verbal descrip-

tions of faces did not improve visual recognition

performance. McKelvie (1976) investigated the effects

of labelling on the encoding and recognition of

Schematic faces. He presented easy to -label and

hard-to-label schematic faces which a previous study

(McKelview, 1973) had provided.

Chance and Goldstein (1976) examined whatever

use of two kind of self generated verbal labels for

Page 30: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

'Of)

faces was related to the accuracy of recognizing

these faces one week later*

McKelvie (1976) showed that labels have their

main effect by focusing attention on specific facial

features during viening and that the labels may be

stored along with the visual trace to be used at

output. Also important is the fact that some of the

results show that labels will only be effective if

actively generated by the subject - a finding which

we have argued to have important theoretical

implications. Yarmey A.D. (1970) 96 undergraduates

(4 experimental and 2 central groups) learned 2 lists

of 10 pairs of number - faces or number names.

Experimental Ss learned the 2nd list under a mnemonic

rhyme condition. 1/2 of the experimental Ss used

imagery as a mediator while the remainder were given

verbal mediation sets.

Names were recognized better than faces, mnemonics

improved recognition scores and recognition of faces

but not of names was facilitated by the mnemonics

control group data show that results are not attri-

butable to practice effects.

Klimesch Wolfgang (1982) examined whether

pictures and words activated identical semantic

information in a common memory system or whether the

semantic encoding of pictures and words took place in

Page 31: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

two separate memory systems. A third possibility

was that the semantic encoding of pictures and words

took place in a single memory system, but was enacted in

different ways.

Findings show that pictures and words do not

activate identical semantic information and that

the semantic encoding of pictures and words do not

take place in two different memory systems. There

was a strong interdependence between the encoding

of pictures and of words. When two items were shown

successively with an interval of 1.6 sec, equally

strong semantic relatedness effects were obtained.

Thus, the semantic encoding of pictures and words

takes place in the same memory system but is enacted

in different ways,

Clarke & Morton (1983) studied word recognition

and have shown no effect on subsequent visual word

recognition of naming a picture.

Joseph and Stone (1984) conducted a study

attempted to describe fully in words the picture

stimuli that were presented, bectuse a picture simple

word label obviously is not equivalent in complexity

to the picture itself. Pictures^their simple labels,

and a series of words that completely described the

pictures were presented to 90 undergraduates.

Page 32: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

Results indicate that stimuli in the picture form

were remembered better than both types of word

stimuli and support the double encoding

hypothesis.

McCorty D.f, (1980) conducted study with

college Ss on a face - name mnemonic containing

several components: a prominent facial feature;

a concrete high imagery transformation of the

person's name.

Ss were given one of strategies for learning

face - name associations^ the strategies differed

with respect to which on the 3 or combination of

the 3 components were incorporated.

Results indicate that all 3 components of

the face - name mnemonic were essential for its

effectiveness.

Page 33: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

< • /

B. Problem And Hypothesis

It is quite clear from the first chapter that

most of the researches have been conducted in recog-

nition memory related to the single code or the

double code of human information processing mechanism.

Upto mid 1960s it was believed that there was a single

code i.e. the verbal code, which was responsible for

both the verbal and the non-verbal material. The

dual coding theory proposed by Allan Paivio (1975)

state that there are two types of codes in our memory

system, they are verbal and hon-verbal. These two

codes are independent as well as inter-related to

each other.

The dual coding theory was criticized by Banks

and Flora (1978) finding the verbal codes are having

pictorial information similarly the nonverbal codes

are having some verbal information. Allan Paivio

proposed orthogonality hypothesis for dual coding

theory which is supported by te Linde (1982),

The present investigator formulated the problem

in order to extend the dual coding theory and to assess

its generalizability in the context of memory for

names and faces.

Page 34: Chapter i A. Introductio And Historica Resuml e n The ...shodhganga.inflibnet.ac.in/bitstream/10603/44034/6/06_chapter 1.pdf · zation is memory for music and musica notesl Althoug

Problem:

Validation of Dual Coding theory in Recog-

nition Memory for names and faces into four conditions

(i.e. name - name)# (picture - picture), (picture +

name - name) and (picture + name - picture). The

recognition memory would be less when only name or

picture is recognized, by presenting the name or

picture along (in a single manner) in comparison to

the presentation of both (dual) (name with picture).

The following hypotheses were made.

(a) The Dual presentation of name and picture would

be better recognized than the single presentation

of name or picture. In elaborative terms, the

memory of faces will be facilitated with

additional information of names and otherwise.

(b) There would be no difference between the four

presentation conditions (i.e. name - name;

picture - picture; picture + name - name;

picture + name - picture).