Musicae Scientiae

http://msx.sagepub.com/

Mothers and Infants and Communicative Musicality

Stephen N. Malloch
Musicae Scientiae 2000 3: 29
DOI: 10.1177/10298649000030S104

The online version of this article can be found at:

http://msx.sagepub.com/content/3/1_suppl/29

Published by:

http://www.sagepublications.com

On behalf of:

European Society for the Cognitive Sciences of Music

Additional services and information for Musicae Scientiae can be found at:

Email Alerts: http://msx.sagepub.com/cgi/alerts

Subscriptions: http://msx.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://msx.sagepub.com/content/3/1_suppl/29.refs.html

>> Version of Record - Sep 1, 2000

What is This?

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

Musicz Sclentiz =1999 by ESCOM European Society
Special Issue 1999-2000, 29-57 for the Cognitive Sciencesof Music

Mothers and infants

and communicative musicality

STEPHEN N. MALLOCH
Macarthur Auditory Research Centre Sydney, University ofWestern Sydney Macarthur. Austra1ia

• ABSTRACT
Using music as a model, mother/infant vocalisations are examined using computer-
based acoustic analysis. Past research is summarised which demonstrates the
importance of both parties in the mother-infant dyad. Methods are then introduced
for analysing pulse, quality and narrative in mother/infant vocalisations. These
three elements comprise "communicative musicality": those attributes of human
communication, which are particularly exploited in music, that allow co-ordinated
companionship to arise. The analysis of pulse is based on spectrographic analysis,
and regular timing intervals are discovered that serve to co-ordinate the mother's
and infant's joint vocalisations. Quality consists of both the pitch-contour of the
vocalisations, and their timbre. Pitch plots are derived using software developed for
this project using a constant Q spectral transform. I examine how the infant and
mother structure their joint exploration of pitch space on the small and large scale.
Timbre is measured with a variety of acoustic measures - tristimulus values,
sharpness, roughness and width. It is found that the mother's voice changes its
quality in response to the infant's. Narrative combines pulse and quality - it allows
two persons to share a sense of passing time - and the musical companionship is
examined that is created between a mother and her baby as she chants a nursery
rhyme. It is concluded that communicative musicality is vital for companionable
parent/infant communication.

1
INTRODUCTION

A mother and her young baby are playfully interacting. We hear the mother speak
in shan bursts. talking in an inviting sing-song manner. and the baby occasionally
"answers back". It appears that communication is taking place. but communication
based in what? The baby cannot understand the meaning of the words the mother
is using. and the baby often answers in "gliding-type" sounds. The communication
must be "held" by means other than lexical meaning. grammar and syntax.

29

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 2
BACKGROUND

Infants usually stimulate an affectionate adult, male or female, to extended poetic or

musical speech, which often moves into wordless song, or imitative, rhythmic and
repetitive nonsense sounds. This distinctive style ofadult speech is called "rnorherese",
"parentese", or "infant directed speech" (IDS), and is attended to and responded to
with much pleasure by infants. It varies with the age and state, and motives and
emotions of the infant partner (Fernald, 1992; Papousek, M., 1992; Stern et el;
1985).
Experimental studies of infants' reactions to different elements of human
expression have revealed that infants possess complex endowments for perceiving
and stimulating parental communicative signals. It has been shown that infants can
discriminate timing patterns, pitch, loudness, harmonic interval and voice quality
(Trehub, Trainor and Unyk, 1993). An infant can learn its mother's voice from
before birth, and can recognise melodies or poetic verses that were presented for it
to hear prenatally (DeCasper and Fifer, 1980; Fifer and Moon, 1995; Hepper,
1995). Reactions of newborns to the human voice and their imitations of facial
expressions, vocalisations and hand movements, show that their awareness of human
signals, while slow and rudimentary, is already comprehensive, multimodal and
coherent at birth (Nadel and Butterworth, 1999). In fact, in the first few months,
infants' face movements and hand gestures are similar in form and timing to the
unconscious expressions that accompany speech in conversation between adults, and
the hand gestures of newborns can be entrained to the rhythms of adult speech
(Condon and Sander, 1974; Fogel and Hannan, 1985; Weinberg and Tronick,
1994; Trevarthen, 1986).
Complementary research on mothers' expressions when they are addressing their
infants shows that the mother/infant pair involves special abilities on both sides.
Mothers' speech to infants has unconscious or intuitive forms that have many of the
same characteristics in many different languages. For example, while Mandarin
Chinese is inronarionally a highly inflected language and American English is not,
it has been shown that mothers in both cultures speak in closely matching IDS
(Greiser and Kuhl, 1988). The tone of a mother's voice (the "voice quality"), and its
rhythms and melody, are all regulated in predictable ways, and these features match
the demonstrated preferences that young infants seek in a human partner (Trehub
and Trainor, 1993). Typically, mothers repeat short, evenly spaced words with
simple, sing-song intonations in a resonant yet relaxed and "breathy" moderately
high-pitched voice (Stern et al., 1982). Baby and mother listen to one another's
sounds, creating co-operative patterns of vocalisarions, Micro-analyses of these
vocalisations, and of the accompanying gestural movements, have shown that
mother/infant behaviour can generate a coherent system which is constrained by
matching rhythmic and emotional factors in the two subjects (Beebe, Stern and
Jaffe, 1979; Fogel and Thelen, 1987; Murray and Trevarthen, 1985; Stern, 1985;
Trevarthen et al., 1981; Tronick and Weinberg, 1997). Indeed, in optimal "proto-
30

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 Mothers and Infants and communicative musicality
STEPHEN N. MAlLOCH

conversations", expressive phrases are precisely alternated or synchronised between

the infant and the adult (Beebe et al., 1985; Jaffe et al., 1973; Tronick et al., 1980).
It appears that the mother's intuitive behaviour suppOrts the infant's innate
communicative capacities (Papousek and Bornstein, 1992; Papousek and Papousek,
1987; Stern et al., 1985).

3
THE QUESTION

While mother/infant interactions have often been perceived as "musical", there is

currently very little discussion or data on the precise similarities between IDS and
what we generally understand to be music. What is musical about infant/parent
communication? The remainder of this. paper will examine the musical nature of
IDS, and describe the methods that we have used to draw-out this information.

4
THE METHODS AND THE FINDINGS

When mothers and babies communicate effectively, it is clear that each is highly
"attuned" to the vocal and physical gestures of the other. This attunement is critical.
Without it, as can occur in an interpersonal disorder such as maternal post-natal
depression (Beebe and Lachmann, 1994; Bettes, 1988; Murray and Stein, 1989;
Murray et al., 1996; Papousek and Papousek, 1997), or sensory and motor disorders
of the infant (Burford and Trevarthen, 1997; Fraiberg, 1979; Hauge and Hallan
Tensberg, 1996; Preisler and Palmer, 1986; Trevarthen and Burford, 1995), both
parties will suffer. Failed attunernent has been demonstrated in perturbation
experiments, in which, for example, the mother is asked to keep a still face and to
remain silent in front of her infant for one minute. The baby protests (Tronick et al.,
1980; Murray and Trevarrhen, 1985). In another important test, communication by
video was used (Murray and Trevarthen, 1985). First, a mother and her two-month-
old infant were set up to interact by means of a video-sound link in which each was
photographed by a hidden camera and each saw the other as a video image in real
time. A recording of the mother's behaviour when she was communicating happily
with her infant "live" was replayed to the baby. The infant's immediate reaction to
the active but unresponsive replay of the mother was to make signs of protest, and
then to withdraw. This experiment demonstrates how vital it is that an infant receive
vocal and gestural responses that fit with its innate predisposition to interact with an
other. An infant seeks not just encouraging communicative forms of signal from its
mother - the signals must be appropriately timed and inflected (Nadel et al.,
1999).
The elements of the co-operative and co-dependent communicative interactions
between mother and infant combine to make-up what I have called
"Communicative Musicality". This term recognises that the mother and her infant
31

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

are partners in a musical dialogue. Communicative musicality consists of the
elements pulse, quality and narrative - those attributes of human communication,
which are particularly exploited in music, that allow co-ordinated companionship to
arise.

4.1.
Pulse is the regular succession of expressive "events" through time. An event in vocal
expression may be the beginning or end of a vocalisation, a louder moment, a turn
in the pitch-shape of the mother's voice. The most direct way to measure pulse is
through examination of a spectrograph of the mother/infant vocal interaction.
A spectrograph shows the beginning and end of the vocalisations, their general
pitch movement, and the amplitude of the vocalisations through time. Figure 1 is a
spectrographic analysis of Laura, a 6-week old infant, and her mother vocalising
together). The vocalisations are represented by their harmonic components. the
amplitudes of which are indicated by a grey-scale. calibrated as shown on the top-
left on the spectrograph. The mothers words are shown at the bottom of the
spectrograph. and the baby'svocalisationsare enclosed by rectangles.The vocalisations
are numbered below the spectrograph to allow other measures, which will be
introduced later in the paper, to be matched with the spectrograph. The infant's
vocalisarions are numbered in italics. There are three ten-second sections, each
overlapping the other by one or two seconds, as indicated. The pitch-level of C4
(middle-C) is shown by a broken horizontal line-,

(1) This recording was made in 1979, in Prof. Trevarthen'slaboratory at the University of Edinburgh,
during a project entitled "Co-operative understanding in infancy", funded by the Spencer
Foundation of Chicago. Mother and infant were seated in a room by themselves. The infant was
comfortably seated and strapped into a seat, so that its limb movements were free and it was able
to have eye contact with the mother. During the section shown in the spectrographic analysis, the
mother was instructed to freely chat and play with the infant, and to encourage the infant to smile.
The overall length of this "treatment" was around three minutes.
Simultaneous video, audio and still photographs were recorded. The mother and infant each had an
AKG C451E directional microphone pointing towards their mouth at a distance of around 30 cm.
The audio was recorded in stereo onto a Uher 4200 "Report" Stereo reel-to-reel tape-recorder,
using19cm/sec tape speed.
(2) The generation of spectrographsfor all mother/infant interactions discussed in this paper proceeded
asfollows: Theaudio recordingwas transferredto the hard disk of a DECpc 425i computer (CPUi486DX2
- 50MHz) using software and sound cards designed by Digital Audio Labs (The EdDitor vs. 2.31,
Waveform Editor and Catalogue Manager; The CardD sound card, sampling at 16 bits, at 44.1 kHz).
The sound file was converted from stereo to mono by averaging between the two channels -
separation between the two channels was not sufficient for any benefit to be gained from separate
spectrographic display. The sound file was then acquired by the Digital Signal Processing package
Hypersignal Workstation v2.02 (manufactured by Hyperception), where a Fast Fourier Transform
(FFT) was executed on the waveform file, and the resultant data on frequency displayed as a
spectrograph. The phase information was discarded.
The FFTparameterswere: Transform size: 4096; Overlap: 2048; Window: Hamming. These parameters
were chosen to provide optimum time and frequency resolution within the constraints of the software.

32

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 Mothersand Infantsand communicative musicality
STEPHEN N. MALLOCH

2 N Cl SPECTReX;RFFHIC fIft..YSIS
_

ijj -5, 3111 dBItDLDR I1M spanned I 0 - 10 seconds

N N
l!J N
o I- 1.53 .I
.J

-:,-
1.53 1.53' 1.58 1.63 1.48

.'
...'
}- "!
U
Z . if!! ,
.
W .. I s
..
.r :
:::IN
WID
____.....
•
: u.._ _....I
__
2 3 4 : 5 6 7 8
1 sec. come on that's is that
A come on again then clever oh yes right

IiM sJIlUIned: 9 - 19 seconds

N
l!J
o I t
t
.J
.. i.53 : 1.58 1.63 1.53 1.53

}- " "! • ,:11 :.: - !

U
Z , '.. .i) i :. 1.w. ..
W ,,: I , ,,- I '-=. 1\ ....
:::J
C4 - C4
W I I I I
I I I I

I.L 8 : 9 : • 10 11 12 : 13:
1 sec.
B that well tell me tell me some orh come on
right some more then more then

IiM spanned: 17 - 27 seconds

N
l!J
0
.J
I
1.53 1.6J H II
1.¥i 1.53 1.53 1.53

.
}-

@-
U t .

-
Z r-':"';. - :
"'=--.
W
:::J
o _ j:: _ ----
W
C

13
I
I
I 14 15 16 17
.
I
I

I
C4

I.L I I

1 sec.
e come on ch ch
ch ch
ch
ch
egoo goo

Figure 1. Laura (6 weeks) and her mother vocalising together.

33

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 The interpretation of the timing of this vocal exchange is shown by the
placement of vertical "bar-lines". The time-interval between each bar is indicated on
the spectrograph in seconds. These bar-lines correspond to important temporal
landmarks that occur during the course of the mother/infant interaction. In the first
ten seconds, the Ist, 3rd, 4th and 6th bar-lines (the Ist bar-line is against the left
edge of the spectrograph) occur at the onset ofan utterance; the 2nd bar-line occurs
at the lowest parr of a pitch bend, and the last bar-line of spectrograph A occurs at
an emphasised word ("is that right"). The dashed-line bar-lines seen during the thirty
seconds occur between events - the time-interval being taken from the most
common bar-length in this thirty second interaction. The dashed-line bar-lines
enclose a phrase whose beginning and ending fall within the bar, but do not define
the limits of the bar. Often, the mother allows timed-space for the infant to reply -
for instance, bar 5 in spectrograph A, and the 3rd bar ofspectrograph B (though the
infant does not vocalise during this opportunity). The infant and mother also
demonstrate very accurate small-scale timings. Note how the infant vocalises at the
end of the fourrh bar of spectrograph B during an interval equal to almost precisely
a quarter of the total length of this bar, and, during the whole of this bar, the pitch-
curve of the mother's voice can be divided naturally by this same interval. This is
shown on the spectrograph by four equally-spaced vertical lines (the interval
between each is 0.41 second). Equality of division in the mother's vocalisations is
also seen in the second full bar of spectrograph B (the interval between each line is
0.38 second). The pitch-curve of the phrase, "tell me some more then" is more
rhythmically marked and at a higher pitch in the second utterance, number 10,
when compared with utterance 9, perhaps indicating greater "engagement" on the
parr of the mother, encouraging her baby to join in. In spectrograph C, the clicking
sounds that the mother makes with her tongue fall into regular groups, again
as shown on the spectrograph by equal-spaced vertical lines. These are spaced at
0.2 seconds - that is, the clicks are uttered at twice the rate of the sub-divisions
marked in spectrograph B. Overall, the bars show a remarkable regularity, and
the whole interaction shows that the mother and infant enter into co-ordinated,
negotiated communication'.
An alternative method of barring mother/infant vocalisations is shown in Figure 2.
These are spectrographs of Sarah, who is 12 weeks old and correspondingly more
vigorous than Laura, vocalising together with her mother', Unlike Figure I, these
spectrographs do not overlap. Here, a smaller basic time unit is shown which fits to

(3) The periodicity of this bar-structure is further demonstrated from the results of performing a
4k FFTon the loudness data of this 30 second excerpt. The results show that the highest spike lies
at around 0.3 s -1, with a lesserspike around 0.6 s -1. This demonstrates periodicities of around
3.3 and 1.6 seconds. Given the non-exact nature of the bar structure, and that the bar-lines
sometimes lie between events, this result further strengthens our initial belief of a bar-structure of
approximately 1.53 seconds - evidence of a regularity that allows the mother and infant to
negotiate their tum-taking.
(4) Recorded under the same circumstances and conditions as described in note 1.

34

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 Mothers and Infants and communicative musicality
STEPHEN N. MALLOCH

the length of mother's and infant's utterances, and is a subdivision of the longer time
units that can be discerned in this thirty second extract.

...
2. Cl SPECTROGRfFHIC Aft..YSIS
_

-Ii. _ dIIICIIUIII JI ............: 0- 10 seconds

N ." (ii) (iii)
l!J ."
o

r
2.03 2.03 2.06"
..J r I
t-
Il
-Q:. -'- - -'
, il
_.J _ _ _ _ _ _
«:...;.''''
"--
"-'
1 sec.
chhh come on chhh aw tell
A

JI_ .pa....... 10 - 20 seconds

N
l!J
o
..J

}-
U
Z
W
:J
(] C4
W
1.L
1 sec.
me a story come on come on a-ha come on xxx
B x =mouth-dick

,.
JI_ .pa....... 20 - 3D seconds
N
l!J 2.03 2.03 2.03 1.7 2.03
0 I

..J .
}-
u
z
W
:J
(] C4
I!-q-----
,",
W
1.L
1 sec.
x x are you still still sleepy a ha ha ha
sleepy
C
Figure 2. Sarah (12 weeks) vocelising with her mother.

35

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 In spectrograph A of Figure 2, we see the 0.68s interval emerge as the time unit
by which the baby's and mother's utterances may be naturally divided. These first
three 0.68s intervals form a larger interval of 2.03 seconds, which is surrounded by
periods of silences. This larger interval of 2.03 seconds is repeated twice more, and
then repeated a third time in a "stretched" version of 2.06 seconds (which extends
to spectrograph B). The first interval of2.03 seconds (i) is defined by the beginning
and end of a group of utterances from baby and mother; the end of the second
interval (ii) is marked by the very next vocalisation by the infant; the third interval
(iii) ends with the second of two vocalisations by the infant; the conclusion of the
"stretched" interval is marked by the conclusion of a phrase by the mother. The
continuation of intervals of0.68s and 2.03s, as well as variations on these time units,
is marked on spectrographs B and C of Figure 2 - the longer time units consisting
of a single phrase by the mother kg., "come on," "are you still sleepy"). As in
Figure 1, a dashed-line bar-line indicates a division lying between vocal events.
In spectrograph C of Figure 2 we find an important time interval not yet
discussed. At the point where the mother says"a ha," and immediately after this, two
intervals of 0.34 seconds are indicated - the first interval is shown as belonging to
the primary bar-structure, and the second is shown as having a subsidiary role. These
two intervals of0.34 seconds couId have been shown as one interval of0.68 seconds.
However, for the interval of 0.68 seconds to lie at the natural boundaries of the
baby's vocalisations of the last three seconds of spectrograph C, it is necessary to
insert this smaller interval, which is exactly half the length of the dominant
time interval. This suggests that a smaller time interval may be in operation "below"
the interval of 0.68 seconds. In addition, this smaller interval, along with smaller
intervals that are not simple fractions of the dominant interval, can be regarded as
"correction intervals" - periods of time that allow mother and infant to shift the
placement of the "beat", allowing one to meet the other in time, or "get in step". If
the mother and infant are thought of as two persons meeting to walk together, these
correction intervals allow their steps to coincide.
From Figures 1 and 2, it appears that the mother and infant tend to vocalise
naturally in a co-ordinated and rhythmic fashion. Two different ways of "barring"
the vocalisations have been presented. It is possible that the two different time
structures - one using much longer intervals than the other, and the appearance of
correction intervals - indicate "layers" of inter-dependent time structures.
Figures 1 and 2 show the vocalisations of infants of 6 and 12 weeks old
respectively. Rhythmic elements are also found in vocal exchanges of premature
infants with their parents. Figure 3, from a video documentary made by Saskia van
Rees (1987) shows a spectrograph of a 2-month premature infant (32 weeks
gestational age) vocalising with her father. In the figure, the infant's vocalisations are

(5) Although 3 times 0.68 is equal to 2.04, the units of 0.68 seconds are grouped into larger units
of 2.03 seconds due to the rounding of the length of this larger unit to two decimal places.

3&

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 Mothers and Infants and communicative musicality
STEPHEN N. MALLOCH

•• ANALYSIS
Il3 -3.5lJI dBlCILDR II"", ....nned: 0 - 30 seconds
I 6.8 I I 4.5 3.3 I Father
o B
-.J
*1 * 1 * 1* 1* 1 1 *1 * *
, •
}-
U
Z ,.,.
W
:IN
C1Irc4
w...
!.J...'" ,Infant
N +---+----!--i---+----t--i---+---+--t-------t
3IBCS.

Tlo,o spanned: 10 seco nds

2.2 11.35 1.4 Father
0.75 1.5 0.65 0.7 0.7 0.7 0.85 TURN-TAKING
* 1 • I * I * I
,
}-
U _fi..... ;;--t-
Z
W
J--- ___ ...:.. i' t:"-
...
:J
QI C4
_. ,p,; -- .- ..... -
W '".. \Jlop

t, 2.15 1.4 I 1.55 I Infant

I
1 sec.
B

Figure 3. A two-month premature infant vocalising with her father.

marked by *, the father's by !. The father is kangarooing his baby - he is holding

the baby inside his shirt. During this intimate physical contact, father and baby
exchange "coos" in alternation. A hierarchy of time elements, described as "syllables",
"utterances" and "phrases", has been reported in the spontaneous vocalisations of
infants by Lynch (Lynch et al; 1995). In spectrograph A of Figure 3, it can be seen
that both father and baby are spacing intervals in what Lynch calls phrase units of
about 4 seconds, and in spectrograph B, which is a detail of the region marked B on
spectrograph A, we see that father and infant are imitating one another at syllable
intervals of about 0.75 seconds - again, this time interval concurs with Lynch.
Thus, even such an immature human has a complex periodicity of vocal expression
comparable with that of adult expression, and is similar to the rhythmic patterns
shown by infants in spontaneous vocalisations weeks and months after full-term birth.

37

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

4.2.
Quality is the second dimension of Communicative Musicality. Quality consists of
the melodic and timbral contours of the vocalisations (equivalent to the contour and
speed of the bodily gestures). Melodic contours are measured from pitch-plots, and
timbre contours are measured from the combination of methods that will be
introduced later in the paper. Let us look first at melodic contours.
While spectrographs are very good at allowing us to measure the time-intervals
between vocalisations, they are not good at revealing high resolution fundamental
frequency information, especially at lower frequencies. In order to investigate how
the expressive movements of the mother's and baby'svocalisations move in frequency
space during their interactions, a method is needed for graphing the pitch of the
vocalisarions,
Brown (1991) describes the calculation of a constant Q spectral transform that
produces a constant pattern in the log frequency domain for sounds with harmonic
frequency components. In Brown (1992), a method is described whereby this log
frequency pattern is correlated with an ideal harmonic pattern, and thus the
fundamental frequency can be determined. Brown's methods have been implemented
by us in software, enabling the pitch of the vocalisations to be plotted.
Figure 4 shows a pitch plot of the data represented in Figure 1. A data point is
calculated every 0.01 seconds, and the frequency resolution is one quarter of a tone.
Pitch is represented on the vertical axis (in letter names - C4 is equivalent to
middle-C) and time is represented along the horizontal axis. Numbers, which number
the mother's and baby's vocalisations, are placed below the graph, and refer to the
numbers in Figure 1. As in Figure I, each of the baby's vocalisations is enclosed by
a rectangle. Figure 4 also shows the strength of correlation between an "ideal"
harmonic spectrum and the actual spectrum - roughly equivalent to how strongly
pitched an event is. The key to this is shown on the right of the graph - the darker
the small circle that indicates the pitch, the stronger the correlation. Like the
spectrograph of Figure 1, these pitch-plots overlap one another by 1 or 2 seconds, as
indicated.
Pitch plots allow us to observe how the pitches of the mother's and infant's
vocalisations progress during the course of an interaction. The Ll-shaped pitch
curves of the first three vocalisations in pitch-plot A are characteristic ofa vocalisation
that invites activity. Note how after the infant vocalises twice (numbers 5 and 6),
with the second having a higher pitch than the first, the mother continues and
exaggerates this upwards pitch movement with her very next vocalisation. In her
vocalisation number (9), she summarises this pitch movement with an upwards
sweep,which she follows with U-shaped pitch movement in this higher register (10).
The infant then vocalises a short downwards moving pitch shape, which the mother
immediately follows with an exaggerated downwards pitch movement. With the
infant's flatter pitch shapes at (14) and (15), the mother replies with a flatter pitch
shape. We see the mother being acutely aware of the pitch level and shape of the

38

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 Mothers and Infants and communicative musicality
STEPHEN N. MALlOCH

A
(data point.: 0.01&)

..
o
• very
good
F"
Ii:
• good
D

.
04
• • poor

o
F" overy
It
poor
D

C3
O.DO a.DD • . IXJ :a.0a 4.DD II.... 5.DD .... DO • • DD '.DO .La.DO

come on again come that's clever oh, yes is that right

on then
2 3 4 5 tl 7 8

B
.... tc:h (dala points' 0.01.)

• very
o good
F"
Ii:
• good
D

·..
"4

o
• poor

""D '
o ve'"
poor

c:II
'.IXI SO.DO .L.... DO .&.a.DO .1.3.00 ..... 00 .&5.110 .1.5.011 .1.. ... 00 ..... DO .1.'.00

is that right well tell me orh come on

8 9 some more then 10 11 12 13

c
(data points: 0.01s)

.
o
• vllry
good
...
E good
D

"4
..• e- poor

.."'
a
overy
poor
D

C3
...... 00 l.a.DD ..'.DD aD.DD 8a.ao _.00 a:I.Da ".DD 1aS.00 a&.DO ..... DO

come on egoo goo

13 14 15 16 17

Figure 4. Pitch plot of Laura (6 weeks) and her mother, taken from data represented in Figure 1.
39

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 infant's vocalisations, and she acts to create a balance in their joint exploration of
pitch space.
This last point - the creation of balance - is borne-out strongly in Figure 5.
This figure shows the 26 seconds of vocalisations shown in Figure 4 condensed into
a single pitch-plot. Here we see evidence ofa ca. 20 second cycle.The mother's pitch
rises at 8 seconds. and returns to middle-C at 26 seconds. In subsequent pitch-plots,
this mother continues this 20 - 25 second cycleof pitch movement. Thus, it appears
that mother and infant explore pitch-space in a methodical manner over both short
and long intervals of time. The examination of pitch plots of other mothers over a
span of 5 minutes, also reveals regular 20 - 25 second "waves" of pitch movement by
the mother, with the infant following this trend in its own vocal pitches. We have
also found evidence of very precise pitch and rhythmic matching by infants to their
mother's vocalisations (both between and with maternal vocalisations). These infant
vocalisations are often "musically logical" - particularly during songs sung by the
mother (see section 4.3 below). Pitch, rhythmic and loudness movements by the
mother and infant will have important emotional and motivational roles within
their co-operative interaction.
The second element of Quality is that of timbre contour. Timbre is a
multidimensional attribute of sound - unlike pitch and loudness, timbre cannot
be adequately described by anyone measure.
How can we measure timbre? One technique is the Tristimulus Method
developed by Pollard and Janson (1982). The relative loudness of three spectral areas
of a harmonic sound - the fundamental, harmonics 2-4, and harmonics 5-n - are
compared. The results are plotted in a triangular space (which omits a time axis).
Figure 6 shows a tristimulus plot of the five vowel sounds sung to a fixed pitch.
The horizontal axis measures the relative loudness of the middle harmonics (2-4),
and the vertical axis measures the relative loudness of the upper harmonics (5-n).
Thus, the placement of a sound within the tristimulus space gives an indication of
its sound quality.
Figure 7 shows a rristimulus graph of the vocalisations represented in the
spectrographs of Figure 1 and the pitch-plots of Figure 4. The numbers refer to the
numbers found in these previous figures.
The infant's utterances are enclosed in a rounded rectangle. We can see that
the timbre of the infant's vocalisations is very different from that of the mather -
they lie in a different section of the tristimulus space. On the diagram, the mother's
utterances have been loosely classifiedon the basis of their communicative function.
We can see that an the basis of this grouping, utterances that share a similar
funcrion are found within a particular area of the rristimulus space6.

(6) The mother's utterance "Is that right?" (number 8), classified here as "inviting", might be
seen instead as "affirming". However, its upwards pitch movement aligns it with other "inviting"
utterances.

40

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 Pitch (data points: 0.01 s)
o
F
E
D

A • very
• ..
ie good
0 e a
F
E
'-
lID
IL. -
,- Q
,- &
w. ;- ••• D • good
...
.. D - G) • I rllI ...-:: Ii
C4 I.... 'T)
• •
..a iI
B - • .-
..• -..
.. t= poor ::r
.f
iil
A eD
0
•
•
lil
-
G 0
- •
:f- ...s
;,:'
::I
F overy lit
E !:
poor ...
D S

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

3
C3
.,,::1
0.00 .10.00 2:0.00
z<!:
TII1lI!!/5I!!C

Figure 5. Pitch plot of Laura (6 weeks) - entire extract.

 06 H: harmonics 5 - n
M: harmonics 2 - 4
H A.
04 E
•
I•
U

02

02 04 Q6 08
M

Figure 6. Tristimulus values of sung vowel sounds.

It was said that timbre is a multidimensional attribute - a number of different

measures can be applied in order to gain a fuller understanding of the quality of the
sound. Three other timbre measures that we have used are measures of sharpness,
width and roughness. Sharpness, measured in acums, is calculated according to the
formula proposed by Aures (1985). It is related to the position of the loudness
centroid in a sound's specrrum. Timbral width is a measure of the fraction of
loudness that lies outside of the loudest 1/3 octave band. It is a measure of how
"expansive" or "narrow" a sound is heard to be. This measure is derived by Malloch
(1997; and in press) from the Tristimulus Method. Roughness is caused by beating
between partials. The model of roughness measurement we use is that proposed by
Hutchinson and Knopoff (1978)7.
Figure 8 shows these three measures of timbre combined for the vocalisations of
Laura (6 weeks) and her mother. Because these three measures have been combined

(7) The computer programs used to calculate these measures of timbre were all written at the
University of Edinburgh for the author's PhD, and later adapted for this project.

42

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 Mothers and Infantsand communicative musicality
STEPHEN N. MAlLOCH

Q9
H: harmonics 5 - n
Q8
M: harmonics 2 - 4
Q7

Q6
o = infant

Q5
H
Q4

Q3

02

Q1

o Q1 02 Q3 Q4 Q5 Q6 Q7 Q8 Q9
M
* coming immediately after a vocalisation by the infant
Figure 7. Ttisiimulus values for utterances shown in Figures 1 and 4 (Laura, 6 weeks).

on the one graph, the y axis is dimensionless. We are interested in the movement of
the three measures relative to one another, within the context of the mother/infant
vocalisations,
In Figure 8, each unbroken horizontal bar represents the average value of that
quality (roughness, sharpness or width) for the duration of each vocalisation. To
differentiate the different measures on the graph, the unbroken bars are "patterned",
and are joined by thin broken lines. As in previous graphs, a rectangle placed around
the data shows that it is a representation of a vocalisation by the infant. The numbers
below the vocalisations correspond to the numbering found in Figures 1,4 and 7.
A particular trait of the mother that is shown by these graphs is the way in which
her voice quality changes after each of the vocalisations by the infant. After all three
sets of vocalisations by the infant, the timbral measures for the mother's voice drop
from the levels immediately prior to the infant's vocalisarions, This may indicate

(8) The one exceptionis the mother'svocalisation number 12. Here, there is a very slight rise in the
measure for width. The other two measures, however, fall.

43

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 - .
......... _--------------_..

-
roughness _

<: -
I ..
\
I.- "\'"
...
'.
IUUI.-------IIIII---------- 111I11101-------- -IIIIIIIIIPIll
width
!IUIIUU
IIPlllllllnnt!
,."'-'::
..... -----_._- ... .........
.. ...
....•... ... i
I
i

sharpness
1 2 3 4 5 6 7 8
seconds
o 1 2 3 4 5 6 7 8 9 10

roughness
-_ -
1111
.•..
.'
...•

----- -

---I1I11III11PIIIIII 1111111111111111I"-------11I11111111
width

sharpness -, -;

---- .-... ...---------""-.::

8 9 10 11 12 13
9 10 11 12 13 14 15 16 17 18 19

-
roughness

1-·------- -------------
------
I
IYUU_-------il1I1I1II1I1I1
111111I1111
width

sharpness ... . . -----------------:o..,"'.,..

13 14 /5 16 17
17 18 19 20 21 22 23 24 25 26 27

Figure 8. Measurements of Timbre for utterances shown in figures 1, 4 and 7.

44

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 Mothers and Infants and communIcative musicality
STEPHEN N. MALLOCH

that, by changing her voice quality. the mother signals to her infant that she has
heard her. These measures may also indicate that the mother wishes to make her
voice more similar to that of her infant. Referring back to Figure 7, we can see that
the mother's vocalisations that occur after the infant's (numbers 7. 12 and 16) lie in
the same horizontal region of the tristimulus space as all but one of the infant's
vocalisations, indicating a timbral similarity.
We have discussed the elements of pulse and quality. All these attributes of the
communicating mother/infant dyad combine to create the Narrative structure of
their companionship.

4.3.
Narratives of individual experience and of companionship are built from the units
of pulse and quality found in the jointly created gestures of vocalisations and bodily
movement. Narratives are the very essence of human companionship and
communication. Narratives allow two persons to share a sense of passing time. and
to create and share the emotional envelopes that evolve through this shared time.
They express innate motives for sharing emotion and experience with other persons
and for creating meaning in joint activity with others (Stern, 1993; Trevarthen,
1988; Trevarthen, 1998).
Figure 9 represents a mother saying a nursery rhyme with her 4 month old baby
girl 9. Mother and infant are in a naturalistic surrounding. but the mother is not
holding the baby so the infant is not being "bounced" in time with the rhyme. It is
obvious. from the infant's reactions. that this is a rhyme the baby has heard many
times before. The baby shows signs of pleasure. and smiles as soon as the mother
begins. For the first verse. the baby does not vocally join in. In the second. the mother
ceasesto articulate the words - rather. she saysa rhythm - "di dum. di dum. di dum".
This rhythm is represented in musical notation in the figure. Below the mother's
rhythm the vocalisations of the baby are also represented in musical notation, and the
time of each bar is indicated in seconds above the bar. (For reasons of visual clarity,
rests are omitted in the transcription where it does not create ambiguity.) In the
second verse of the rhyme. we can see that the baby is joining in with the mother in
a musical fashion. Her vocalisations are in time with the mother. Notice how the
baby's contribution shows musical variety, as well as consistency. In the first bar of
the second verse. the baby usesthe "up-beat" idea that is so prominent in this rhyme. In
the second and third bars, the baby vocaliseson the beat. In the third bar. the baby
appears to make use of the rest by introducing a new rhythmic idea in the form of
a triplet (the baby laughs in a triplet rhythm). In the third verse. the baby consistently
vocalises on the last beat of each bar. and provides the up-beat to the beginning of
the verse. which the mother omits. This is very different to what the baby did during
the second verse. It appears that the baby is changing her musical style from verse to
verse.

(9) Recorded under the same circumstances and conditionsas described in note 1.

45

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 mother 2.2 s 2 2.3 2.2

I'll 1'; S 1'; nnn SJ'J'J'

I' ;l I' 1 I' 1'1' m 3
'-----'"
infant SECOND VERSE

...
CII

mother
mother 2.5 2.3 laughs
2.4

nnn. nnn. n 1 n n nn.S

infant
I'
1n 1·
'---.-J

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

THIRD VERSE

Figure 9. Musical notation representing the participation of mother and baby (4-month old) in a rhythm derived from the nursery rhyme Clap-a-c1ap-a-
handies (rests are omitted where this does not causeambiguity),
 Mothers and Infants and communicative musicality
STEPHEN N. MALLOCH

In the third bar of the third verse, the baby makes what can be described as a
"musical joke". After vocalising precisely on the last beat of bars 1 and 2 of verse 3,
the baby still provides this beat in bar 3, but enters a semiquaver early. The baby
vocalises particularly vigorously at this point - it may be that the baby wants to
emphasise the fact that she is entering early. However, whether or not this is a
deliberate act on the part of the infant, that this "early" vocalisation is sensed and
appreciated by the mother is suggested by the mother's laughter immediately
following it. After this shared tour-de-force by both musical partners, the rhyme
loses energy, and other play takes over.
From this example, we can see that an infant at 4 months of age is quite capable
of entering into the "structure" of a musical game with an other, participating in a
musically logical way. Her vocalisations during her mother's rhyme show a true
musical feeling - they support the musical structure of the rhyme - they never
work against it.
In this last example, we have seen what can be described as a musical narrative
created through the companionable interaction of a mother and her infant. We have
also seen how timing is a vital element in mother/infant vocalisations, and how this
timing can be represented as a series of bars, or as a series of smaller time-units. We
have seen how mother and infant can use pitch imitation and complementation
during their vocal exchanges - how both appear to have a strong sense of the pitch
space that their joint vocalisations are covering, and how each relates to the other in
a musical manner. We have seen how the mother is intuitively aware of the timbre
of her voice, and how her modification of this timbre is an integral part of her vocal
communication with her infant. Lastly, we have seen how a mother and her infant
can jointly create a musical piece - both are musical partners within their
communicative space.
In his article in this issue, Bjorn Merker (l999) suggests that synchronous
chorusing and bodily gesture in our common ancestor with the chimpanzee,
coming about as a result of sexual selection, may be the vital evolutionary step
making the development of language possible. He suggests that the ability to
vocalise and gesture together in time may underlie language. Similarly, it is our
contention that the ability to act musically underlies and supports human
companionship; that the elements of communicative musicality are necessary for
joint human expressiveness to arise, and lie beneath, to a greater or lesser degree, all
human communication.
At the start of this paper, we introduced the concept of communicative musicality,
and we have emphasised the musicality of motherlinfant vocalisations. However, as
has been suggested above, this term is not to imply so much that mother/infant
communication is musical - rather that what we generally call music is one particular
drawing-together of the elements of pulse, quality, and narrative - elementsthat are
intrinsic to all human communication. The elements of communicative musicality are
the tools by which emotion is conveyed and thus companionship formed. What

47

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 happens when these elements are present to a lesserdegree. or are absent? In her article
in this issue (also see Robb, 1998), Louise Robb shows that in vocal exchanges
between mothers suffering from post-natal depression and their babies, regularity of
timing is present to a far lesser extent. and when it is. the regular units are slower
than in happy mother/infant dyads. As well as this, the pitch of the mother's voice
is generally lower. the pitch shapes that her vocalisations form are "flatter". and when
the vocalisations do take-on a more "shapely" appearance, they are generally more
drawn-out through time. Further research on the effecton the elements ofcommunicative
musicality due to post-natal depression is currently being undertaken by the author.
The movement from mother/infant vocalisations towards song is illustrated by
Figure 10. This figure shows a pitch plot of a favourite nursery song Rock-a-byt-baby.
As with the preceding pitch-plots. time is represented along the horizontal axis. and
pitch along the vertical. Comparing the pitch shapes here to the pitch shapes we
have seen in Figures 4 and 5, we can see a certain similarity - the pitch shapes in
Figure 9 are more stylised, but we see the same operations of imitation and
complementation in both instances. For example, the first three utterances by the
mother in spectrograph A of Figure 4 show imitation. Utterance 4 introduces a new
idea. The baby introduces yet another idea - a rising figure - which the mother
completes with utterance 7. and then elaborates with utterance 8. Utterances 9 to
17 show at first a big sweep of pitch upwards. and then a gradual coming to rest.
We can see very similar processes at work in the poetic and musically "dramatic"
lullaby. The first two phrases of Figure 10 (0 - 14 seconds) show imitation. along
with some variation. The last phrase (22 seconds to the end) shows a new idea.
which is completed by the downward movement to the end of the piece. The way
in which the lullaby has a shape which leads to a climax at 22 seconds and then
resolution at its end. is very similar to the overall shape of the vocalisations as
shown in Figure 5 - a climax at around 7 seconds. and then a movement towards
a resolution.

5
CONCLUSION

Communicative musicality is the art of human companionable communication. It

consists of our innate abilities, which function from birth. for being able to move
sympathetically with an other. It is the vehicle which carries emotion from one
to the other. When our ability to share emotions is impaired. it appears that
the elements of communicative musicality change in ways that make them less
"musical".
We have discussed communicative musicality in terms of pulse. quality and
narrative. We have seen that in these three areas systematic movement occurs
between mother and infant - movement that allows mother and infant to express
themselves in ways that are sympathetic with the other.
48

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 Plt:ch (data points: 0.05 s)
G

F
E

C:S
B

A • • very
G
- good
F
... -- ....,.
E .-
--•-- - • good
D :i •
I II
.. ........ .• •--- .....- ... -
• •
C4
B
1- IlL
---• .,.. . ...-- •
- •
A --- • • poor
•• ii1
G ::I
i.
...
F
E Overy
poor
D ...

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

C3 S
3
0.00 4.00 8.00 1.2.00 1.6.00 20.00 24.00 28.00
",,::I
Til"le/• •c
z=
Figure 10. Pitch plot of Rock-a-bye-baby. ;z:i
The words are: Rock-a-bye baby on the tree top I When the wind blows the cradle will rock I When the bough breaks the cradle will fall I Down will come
baby cradle and all.
 Movement - gestural, vocal and emotional - is what allows communicative
musicality to occur. When this movement is constrained or impeded, communicative
musicality suffers, and companionship suffers. Acoustic analysis of vocalisations
allows us to see into this movement is a very precise and illuminating way 10.

(10) I wish to dedicate this paper to Prof. Colwyn Trevarthen, without whose passion and drive
for understanding I would never have been inspired to enter the fascinating world of infant
communication.
I also wish to thank Dr. David Sharp for his work in computer programming, and Dr. Murray
Campbell for his advice on acoustic analysis. The research for this paper was conducted in the
Department of Psychology at the University of Edinburgh. and was funded by a grant from the
Leverhulme Foundation.

Address for correspondence:

Stephen N. Malloch
Macarthur Auditory Research Centre Sydney
University of Western Sydney Macarthur
PO Box 555. FASS 4B
Campbelltown NSW 2560
Australia
Phone: +61 2 9772 6582
e-mail: [email protected]
50

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 Mothers and Infants and communicative musicality
STEPHEN N. MALLOCH

• REFERENCES

Aures, W. (1985). Berechnungsverfahren fur den sensorischen WohlkJang beliebiger Schallsignale.

Acustice; 59, 130-41.
Beebe, B., Jaffe, J., Feldstein, S., Mays, K., & Alsen, D. (1985). Inter-personal timing: The
application of an adult dialogue model to mother-infant vocal and kinesic interactions.
In F. M. Field and N. Fox (eds.), Social Perception in Infants. Norwood, NJ: Ablex,
Beebe, B., & Lachmann, F. M. (1994). Representation and inrernalisation in infancy: three
principles of salience. Psychoanalytic Psychology, 11(2), 127-65.
Beebe, B., Stern, D., & Jaffe, J. (1979). The kinesic rhythm of mother-infant interactions. In
A. W. Siegman and S. Feldstein (eds.), OfSp((ch and Time: Tmlporal Sp((chPattern: in
Intnpmonal Contats. Hillsdale, NJ: Erlbaum.
Bettes, B. (1988). Maternal depression and mctherese: temporal and intonational features. Child
Development; 59, 1089-96.
Brown, J. C. (1991). Calculation of a constant Q spectral transform. journal of tb« Acoustical
Soci(ty of'America, 89, 425-34.
Brown, J. C. (1992). Musical fundamental frequency tracking using pattern recognition method.
journal ofth( Acoustical Soci(ty of'America; 92, 1394-1402.
Burford, B. (1992). Communicating through movement and posture. In W. MacGillivray,
W. I. Fraser and A. Green (eds.), Hallas' Caringfor Peapl« with Mental Handicap
(pp. 94-102). London: Butterworth-Heinemann Ltd.
Burford, B., & Trevarrhen, C. (1997). Evoking communication in Rert syndrome: Comparisons
with conversations and games in mother-infant interaction. European Child and
Adolescent Psychiatry, 6, 1-5.
Condon, W. S., & Sander, L. W. (1974). Neonate movement is synchronized with adult speech:
interactional participation and language acquisition. Science, 183,99-101.
De Casper, A. J., & Fifer W. P. (1980). Of human bonding: newborns prefer their mother's
voices. Science, 208,1174-76.
Fernald, A. (1992). Meaningful melodies in mothers' speech to infants. In H. Papoufek,
U. Jurgens, and M. Papoukk (eds.), Nonv(rbal\lOcal Communication: Comparativ( and
Deuelopmental Aspect: (pp. 262-82). Cambridge: Cambridge University Press/Paris:
Editions de la Maison des Sciences de l'Homme,
Fifer, W. P., & Moon, C. M. (1995). The effects offetal experience with sound. In J.-P. Lecanuet,
W. P. Fifer, N. A. Krasnegor and W. P. Smotherman (eds.), htal Deuelopment: A
Psychobiological Pmputiv( (pp, 351-66). Hillsdale, NJ: Erlbaum.
Fogel, A., & Hannan, T. E. (1985). Manual actions of nine- to fifteen-week-old human infants
during face-to-face interaction with their mothers. ChildDevelopment, 56, 1271-79.
Fogel, A., & Thelen, E. (1987). Development of early expressive action from a dynamic systems
approach. Developmenta! Psychology, 23, 747-61.
Fraiberg, S. (1979). Blind infants and their mothers: an examination of the sign system. In
M. Bullowa (ed.), &for( Spuch: The B(ginning of Interpersonal Communication
(pp. 149-69). London: Cambridge University Press.
Greiser, D. L., & Kuhl, P. K. (1988). Maternal speech to infants in a tonal language: support for
universal prosodic features in motherese, Deuelopmente! Psychology, 24, 14-20.
Hauge, T. S., & Hallan Tensberg, G. E. (1996). The musical nature of prelinguistic interaction.
Nordic journal ofMusic Th(rapy, 5(2), 63-75.
51

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

Heimann, M. (1989). Neonatal imitation, gaze aversion and mother-infant interaction. Infant
&haviour and Development; 12, 495-505.
Hepper, P. G. (1995). The behaviour of the fetus as an indicator of neural functioning. In
J.-P. Lecanuer, W. P. Fifer, N. A. Krasnegor and W. P. Smotherman (eds.), Fttal
Development: A Psychobiological Penpectiv« (pp, 405-17). Hillsdale NJ: Erlbaum,.
Hutchinson, w., & Knopoff L. (1978). The acoustic component of Western consonance.
Inttrj"act, 7(1).
Jaffe, J., Stern D. N., & Peery J. C. (1973). Conversational coupling of gaze behaviour in
prelinguistic human development. JournalofPsycholingvistic Restarch, 2, 321-30.
Lynch, M. P., Oller, D. K., Steffens, M. L., & Buder, E. H. (1995). Phrasing in prelinguistic
vocalisations. Dtvtlopmmtal Psychobiology, 28, 3-25.
Malloch, S. (1997). Timbre and Ttchnology: an analytical parmmhip. Ph.D. thesis, Faculty of
Music, University of Edinburgh.
Malloch, S. (in press). TImbre and Technology: An Analytical Partnership, Conttmporary Music
Rtvitw.
Meltzoff, A. N., & Moore, M. I<. (1977). Imitation of facial and manual gestures. Science; 198,75-80.
Merker, B. (1999). Synchronous chorusing and the origins of music. (This volume).
Murray, L., & Stein, A. (1989). The effects of post-natal depression on the infant. Baillim's
ClinicalObstttrics and Gynatcology, 3(4),921-33.
Murray, L., & Trevarthen C. (1985). Emotional regulation of interactions between two-month-
olds and their mothers. In T. Field and N. Fox (eds.), Social Perception in Infants.
Norwood, N.J.: Ablex.
Murray, L., Fiori-Cowley, A., Hooper, & Cooper, P. (1996). The impact of postnatal depression
and associated adversity on early mother-infant interactions and later outcome. Child
Development, 67, 2512-26.
Nadel, J., & Butterworth, G. (eds.) (1999). Imitation in Infancy. Cambridge: Cambridge
University Press.
Nadel.L, Carchon, I., Kervella, c., Marcelli, D., & Reserbat-Plantey; D. (1999). Expectancies for
social contingency in 2-monm-olds. Developmental Science; 2(2), 164-73.
Papousek, H., & Bornstein, M. H. (1992). Didactic interactions: intuitive parental support of
vocal and verbal development in human infants. In H. Papousek, U. Jurgens and
M. Papousek (eds.), Nonvtrbal Vocal Communication: Comparativt and Dtvtlopmmtal
Asptets. Cambridge: Cambridge University Press/Paris: Editions de la Maison des
Sciences de l'Homme, 209-29.
Papoukk, H., & PapouSek, M. (1987). Intuitive parenting: A dialectic counterpart ro me infant's
integrative performance. In Osofsky (ed.), Handbook of Infant Development
(2nd edition). New York:Wiley.
Papousek, H., & Papousek, M. (1997). Fragileaspects ofearly social interaction. In L. Murray and
P.J. Cooper (eds.), Postpartum ,uprtssion and childtkvtlopmmt (pp. 35-53). New York:
Guilford Press.
Papousek, M. (1992). Early ontogeny of vocal communication in parent-infant interactions. In
H. Papousek, U. Jurgens and M. Papousek (eds.), NonVtrbal Vocal Communication:
Comparativt and Dtlltlopmmtal Asptets (pp. 230-61). Cambridge: Cambridge
University Press/Paris: Editions de la Maison des Sciences de l'Homme,
Pollard, H., & Janson, E. V. (1982). A rristimulus method for the specification of musical timbre.
Acustica, 51, 162-71.
5:.1

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 Mothers and Infants and communicative musicality
STEPHEN N. MAllOCH

Preisler, G., & Palmer, C. (1986). The function of vocalization in early parent-blind child
interaction. In B. Lindblom and R. Zerrerstrorn (eds.), Precursors of Early Spuch.
Basingrtokt (pp, 269-77). New York: Macmillan.
Robb, L. (1998). Emotional-mulicality in mothtr-infantvocal aiftct:a ltudy ofpoIt-nataldtprmion.
Fourth year honours dissertation, Department of Psychology, University of Edinburgh.
Stern, D. (1993). The role of feelings for an interpersonal self In U. Neisser (ed.), The Ptruivtd
St/f(pp. 205-15). Cambridge: Cambridge University Press.
Stern, D. N. (1985). The Inmpmonal World of tht Infant: A Vitw from Prychoanalysu and
Deuelopment Prych%gy. New York: Basic Books
Stern, D. N., Hofer, L., Haft, w., & Dore, J. (1985). Affect atrunement: The sharing offeeling
states between mother and infant by means of intermoda! fluency. In T. N. Field and
N. Fox (eds.), SorialptrcqJtion in infant;. Norwood, NJ: Ablex.
Stern, D. N., Speiker, R. K., & Mackain, K. (1982). Intonation contours as signals in maternal
speech to prelinguistic infants. Deuelopmenta! Prych%gy, 18(5),727-35.
Trehub, S. E., & Trainor, L. J. (1993). Listening strategies in infancy: the roots of music and
language development. In S. McAdams and E. Bigand (eds.), Thinking in sound.
Oxford: Oxford University Press.
Trehub, S. E., Trainor, L. J., & Unyk, A. M. (1993). Music and speech processing in the first year
of life. Advanw in Child Development and Behaviour, 24, 1-35.
Trevarthen, C (1980). The foundations of intersubjectiviry: development of interpersonal and
cooperative understanding in infants. In D. Olson (ed.), The socialfoundation oflan-
guagt and thought. New York: Norton,
Trevarthen, C (1986). Form, significance and psychological potential of hand gestures of infants.
In J.-L. Nespoulous, P. Perron and A. R. Lecours (eds.), The biologicalfoundation Ofgtl-
tum: Motorand semiotic mptctl (pp.149-202). Hillsdale, N]: Erlbaum.
Trevarrhen, C. (1988). Sharing makes sense: Intersubjectiviry and the making of an infant's
meaning. In R. Steele and T. Threadgold (eds.), Languagt Topia: £Slayl in Honourof
Micbae! Hal/iday (pp, 177-99). Amsterdam and Philadelphia: John Benjamins.
Trevarthen, C (1998). The concept and foundations of infant inrersubjecriviry, In S. Briten (ed.),
Intersubjectio« Communication and Emotion in Early Ontogmy(pp, 15-46). Cambridge:
Cambridge University Press.
Trevarthen, C, & Burford, B. (1995). The central role of parents: how they can give power to a
motor impaired child's acting, experiencing and sharing. European JournalofSptrial
Nm/s Education, 10(2),138-48.
Trevarthen, C, Murray, L., & Hubley, P. (1981). Psychology of infants. InJ. Davis andJ. Dobbing
(eds.), SritntificfoundatiOn! ofc/inicalpatdiatricr (2nd edition). London: W. Heinemann
Medical Books Ltd.
Tronick, E. Z., & Weinberg, M. K. (1997). Depressed mothers and infants: Failure to form dyadic
states ofconsciousness. In L. Murray and P. J. Cooper (eds.), Postpartum Depression and
Child Development (pp.54-81). New York: Guilford Press.
Tronick, E. Z., A1s, H., & Brazelton, T. B. (1980). Monadic phases: A structural descriptive analysis
of infant-mother face-to-face interaction. Merrill-Palmer Quarurly of &haviour and
Dtvt/opmmt, 26, 1-24.
van Rees, S., & de Leeuw, R. (1987). Born Too Early: TheK4ngaroo Mtthod with Prematur« Babie:
Video by Stichting Lichaarnstaal, Scheyvenhofweg 12, 6093 PR, Heythuysen, The
Netherlands.

53

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

Weinberg. M. K., & Tronick, E. Z. (1994). Beyond the face: An empirical study of infant
affective configurations of facial. vocal, gestural and regulatory behaviors. Child
Development, 65. 1503-15.

54

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 Mothers and Infants and communicative musicality
STEPHEN N. MALLOCH

• Madres e hijos y la musicalidad comunicativa

Utilizando la rnuslca como modelo, examinamos las vocalizaciones madres/hijos

utilizando anallsls acusticos basad05 en procedimientos lnformatlcos,
Primeramente, resumimos la investigaci6n anterior, que demuestra la importancia
de ambas partes en la comunicaci6n madre/hijo. Nuestros metodos son utilizados
para analizar entonces pulso, calidad y narrativa en las vocalizaciones madres/hijos.
Estos tres elementos comprenden la "musicalidad comunicativa" : aquellos atributos
de la comunicaci6n humana que son particularmente explotados por la muslca que
permite surgir la compafila coordinada. EI pulso es estudiado mediante analisis
espectrograficos, descubriendose que 105 intervalos regularesde tiempo sirven para
coordinar el conjunto de vocalizaciones de madres e hijos. La calidad consiste en el
estudio del espectro de alturas de las vocalizaciones y su timbre. EI estudio de la
altura es posible gracias a un software desarrollado por nosotros, que utiliza una
constante 0 de transformaci6n espectral. Examinamosc6mo los ninos y las madres
estructuran su exploraci6n total del espacio de alturas a pequefia y a gran escala.
EI timbre se describe mediante una variedad de medidas acusticas - agudeza,
nitidez, dureza y amplitud. Constatamos los cambios de calidad de la voz de la
madre en respuesta a 105 del nino. La narrativa combina pulso y calidad, 10 que
permite a dos personas compartir un sentido del paso del tiempo. Esto se i1ustra
mediante el examen de /a compania musical que secrea entre una madre y su bebe
cuando ella canta 'una canci6n de cuna. Concluimos que la musicalidad
comunicativa es vital para la cornunlcaclon padres/hijos.

• Madri e bambini e la musicalitA comunlcativa

Usando la musica come modello, esaminiamo la vocalizzazione madre/bambino

utilizzando analisi acustiche computerizzate. Innanzitutto riassumiamo Ie ricerche
passate, che dimostrano l'importanza di entrambe Ie parti nella diade madre-
bambino. Introduciamo quindi i nostri metodi per analizzare pu/sazione, qua/ita e
narrativita nelle vocalizzazioni madre-bambino. Ouesti tre elementi costituiscono la
"musicalita comunicativa" : gli attributi cioe della comunicazione umana,
particolarmente sfruttati in musica, che fanno nascere una sintonia coordinata. La
pulsazione viene analizzata utilizzando analisi spettrografiche, e vengono scoperti
intervalli temporali regolari che servono a coordinare Ie vocalizzazioni di madre e
bambino. La quallta consiste sia nell'altezza delle vocalizzazioni, che nel lora
timbro. Diagrammi dell'altezza dei suoni vengono otten uti con un software da noi
sviluppato, che utilizza una costante trasformazione spettrale (O-spectral
trasform). Esaminiamo come il bambino e sua madre strutturano la lora
esplorazione congiunta dello spazto sonoro su piccola e grande scala. II timbro
viene descritto con una varlets di misure acustiche - valori di acurezza, intensita,
durezza e altezza. Abbiamo riscontrato che la voce della madre modifica la propria
qualita in funzione di quella del bambino. l'elemento narrativo combina insieme

55

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

pulsazione e qualita : esso permette a due persone di condividere la sensazione del
trascorrere del tempo. Ci6 viene iIIustrato dall'esame della sintonia musicale che si
crea tra una madre ed iI suo bambino quando lei gli canta una filastrocca.
Condudiamo che la musicafitA comunicativa il vitale per un'arrnoniosa comunicazione
genitori/bambino.

• Mere et nourrisson : la muslcalite de communication

Prenant appui sur la musique, nous analysons par ordinateur les elements
acoustiques des vocalisations entre mere et nourrisson. Aprils un rappel des
travaux qui ont eclaire Ie rOle de chacun des parten aires dans la dyade mere-
nourrisson, nous presentons les methodes sur lesquelles nous fondons I'analyse de
la pulsation rythmique, de la quallte et de la narration des vocalisations entre
la mere et son hebe. Ces trois elements constituent ce que nous qualifions de
"musicalite de communication" : les attributs de la communication humaine,
slngullerernent exploltes en musique, qui autorisent l'ernergence d'une camaraderie
coordonnee. L:analyse spectographique de la pulsation rythmique revele la
presenced'intervalles temporels regullersautorisant la coordination des vocalisations
communes de la mere et du nourrisson. La quallte se rapporte au contour de
hauteur des vocalisations et a leur timbre. Le trace des hauteurs est derive par un
logiciel developpe par transformation spectrale de la constante Q. Nous etudlons
comment Ie nourrissonet sa mere structurent, aux echelles macro et microscopiques,
leur exploration mutuelle de I'espacedes hauteurs. Le timbre est decrit par diverses
mesures acoustiques qui constituent les trois valeurs du stimulus : hauteur,
approximation et ampleur. II est a noter que la mere modifie la qualite de sa voix
en reponse a celie de I'enfant. La narration associe pulsation rythmique et quahte:
elle permet a deux individus de partager Ie sensde l'ecoulernent du temps, comme
Ie montre l'etude de la camaraderie musicale qui s'instaure entre une mere et son
bebe lorsqu'elle lui chante des comptines. Nous en deduisons que la musicalite est
vitale a I'attrait de la communication entre l'un des parents et le nourrisson.

• MOtter und Klelnkinder und die kommunikative Musikalitat

Mit Musik als unserem Modell untersuchen wir auf der Basis computerunterstotzter
akustischer Analyse Mutter/Kleinkind·Vokalisationen. Wir geben zunachst eine
Zusammenfassung der bisherigen Forschung, welche die Wichtigkeit beider Teile
in der Mutter/Kind-Dyade zeigt. Unsere Methoden werden dann fOr die Analyse
von Puis, Qualitiit und Narration der Mutter/Kind-Dyade eingesetzt. Diese drei
Elemente bilden die "kommunikative Musikalitiit", namlich jene Attribute der
menschlichen Kommunikation, welche besonders in der Musik ausgenOtzt werden

56

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014

 Mothersand Infants and communicative musicality
STEPHEN N. MALLOCH

und die eine koordinierte Partnerschaft aufzubauen erlauben. Puis wird mittels
spektrographischer Analyse untersucht, und es werden regelrnalllge Zeitintervalle
entdeckt, welche der Koordination der StimmauBerungen von Mutter und Kind
dienen. Qualitat besteht sowohl aus der Tonhohenkontur der StimmauBerungen
als auch aus deren Klangfarbe. Tonhohenplots werden mittels einer von uns
entwickelten Software abgeleltet, welche eine konstante Q-Spektraltranstormation
verwendet. Wir untersuchen, wie das Kleinkind und die Mutter ihre gemeinsame
Erkundung des Tonraumes im kleinen und groBen strukturleren, Klangfarbe wird
beschrieben durch eine Mehrzahl akustischerMeBdaten: Tristimuluswerte, Scharfe,
Rauhigkeit und Weite. Wir finden, daB die Stimme der Mutter in Reaktion aut jene
des Kindes ihre Qualitat andert. Narration verbindet Puis und Qualitll.t: sie
ermOglicht zwei Personen einen gemeinsamen Sinn tur den Zeitablaut. Dies wird
durch die Untersuchung der musikalischen Partnerschaft iIIustriert, welche
zwischen einer Mutter und ihrem Baby beim Singen eines Kinderreims erzeugt
wird. Wir kommen zu dem SchluB, daB die kreative Musikalitat fur eine harmonische
Eltern/Kleinkind-Kommunikation grundlegend ist.

57

Downloaded from msx.sagepub.com at Yale University Library on October 26, 2014