the referent of children's early songs
TRANSCRIPT
This article was downloaded by: [Ams/Girona*barri Lib]On: 27 November 2014, At: 00:26Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Music Education ResearchPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/cmue20
The referent of children's early songsEsther Manga Hong Kong Baptist University , Hong KongPublished online: 23 Jan 2007.
To cite this article: Esther Mang (2005) The referent of children's early songs, Music EducationResearch, 7:1, 3-20, DOI: 10.1080/14613800500041796
To link to this article: http://dx.doi.org/10.1080/14613800500041796
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoeveror howsoever caused arising directly or indirectly in connection with, in relation to orarising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions
The referent of children’s early songs
Esther Mang*Hong Kong Baptist University, Hong Kong
Musical creativity during early childhood is readily exemplified in vocal behaviours. This paper is a
discussion of observations on children’s performance of learned songs and self-generated songs.
Longitudinal observations suggest that self-generated songs may be seen as referent-guided
improvisation using source materials derived from learned songs. It is hypothesised that
interactions with the environment, such as a need for the child to communicate emotions in
play settings, could autonomously abstract novel sensory input into existing categories or distort
aspects of existing ones to assume novel parametric combinations. Hence, using learned songs as
referent, a child possesses a repertoire of learned musical features that allows her to create and
improvise as a reaction to environmental stimuli. The resultant early songs, therefore, exhibit a
distinctive sense of ownership endowed with rich creative instincts.
Introduction
Young children are often found actively engaged in vocalising, singing and moving to
musical stimuli. As early as 18 months, children may begin singing spontaneously,
and at around the age of 2 years, they may gradually incorporate snatches of learned
songs into their self-generated songs. Early songs of children are marked by an
improvisatory character and resemblance to variants of songs modelled after those
taught to them. Since the array of configurations in early songs is abundant, analyses
of the relationships between the self-generated songs and those that were taught to a
child could provide adults a glimpse of her musical world. Hence, observations of a
young child’s early songs may help deepen understanding of early childhood musical
creativity encoded in vocal behaviours. The present paper is an exploratory
discussion of some observed affiliation between young children’s learned song
achievement and the multiplicity of their self-generated song performance.
Speech, songs and vocalisations
Researchers have described different strains of early songs according to their musical
characteristics and the context in which they were observed. The pioneering study
conducted by Moorhead and Pond in 1941 reported spontaneous vocalisations that
were ‘unfettered and free rhythmically, like plainsong’ and another type of ‘chant’,
*Department of Music and Fine Arts, Hong Kong Baptist University, Kowloon Tong, Hong Kong.
Email: [email protected]
ISSN 1461-3808 (print)/ISSN 1469-9893 (online)/05/010003-18
# 2005 Taylor & Francis Group Ltd
DOI: 10.1080/14613800500041796
Music Education ResearchVol. 7, No. 1, March 2005, pp. 3�/20
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
which was ‘rhythmic, of limited compass and repetitious’ (1978, p. 8). According to
the researchers, those spontaneous songs of young children were often narrow in
range and not diatonic; they were usually uttered in undulating contours without a
definite tonal centre; and had flexible but occasionally patterned rhythm. Because
those songs are not confined to the Western musical idiom, Moorhead and Pond
described them as unpredictable and unmemorable. It should be noted that the
researchers excluded ‘fragmentary songs’ in their description of spontaneous songs
(Moorhead & Pond, 1978, p. 9).
A similar emphasis in rhythmic interest was reported by Fujita (1990), who
identified a prevalent form of spontaneous vocalisation as ‘intermediate performance
between talking and singing’, in which the phrase or words were uttered metrically,
or melodically with some fixed pitch (p. 146). As found in previous studies,
rhythmical talking was observed during free play, produced when children wished
to communicate with others (ages 4 or 5), while 2�3-year-olds talked rhythmically
for their own satisfaction rather than for communication (p. 143). Fujita pointed out
that much of the rhythmical talking was accompanied with movement and bore a
close affiliation to the inflection of the Japanese language and its culture as a whole.
Fujita maintained that ‘the children refined their musical performance by them-
selves, conforming to the form of musical expression, and never created a musical
performance which exceed their cultural context’ (p. 142).
Contrary to the emphasis on rhythmic interest as reported in the literature, Ries
(1987) reported that the tonal centre was a prominent feature in the spontaneous
songs that she observed and collected. Ries regarded spontaneous songs as ‘bits of
singing that stood out from the surrounding material because of their cohesiveness
and usually slightly greater length. These renditions were consistent in mood and
always maintained a tonal center, even at 7 months. They seemed to be more than
just a reply to the stimulus or prompt and appear to be a special musical effort’
(p. 15).
Moog (1976) provides an extensive analysis of spontaneous songs as they emerge
at different age level during childhood. At ages 2�4, spontaneous singing dominated
the musical activities of children. Moog identified three different forms*/
‘imaginative’, ‘narrative’ and ‘pot-pourri’ songs. Imaginative songs form only a small
repertoire of spontaneous songs ; they bore no resemblance to a known song, and were
often hummed or sung to a single syllable. Narrative songs seemed not to be intended
for an audience, although they often told stories. They were sung monologues
composed mainly of nonsense snatches of words and tunes, and occasionally
included excerpts of learned songs. Pot-pourri songs were combinations of learned
songs with mixed up words and melodies, and they might contain original
improvisations.
It is noteworthy that Moog observed that children adapted the same formal design
from their learned songs into songs that they created. Interestingly, for children who
did not demonstrate singing any learned songs in the study, no original song was
observed.
4 E. Mang
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
Young (2002) documented children singing during free-play and analysed
spontaneous vocalisation according to its context and acoustical characteristics.
Young observed that some known songs resurfaced in child’s free-flowing singing as
discrete fragments in dramatic play. This form of vocalisation was labelled as
reworkings of known songs, which among her six categories of spontaneous vocalisa-
tion, relates most closely to Moog’s conception of spontaneous songs. Similarly, the
contexts of the spontaneous vocalisations collected by Young were dominated by
vocal play triggered by verbalisations in role-play situations.
Spontaneous songs are also regarded as a key building block in song acquisition.
Davidson et al . (1981) proposed that a spontaneous song transforms musical events
from the environment (assimilation) to fit into a scheme, made up of the properties
of songs in the culture, of which the child has already developed (accommodation)
(p. 303). As the researchers explained:
In many child-versions of standard songs, those bits that most closely approximate the
model appear at the same time as identical melodic structures in the child’s spontaneous
repertoire . . . Now the child can pick and choose from among his repertoire of fragments
and bit those most appropriate to a certain spot in a certain song. This ability to handle
one’s own repertoire with increasing flexibility so that it can lead to the production a
number of standard tunes leads us to characterize the child’s competence by the age of
three or so*/as the mastery of the outline song. The capacity to master a number of basic
fragments, and then to compose them into a reasonable approximation of a target
model, is by no means restricted to the child’s progress in the musical sphere. Indeed,
this kind of ‘outline knowledge’ of a symbolic domain seems quite characteristics of the
child around the age of three. (p. 306)
The abundant documentation of young children’s early songs brings about an array
of diverse viewpoints. In summary, researchers have consistently raised issues such as
singing versus speaking, and performance of learned songs versus self-generated,
improvisatory songs.
Data collected form longitudinal observations suggest that young children’s
speech and song acquisition are developmental behaviours that exhibit phases of
mutations. Mang (2001) proposed that young children undergo a three-stage
development, whereby children under 2 years old demonstrate acoustical intermediate
vocalisations , which facilitate their initial acquisition of vocal control. By the age of
2 years, children are capable of making productive distinctions between singing and
speaking, and by 3½ years, children may use novel forms of vocalisations freely to
facilitate expression. This form of vocalisation is observed after children had
mastered the skill needed to perform songs accurately and communicate effectively
in speech but purposefully alternate easily between singing and speaking to
communicate in novel forms of contextual intermediate vocalisation (Mang, 2002).
Referent in improvisation
The present paper is a discussion of the forms of early songs collected longitudinally
between the ages of 2 and 4 when children already possess a conception of song and
The referent of children’s early songs 5
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
begin to acquire fundamental singing skill. Published findings concur that some early
songs at this age level appear to be distant mutations of song material familiar to the
child; others are noticeably reproductions of learned songs with impetuous
adjustments. Following this line of inquiry, it seems pertinent to delineate possible
relationship between young children’s learned and self-generated songs. One possible
approach to examining early songs is to contemplate the notion of ‘referent’ in
improvisation. Pressing (1984) explains that a ‘referent’ has direct ramifications for
the resultant improvisation.
Central to improvisation is the notion of the ‘referent’. The referent is an underlying
formal scheme or guiding image specific to a given piece, used by the improviser to
facilitate the generation and editing of improvised behaviour on an intermediate time
scale. The generation of behaviour on a fast time scale is primarily determined by
previous training and is not very piece-specific. If no referent is present, or if it is devised
in real-time, we speak of ‘free’ or ‘absolute’ improvisation. This is much rarer than
referent-guided, or ‘relative’ improvisation. (p. 346)
In the case of children’s self-generated songs, their referent is often learned songs.
Key features of a learned song, such as repeated lyrics or melodic motif, are often
reconfigured and integrated into another song. As Pressing (1984) explains, ‘the
referent typically functions either as a source for material, which is then repeated,
transformed, varied or developed, or as a focus for the production and organization
of material from other sources’ (p. 347).
Hence, children’s self-generated songs could be seen as referent-guided improvisa-
tions because the source materials are often derived from learned songs. According to
the Pressing (1988) model, the source materials (i.e. in Pressing’s terminology,
object, features and process array types) are basically fixed, but innate ecological
capacity could autonomously abstract novel sensory input into existing categories or
distort aspects of existing ones to assume novel parametric combinations. Hence,
interaction with the environment could result in aggregation of memory constituents
to extract or create new cognitive assemblages. In the case of children’s early songs,
a strong affiliation between learned songs and self-generated songs might be
inferred from Pressing’s model (Figure 1). Since a child could manipulate the
source material by freely combining selected aspects of different learned songs, early
songs display different levels of novelty. Hence, researchers use labels such as
‘narrative song’, ‘musical monologue’, ‘pot-pourri song’, ‘imaginative song’ and
‘private song’, to depict the multiplicity of children’s early songs (Moog, 1976;
Papousek & Papousek, 1981). Applying the principals of the Pressing (1988) model,
the array of self-generated songs might well illustrate the richness of novel sensory
inputs.
The purpose of this article is to report and discuss data concerning children’s early
songs collected in a longitudinal study on young children’s vocal development.
Specifically, it explores the notion of children’s use of learned songs as the referent
for their self-generated songs. Since self-generated, improvisatory singing responses
are most prominently observed between ages 2 and 4, findings and discussions are
6 E. Mang
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
limited to this age phase, although more comprehensive investigation was conducted
in the original study. This article focuses on the examination of musical and
singing achievement exhibited in the performance of learned songs and analysis of
how children adapted this to self-generated songs. It is therefore inferred that
analysing the performance of learned and self-generated songs will facilitate
understanding of early childhood musical creativity as it is encoded in vocal
behaviours.
Method
The sample and subjects
Subjects were recruited through already established personal contacts. The sample
consisted of only girls because no parent with a male child expressed interest in the
study. Seven of the eight girl participants were aged 2�4 years at the start of the
study. (See Table 1 for more information about the subjects.) Six of the subjects had
had some informal musical training with the investigator before the study
commenced. Throughout the 42 months of the study, contact was maintained
with the parents to update them on progress made by each child and to obtain more
in-depth views about consistency and changes their child’s vocal behaviours.
Ongoing confirmation and synthesis of data allowed the investigator to compare
and corroborate observations of each child.
To obtain information about the type of songs the children had acquired, they were
asked to sing their favourite songs and games. It seemed that all children had
acquired traditional nursery rhymes and popular North American children’s songs
such as: Eency Weency Spider; The Wheels on the Bus; Bingo; It’s Raining, It’s Pouring;
Pop Goes the Weasel; and This Old Man . It was not clear whether children had
acquired these songs at their daycare centres or elsewhere; nevertheless, their song
performances could provide an overview of their repertoires.
“Referent”
Decompose andreconfigure
key features oflearned songs
Learned Songs
Songs recallingpreviouslyintroduced songs
Referent-guided Self-generated Songs
(e.g., “narrative song”,“musical monologue”,“pot-pourri song”,“imaginative song”,“private song”)
Novel sensory input
Novel sensory input
Figure 1. Affiliation between learned songs and self-generated songs
The referent of children’s early songs 7
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
The role of the researcher
Six of the subjects had attended a music programme offered by the investigator for
eight or 16 45-minute sessions before the study commenced. These children had
joined the programme at some time between the ages of 18 and 20 months. They
were invited to participate in the study only after the child had left the music
programme. The types of musical experiences those children had, before the study
commenced, included movement, listening and playing instruments.
For the two other subjects who did not attend the music programme, the
investigator was a family friend. For the present study, the investigator assumed the
role of an observer-participant and was treated as a visiting family friend.
Data collection
Each child was visited every 4�6 months over 42 months. A typical home visit
consisted of approximately 45 minutes to tape the child speaking and singing, and at
least 15 minutes of an informal conversation interview with the parents.
All data collection took place at each subject’s home. Whenever possible, the child
was asked to sit on the carpet in her bedroom where she was surrounded by familiar
things to ensure that interaction could take place in a comfortable manner.
Sometimes, data were collected in the living room, where the child would be asked
to bring along some favourite toys and books.
The parents sometimes sat beside or within several feet of the child to show
support. The mothers would typically assume the role of an observer at the beginning
of the recording session, but would soon join as a participant. Other parents
participated actively throughout the recording session, encouraging their children to
sing and speak. Explanations and reminders were given to parents to avoid
prompting the child to sing.
Table 1. Some basic information about the subjects
Name of subject Age (at start of the study) First spoken language Place of birth
Heidi 18 months Mandarin (English) Taiwan
Clare 25 months English Canada
Vicki 27 months Cantonese (English) Hong Kong
Jodi 29 months Mandarin (English) Canada
Wendy 31 months Cantonese (English) Hong Kong
Amber 33 months English Canada
Gina 34 months English Canada
Polly 38 months English (Spanish)a Peru
The language given in parentheses is the secondary language spoken by a child at home.aBoth of her parents are native speakers of Spanish from Peru. English is the primary language
spoken at home, as the mother explained, because the child became more reluctant to
communicate in Spanish after they moved to Canada when she was 2 years old.
8 E. Mang
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
Vocal responses were recorded during child�adult interactions and in various
play settings. During each visit, the children were asked to sing favourite songs,
regardless of the language they preferred. Seldom would a child refuse to sing, but
would at times find it difficult to choose a song without assistance. It appeared more
natural to first engage the child in a conversation, then have the parents gradually
encourage the child to ‘teach me’ her favourite song by performing it. All vocal
responses were recorded with a portable professional standard cassette recorder
(Marantz PMD 201) using a unidirectional professional microphone (Shure, SM48-
LC). ‘Digital accuracy’ cassette tapes (TDK SA-X) were used to store the vocal
signals.
Descriptive notes were written at the end of each visit to record the setting and the
observed activities. The following observations were made on the context of the
children’s vocalisations: (i) the source of stimulus (a storybook, an invitation to sing,
etc.) versus the type of vocal response; (ii) other corresponding behaviours (body
movements, use of toys, etc.) observed when the vocalisations occurred; (iii) the
child’s communication intention as inferred from the scenario when the vocalisation
occurred (items i and ii) or verification with the child; and (iv) the parents’
interpretation of the vocal responses.
I usually chatted with the parents after each recording session to facilitate data
interpretation. Since a more literal description would provide a relatively precise
account of the observations, some of the conversations were also taped to enable
citing the exact wording used by parents.
Data analyses
Three judges, who were graduates of music and had vocal training, analysed the
singing responses. Copies of the recorded singing responses were given to each judge,
allowing the tapes to be replayed as many times as needed. The judges were asked to
categorise singing responses into learned song performance, self-generated songs, or
other novel forms of vocalisations.
The judges then provided perceptual analyses of the learned songs to evaluate the
child’s musical and singing competence. Analysis of children’s achievement in
learned song is important because memory serves as a repository for the referent.
Hence, skill achievement of the child as she demonstrated in recalling familiar songs,
could be taken as the readily available resources that were at the disposal of the child
for self-generated songs.
The judges were also asked to examine how the child used learned songs as
referent to organise elemental structures of her early songs. Hence, musical elements
such as cadential material, melodic and rhythmic motifs, repetition and contrast, and
arrangement of words in early songs were the key features under scrutiny. The
following is a discussion of the best examples that captured the typical behaviours
observed in the age levels concerned.
The referent of children’s early songs 9
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
Findings
Ages 2�3
Performance of learned songs. At the 2�3-year level, about one-third of the children’s
song repertoire seemed to be modelled after songs previously introduced to them.
However, those songs were seldom rendered in their complete form and omission of
parts was frequent. The words of the songs were also simplified; some key words of
the song remained intact but other words were freely replaced with nonsense
syllables or words that appeared to fit the sound. Frequently, only short segments of
melody were intact, while other parts of the song were, according to the child,
forgotten. However, it appeared that the child had forgotten only the words but not
the overall musical structure of the song.
The relationships between performance of words and melody were particularly
interesting. When a child said that she had forgotten how to sing a song, she was
often able to sing the learned song accurately after being prompted with the words.
However, a child often failed to ‘remember’ a song even after being prompted with
the melody alone on monosyllables such as ‘doo’, or simply hummed the melody.
This suggests that the words (or specifically, the sounds that made up the words
because they could understand very little about the meaning of the words) appeared
to operate as an ‘anchor’ that secured the musical elements of a song. Hence, when
words were forgotten, some children considered that the song as a whole was
forgotten. Nevertheless, some children at this age were able to render a learned song
without perfect recall of all the words. For example, when Wendy performed a
learned song, she managed to accurately render all the words in the first verse, but
failed to sing the melody accurately. After singing the first verse, she paused briefly
before continuing with the second verse; that is, she appeared to have forgotten the
latter part of the words in the second verse. However, without noticeable
interruption, she substituted nonsense syllables, ‘da-dee-da-dee’ for the forgotten
portion, and continued to perform the song. Evidence suggests that although a child
at the 2�3 level possesses the ability to perform a learned song without accurate recall
of all the words, she may sometimes consider the entire song forgotten when only
some of the words are forgotten.
The word meanings of a learned song often appeared to be more attractive to a
child when sung with body movement. For example, when Clare was invited to sing
an action song, Teddy Bear, she acted out all the body movements corresponding to
the correct sequence of the song but did not attempt to sing the song at any time.
Since the whole song was performed as if it was a ‘silent movie’, the body movements
were acted out in their correct sequence, possibly because Clare was mentally
rehearsing the song from the beginning to the end. When Clare performed Teddy
Bear , I also noticed that her movements were acted out in accordance with the
appropriate melodic rhythm of the words. Hence, Clare did not simply remember the
content of the song, she acted it out with body movements; that is, she ‘expressed’
the song without actually singing it aloud. However, this is a complex example of
dissociation, the ramifications of which are not explored in this paper.
10 E. Mang
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
The learned song was sometimes expressed through tactile experience. For
example, just shortly after Clare began singing Mary had a Little Lamb , she stopped
abruptly and walked away to get her toys. She appeared to be anxious and asked her
mother to help her find ‘Mary’ and ‘Mary’s lamb’. With the mother’s assistance,
Clare found these two characters and returned to the sofa with her toys. She
completed singing the first verse of the song accurately and then attempted to sing
the second verse. However, she was soon absorbed in manipulating ‘Mary’ and her
‘lamb’ to act out the rest of the song and never completed singing the second verse. It
therefore appears possible that for a young child, singing may not be an isolated form
of experience that involves only vocal production; singing is sometimes integrated
with, or even contrasted from other forms of expression such as body movement and
imaginative play.
Performance of self-generated songs. Early songs by children at the 2�3 level took on a
range of different configurations such as ‘pot-pourri songs’ and ‘vocal babbling’ as
defined by Moog (1976). The children appeared to take pleasure in using their voices
to manipulate learned materials and explore new resources for sounds. For a ‘pot-
pourri’ song, it was often difficult to differentiate whether a child was intentionally
manipulating the musical materials from more than one song or if it was simply a
case of confusion over two songs. This difficulty is illustrated in the following
scenarios.
Scenario 1: At age 2 years 3 months, Clare sang the first phrase of Bingo but
shifted to sing the first phrase of Old McDonald without a pause. After that, she
hesitated for a while as though she had forgotten something or made a mistake. Clare
paused for a moment but did not resume singing either of the two songs. It is possible
that there was some confusion over the two songs, i.e. she ambiguously connected
the final word ‘Oh’, from the first phrase of Bingo to the first word ‘Old’, as in Old
McDonald . This interesting scenario suggests that words often serve as an anchor
securing the musical elements of a learned song.
Scenario 2: At age 2 years 8 months, Wendy sang the first three phrases of Happy
Birthday in English. She was later distracted from the words when pretending to
blow out candles, consequently she did not sing the last phrase of the song. However,
since all judges agreed that the first three phrases were sung in tune, it appeared that
Wendy knew the song well. Wendy then continued to recite a Cantonese nursery
rhyme and sang two Cantonese songs. While Wendy sang the second song, It’s a
Small World (a version with words translated into Cantonese), she paused after
singing the first phrase when she noticed her father had just returned home. Wendy
subsequently continued to sing but modified the words to fit into the melody of
Happy Birthday. She then gradually returned to the melody of It’s a Small World and
continued to sing the refrain. Thereafter, although the tonality was successfully
maintained, fragments of the two songs were chained together. It is not clear whether
Wendy was confused over the melodies of the two songs, whether she was
intentionally mixing the two learned songs , or whether she had forgotten some of
the words of both songs. This ‘pot-pourri’ song demonstrates how a child might
The referent of children’s early songs 11
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
combine melodic material from two learned songs. Since earlier in the same
recording session, Wendy sang almost the entire Happy Birthday song accurately and
in tune, it is unlikely that she was confusing the two songs. Because the two songs
were sung in two different languages (Happy Birthday in English and It’s a Small
World in Cantonese), I speculate that it is unlikely that Wendy confused the two
songs as one.
In summary, inspirations for most of the self-generated songs recorded from
children at the 2�3 level appeared to stem from fragments of learned songs. Children
usually began singing a learned song with repetition of a single syllable or nonsense
words before they sang the rest of the song. Reference to the diatonic system and
metrical rhythm would often loosen when a spontaneous early song began.
Ages 3�4
Performance of learned songs. There was a marked increase in the number of learned
songs collected from children at the 3�4 level. Roughly half of the singing responses
collected in this age level were attempts to reproduce familiar songs that the children
had acquired. Some of the favourite learned songs collected from all children in this
age level were: Bingo; Teddy Bear; Twinkle, Twinkle, Little Star; Old McDonald; I Love
You; Jingle Bells; and The Alphabet Song . When compared with data collected earlier,
most children at this age level were able to sing phrases of learned songs with greater
accuracy in melodic and rhythmic patterns. The general contours of the repetitive
patterns in the songs, as well as most of the narrower melodic intervals were often
sung correctly. Judges were therefore able to identify successfully almost all learned
songs sung by the children, even when they were incomplete.
Maintaining tonality of a song was a problem for all children. Even when able to
demonstrate a sense of tonality at the beginning of a song they often changed the
tonality several times within a song. In traditional Western music, most melodies
contain a clear reference to a defined tonal centre. That children modulated during
the course of singing a song even after they had established a tonal centre, suggests
that they had yet to learn how to retain the primacy of an established key as a
function of musical memory.
The words of learned songs appeared to be a challenge for children at the 3�4
level. They sometimes expressed frustration about forgotten words, which made it
difficult for them to complete a song. Even when a child could successfully recall a
melody but not the words, she would frequently give up singing the song. At this age
level, only occasionally would a child substitute forgotten words with nonsense words
or syllables, and continue singing. Children at this age level often failed to produce
the precise structure of learned songs, thus the order of verse and chorus was
frequently confused.
At age 3 years 4 months, Gina wanted to sing The Whale Song . She began by
humming some melodic fragments of the song. Unfortunately, she failed to recall any
of the words and soon gave up singing the song. Similarly, at age 3 years 6 months,
12 E. Mang
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
Wendy sang the first phrase of a learned song but forgot the rest of the words. She
then repeated the first phrase three times before she eventually gave up singing.
When a child claimed that she forgot the words of a learned song, melodic
accuracy suffered. At 3 years 10 months, Gina successfully sang the first verse of
Baby Beluga in tune. However, she appeared to have forgotten some of the words in
the second verse, from which she sang fragments, and then hummed the melody as
she tried to recall the words she was missing. The humming later faded away as she
tried harder to recall the words. Eventually, she chanted fragments of the words
without attempting to sing the melody. It appeared that Gina had shifted her
attention to recalling the words to such an extent that the melody could no longer be
recalled accurately. This issue is addressed in studies on text�melody integration in
melody recall (see, e.g., Serafine et al ., 1984, 1986).
Performance of self-generated songs. Scenario 1: Heidi’s favourite activity was
drawing. She had an easel, newsprint paper and crayons in her room, so that she
could draw whenever she wanted.
After Heidi’s mother showed me some of her daughter’s drawings displayed on the
walls, Heidi was enthusiastic and said she could draw some eggs for me. She walked
over to the easel, picked up a crayon and announced a title for her drawing. With a
loud and clearly projected voice, Heidi enunciated ‘EGGS. There is a nest . . .’ She
then drew an egg in the centre of the newsprint paper, and sang, ‘and inside the nest,
there is an egg.’ Heidi chose another crayon and drew a nest around the egg. She then
sang the final two phrases of a learned song, The Green Grass Grew All Around .
Heidi moved on to draw a second egg, using another colour. As she drew, she said,
‘There was (were) two eggs, sitting in a nest’. Repeating what she had done before,
she then sang the final two phrases of The Green Grass Grew All Around . The ‘drawing
narrative’ now took on a strophic form exemplified by a recurring pattern of
organisation in Heidi’s speaking and singing. Although Heidi spoke and sang as she
drew, she appeared to be fully absorbed in the drawing activity. Hence, speaking and
singing were observed to be fully integrated during the ‘drawing narrative’.
More detailed descriptions of the activities observed in Heidi’s ‘drawing narrative’
may further illustrate how speaking and singing were integrated. (i) Heidi was
constantly making choices about what she wanted to draw, and how she would
achieve her objectives. Not only did she pause for a while in between the verses to
painstakingly choose different colours for her eggs, she expressed frustration when
she could not find the specific colour she wanted and asked her mother for
assistance. She complained that she could not find the right hue among several
crayons within the same colour range, and even asked for a felt tip pen in order to get
the exact colour she had in mind. (ii) Heidi was telling a story as she drew. She
paused occasionally to look at her mother and me, as if she was trying to establish
some eye contact with her audience. Heidi also raised the volume of her voice and
spoke slowly in a clear manner. Heidi’s mother explained that her daughter
frequently made up stories for her drawings. Even when Heidi was drawing alone
in the room, she would raise her voice and tell a story as she drew. (iii) Heidi was
The referent of children’s early songs 13
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
singing as she drew. It is interesting to see how a 3½-year-old child adapted a learned
song, The Green Grass Grew All Around, as the basis for her ‘drawing narrative’.
Heidi used only the beginning and the ending phrases of the song, omitting the
more complicated middle section. To simplify even further, Heidi spoke the first
phrase but sang the ending. Each verse was given a title sentence, ‘There was (were)
___ eggs’, and followed the same recurring pattern. Heidi ended the ‘drawing
narrative’ after she drew five eggs and could not find any room for more eggs in
the nest.
Scenario 2: ‘Pot-pourri’ songs were abundant among children’s self-generated
singing. For example, at age 3 years 5 months, Heidi sang Jingle Bells spontaneously
but also attempted to fit in fragments of Old McDonald . Nevertheless, she managed
to make her way back to Jingle Bells and rendered the song in an arrangement of
verse�chorus�verse. In common practice, Jingle Bells is sung with the chorus section
as the conclusion of the song. It appeared that Heidi deliberately ended the song with
a verse section because she added a sprightly ‘Hey!’ to conclude the song. As
mentioned earlier in the discussion of ‘pot-pourri’ song, it was often difficult to
discriminate a ‘pot-pourri’ song from a mix of two melodies. Heidi clearly
demonstrated a thorough understanding of the structure of Jingle Bells because she
was able to perform both the verse and the chorus sections and achieve a sense of
unity and contrast. Hence, her attempt to integrate Old McDonald into Jingle Bells
appeared to be her experiment with manipulating learned song materials.
Songs collected at the 3�4 level showed a strong emphasis on the manipulation of
musical elements such as rhythmic patterns, scalic structure and melodic contour
(ascending and descending, and parallel and contrary motion). Songs at this age level
might be seen as musical exercises, which appear to play an important role in
assimilating culturally relevant musical practices.
Some game-like songs were collected from Amber when she was 3 years 1 month.
About 20 minutes into the recording session, she showed me her toy phone and then
shifted her attention to a toy xylophone. Using a mallet, she first played several notes
and then sang ‘doh-re-mi-fah-soh’ while playing the xylophone. Although she sang
the scale pattern one beat per syllable, her singing did not follow the pitches she
played on the instrument; i.e. the pitch movement of Amber’s singing moved in a
different direction from the scalic passage she was playing on the xylophone. From an
adult’s perspective, the choice of a parallel motion would appear to be the initial step
in an attempt to match the singing with an instrument, but to play and sing scalic
passages in contrary motion would appear to require much more sophisticated
perceptual and performance skills.
After that initial attempt to sing while playing the toy xylophone, Amber then
moved on to develop more rhythmic variety in her game-like song. At this point, the
xylophone playing and the singing were no longer synchronised. The song was
largely sung with nonsense words, which included a repetitive phrase, ‘a la-dy Oh!’,
while the earlier ascending and descending scalic pattern recurred only in the
xylophone part.
14 E. Mang
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
There was a brief diversion when Amber suddenly went to pick up her toy phone
for an imaginative play. She talked to her ‘sister’ on the phone, and said, ‘Hello
sister, . . . I’ll meet you at the ball game.’ Immediately after she hung up the phone,
Amber returned to the spontaneous song she created just a short moment ago. The ‘a
la-dy Oh!’ motive was performed in a chant-like manner with strong accent and a
clearly metrical pattern. Although the spontaneous song retained the ‘a la-dy Oh’
motive, the xylophone no longer accompanied it. The song then took on a playful
character. Amber took my hands and led me in a dance while we swung our arms and
skipped around in circular motion. At this point, she extended the repetitive ‘a la-dy
Oh!’ sequence by adding ‘and all fall down’. This game reminded me of the singing
game Ring A-Ring A Rosy. (Amber was familiar with this singing game, which she
had learned at the daycare centre.) The chanting soon became faster and faster, and
Amber even pretended to fall down onto the floor from time to time. Amber became
more excited as she sang, and the game went on until Amber was exhausted. Since
The Sound of Music was one of Amber’s favourite movies, it is speculated that the
source of this ‘a la-dy Oh!’ motive was from the song The Lonely Goatherd . In other
words, Amber was possibly yodelling.
Conclusions
To summarise, the children at age level 2�3 mostly performed incomplete renditions
of learned songs with words freely substituted and nonsense syllables in abundance.
When the words of a learned song were forgotten, a child failed to recall the melody.
In a typical learned song performed by a child at this age level, only fragments of a
melody were intact, other parts of the song had melodic and rhythmic patterns
incorrectly sung. Some children in the study recalled a song through body
movements or imaginative play rather than singing it aloud. Many self-generated
songs were observed, especially ‘pot-pourri’ songs combining fragments of learned
songs and improvised songs, sometimes regardless of different languages. Children
were also observed to make less reference to diatonic system and metrical rhythm in
their self-generated songs than in their learned songs.
A marked increase of learned songs was found in all children at age level 3�4; the
melodic and rhythmic patterns were sung mostly correctly. Although children
demonstrated a growing sense of tonality at the beginning of a song, they frequently
moved through several modulations within a song. However, the entire song was
often considered forgotten when only the words were forgotten. In some cases,
nonsense syllables were used to fill in missing words but the precise structure of
learned songs was not recalled accurately even when all other aspects of the song
were sung accurately. For self-generated singing, ‘pot-pourri’ songs and other novel
form of singing were abundant. These self-generated songs consisted of integrated
narration and adaptation of learned song, and were sometimes improvised with
movements and musical instruments, or resembled singing games and other play
activity such as drawing.
The referent of children’s early songs 15
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
Observations from this study confirm that words present the greatest challenge to
younger children when they are at the initial stage of learning to sing. The sounds of
words often operate as an anchor to the musical elements, especially the melody and
the formal structure of the song. Thus, a large proportion of configurations of
learned songs might be, in fact, a spin-off of efforts to recall the words of learned
songs. The meaning of words, on the other hand, contribute as resources for
imaginative play, whereby children often move into activities such as role play and
movement to express the storyline of the song. Maintenance of tonality still presents
a great deal of challenge, even after the children had acquired a superior level of
melodic accuracy.
A comparison between performance of learned and self-generated songs suggests
that some melodic features were produced more abundantly in songs learned from
adult models, whereas others were produced more abundantly in self-generated
songs. Figure 2 illustrates some melodic features observed in children’s early songs.
The following is a discussion of some possible explanations of the differences.
It seems likely that the cognitive load involved in performing a learned song will be
different from that for a self-generated song. Performing a learned song requires a
child to recall musical material according to a previously introduced model, which
would place a demand on good self-monitoring skills. Such learned vocal behaviours
might also be bounded by cultural conventions, which a child would be expected to
have mastered in order to reproduce the song. Self-generated songs, on the other
hand, are initiated by a child. When a child is not required to reproduce according to
a model, it is speculated that the self-monitoring mechanism could be comparatively
relaxed because the musical features become source material rather than reproduc-
tion target; the child’s aesthetic judgment would seem to be a more important
component when singing spontaneously. Pressing (1998) proposes that, with referent
as material for variation, less processing capacity (attention) is needed in the
improvised performance. Hence, it is speculated that a processing reduction might
Learned Songs Self-generated SongsMelodic features observedin children’s singing Age levels (in years)
2–3 3–4 2–3 3–4Repetitive melodic patternsMelodic contourDescending contour N/A N/AMelodic phrasingDiatonic scale structureMaintenance of tonalityAccurate intervalsTransposition of melody
Abilities not observed in any of the childrenAbilities demonstrated by some of the childrenAbilities demonstrated by all the children
N/A Not applicable (learned songs contain essentially both ascending and descending contour)
Figure 2. Melodic features observed in children’s learned songs and self-generated songs
16 E. Mang
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
result in a relatively inferior melodic performance in self-generated songs when
compared to performance of learned songs. However, Pressing points out that a
processing reduction could free up cognitive load for a child to engage in a higher
degree of integration among different modes of activities (such as the singing and
drawing activity observed in Heidi at age 3) and to enhance the expression of the
communication intent (such as the self-generated singing game observed in Amber at
age 3). Observations from the present investigation suggest that through using
learned songs as referent, a child would vary the musical features of a familiar song to
create and improvise according to their imagination. Hence, the resultant early songs
exhibit a distinctive sense of ownership and rich creative instinct.
Implications
Based on the preceding discussion on observations of young children’s early songs,
the following suggestions aim to enhance singing experience in childhood.
(a) Vocal model. Because children assimilate familiar acoustic structures from their
auditory environment, a good vocal model may facilitate the development of singing
(Goetze et al ., 1990; Green, 1990; Mang, 1997). It is hypothesised specifically that
vocal models, which demonstrate proper vocal control over such parameters as pitch,
volume, timbre, breath and articulations, are advantageous for children learning to
sing.
(b) Singing in monosyllables. Although the words of a song often appear to attract
the initial interest of younger children, evidence from research findings suggests that
young children sing more accurately in monosyllables (Levinowitz, 1989; Goetze et
al ., 1990). To assist older children to acquire more sophisticated melodic features,
songs could sometimes be presented in monosyllables. Some music educators
advocate the use of ‘chinning’ to heighten the musical characteristics of a song and to
facilitate learning musical elements while singing (Richards, 1980; Richards &
Langness, 1982; Bennett & Bartholomew, 1997).
(c) Preserving tonality. Children master the maintenance of tonality in singing at a
relatively later age. In fact, it was reported to be a problem even in school age
children (Goetze et al ., 1990; Green, 1990). At the initial stage of learning to
establish a sense of tonality, children often demonstrate the ability in familiar songs
but not in unfamiliar songs. For this reason, experience with singing a familiar song
using different tonal centres may assist children to conserve the relative tonal
relationships in a song. It may also provide children with abundant stimuli generated
from the same archetype of melodic patterns in order that they may advance
progressively and learn to maintain tonality in less familiar songs. Singing a familiar
song in different tonal centres also provides opportunities to extend the vocal
registers of young children.
The referent of children’s early songs 17
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
(d) Singing as a movement experience. Children were sometimes observed to
voluntarily integrate singing with body movements. A child usually begins by singing
a song, gradually adding body movements. Since movements appear to elevate a
child’s involvement in sensory experience, she may become too absorbed in moving
to continue singing. In other words, my data suggests that body movements may
become the performance mode of a song, whilst singing the melody and the words
may assume a secondary role. Bruner (1960, 1966) describes the initial mode of
knowledge representation as ‘enactive’. Such a mode is found when a child
demonstrates concept acquisition by engaging in overt behaviours in response to a
stimulus. Here, externalising a song in body movement depicts how musical
experience for a child may be in a ‘movement’ or ‘enactive’ mode, which holds the
musical elements of a song together even when it is not sung aloud (Bennett &
Bartholomew, 1997). In Pressing’s model of improvisation, the integration of singing
and movement is seen as novel parametric combinations (Pressing, 1988).
Connecting movements with singing can therefore be beneficial for young children,
especially during the developmental stage when they connect more easily with tactile
experience. Some music teaching approaches advocate movement activities. For
example, the Orff�Schulwerk approach encourages ‘translating body movement into
percussion instrument’ and the Kodaly approach combines vocal and movement
activities. Both approaches are widely used in elementary classrooms and in music
therapy settings.
(e) Singing as playing. Playing is an important component in a child’s singing
activities. Children in the present study displayed a rich spectrum of musical and
linguistic behaviours when they played. In fact, all self-generated songs were
collected when they were immersed in playing. Evidence suggests that the high level
of social interaction in a play context may have induced the need to express and
communicate. Again, with reference to Pressing’s model of improvisation (1988), a
high level of novel sensory input during play could facilitate the extraction or creation
of new cognitive assembles.
How can a singing activity be initiated as playing? Singing games are particularly
effective in engaging children and thus provide excellent settings for learning. A body
of research has been accumulated on the singing games of children in different
cultures (Prim, 1989; Riddell, 1990; Addo, 1995; Blacking, 1995; Marsh, 1997).
Hence, singing games can be employed as a valuable foundation for musical
experience in childhood.
Were children in the present study representative of others at their age level? This
question can perhaps, be more aptly answered through further experimental studies,
which exercise physical or statistical control of variables and measurement to
maximise the validity of findings. My investigation sought to document and interpret
broad observable patterns in early songs of young children, the complexity of which
could not be captured through unmediated knowledge about the contexts.
Qualitative inquirers maintain that ethnography aims to study specific cases rather
18 E. Mang
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
than to develop theoretical knowledge. Hence, ethnography is a training of the
capacity for perception to enhance practical wisdom (Schwandt, 1997, pp. 57�60).
Thus, for the purpose of the present study, it may be more practical to select a small
group of children for a more in depth ethnographical investigation.
Notes on contributor
Dr Esther Mang received her doctorate at University of British Columbia, Canada,
where she also taught music education courses and early childhood music
programmes. Her research interests are inter-disciplinary, encompassing child
development, speech science and choral music education. She is currently
Assistant Professor of the Hong Kong Baptist University and lectures in Choral
and Instrumental Conducting, Choral Pedagogy, Principles of Music Educa-
tion, and Music in Early Childhood.
References
Addo, A. O. (1995) Ghanaian children’s music cultures: a video ethnography of selected singing games .
Doctoral dissertation, University of British Columbia, Vancouver.
Bennett, P. D. & Bartholomew, D. R. (1997) Song works 1: singing in the education of children
(Belmont, CA, Wadsworth Publishing).
Blacking, J. (1995) Music, culture, and experience: selected papers of John Blacking . Reginald Byron
(Ed.) (Chicago, IL, The University of Chicago).
Bruner, J. (1960) The process of education (Cambridge, MA, Harvard University Press).
Bruner, J. (1966) Toward a theory of instruction (Cambridge, MA, Belknap Press of Harvard
University).
Davidson, L., McKernon P. Gardner, H. (1981) The acquisition of song: a developmental
approach. Documentary report of Ann Arbor Symposium: applications of psychology to the teaching
and learning of music (Reston, VA, Music Educators National Conference), 301�315.
Fujita, F. (1990) The intermediate performance between talking and singing*/from an observa-
tional study of Japanese children’s music activities in nursery schools, in: J. Dobbs (Ed.)
Music education: facing the future (Christchurch, New Zealand, ISME), 140�146.
Goetze, M., Cooper, N. & Brown, C. J. (1990) Recent research on singing in the general music
classroom, Bulletin of the Council for Research in Music Education, 104(4), 17�37.
Green, G. A. (1990) The effect of vocal modeling on pitch-matching accuracy of elementary
schoolchildren, Journal of Research in Music Education, 38(3), 225�231.
Levinowitz, L. M. (1989) An investigation of preschool children’s comparative capability to sing
songs with and without words, Bulletin of the Council for Research in Music Education, 100,
14�19.
Mang, E. (1997) The effects of training with male and female vocal modeling on the singing
achievement of grade one children, Canadian Journal of Research in Music Education, 38(4),
25�33.
Mang, E. (2001) Intermediate vocalizations: an investigation of the boundary between speech and
songs in young children’s vocalizations, Bulletin for the Council of Research in Music Education,
147, 116�121.
Mang, E. (2002) Vocal play as agent of musical creativity, Proceedings of the 7th International
Conference on Music Perception and Cognition , Sydney (Adelaide, Causal Productions), 1�4.
The referent of children’s early songs 19
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14
Marsh, K. M. (1997) Variation and transmission processes in children’s singing games in an
Australian playground. Doctoral dissertation, University of Sydney.
Moog, H. (1976) The musical experience of the pre-school child (C. Clarke, Trans.) (London, Schott
& Co. Ltd).
Moorhead, G. E. & Pond, D. (1978) Music of young children: Pillsbury Foundation Studies (Santa
Barbara, CA, Pillsbury Foundation for Advancement of Music Education).
Papousek, H. & Papousek, M (1981) Musical elements in infant’s vocalization: their significance
for communication, cognition, and creativity, in: L. P. Lipsitt & C. K. Rovee-Collie (Eds)
Advances in infancy research (Vol. 1) (Norwood, NJ, Ablex).
Pressing, J. (1984) Cognitive processes in improvisation, in: W. R. Crozier & A. J. Chapman (Eds)
Cognitive processes in perception of art (North Holland, Elsevier Science Publishers B.V.),
345�363.
Pressing, J. (1988) Improvisation: methods and models, in: J. Sloboda (Ed.) Generative processes in
music (Oxford, Clarendon), 129�178.
Pressing, J. (1998) Psychological constraints on improvisational expertise and skill, in: B. Nettl
(Ed.) In the course of performance (Chicago, IL, University of Chicago Press), 47�67.
Prim, F. M. (1989) The importance of girl’s singing games in music and motor education,
Canadian Journal of Research in Music Education, 32, 115�123.
Richards, M. H. (1980) Aesthetic foundations for thinking, rethought, part 1: Experience (Portola
Valley, CA, Richards Institute).
Richards, M. H. & Langness, A. (1982) Music language, Section 2 (Portola Valley, CA, Richards
Institute).
Riddell, C. (1990) Traditional singing games of elementary school children in Los Angeles . Unpublished
doctoral dissertation, University of California, Los Angeles, CA.
Ries, N. L. (1987) An analysis of the characteristics of infant-child singing expressions: replication
reports, Canadian Music Educator, 29(1), 5�20.
Schwandt, T. A. (1997) Qualitative inquiry: a dictionary of terms (Newbury Park, CA, Sage
Publications).
Serafine, M. L., Crowder, R. G. & Repp, B. H. (1984) Integration of melody and text in memory
for songs, Cognition, 16, 285�303.
Serafine, M. L., Crowder, R. G., Davidson, J. & Repp, B. H. (1986) On the nature of melody-text
integration in memory for songs, Journal of Memory and Language, 25, 123�135.
Young, S. (2002) Young children’s spontaneous vocalizations in free-play: observations of two- to
three-year-olds in a day-care setting, Bulletin of the Council for Research in Music Education ,
43�53.
20 E. Mang
Dow
nloa
ded
by [
Am
s/G
iron
a*ba
rri L
ib]
at 0
0:26
27
Nov
embe
r 20
14