Exp Brain Res (2008) 184:599–603
DOI 10.1007/s00221-007-1256-x
R ES EA R C H N O T E
Manual actions aVect vocalizations of infants
Paolo Bernardis · Arianna Bello · Paola Pettenati ·
Silvia Stefanini · Maurizio Gentilucci
Received: 14 June 2007 / Accepted: 12 December 2007 / Published online: 9 January 2008
© Springer-Verlag 2007
Abstract Upper limb gestures, as well as transitive
actions (i.e. acted upon an object) when either executed or
observed aVect speech. Broca’s area seems to be involved
in integration between the two motor representations of arm
and mouth (Bernardis and Gentilucci, Neuropsychologia,
44:178–190, 2006, Gentilucci et al., Eur J Neurosci,
19:190–202, 2004a, Neuropsychologia, 42:1554–1567,
2004b, J Cogn Neurosci, 18:1059–1074, 2006). The relevance of these data is in relation with the hypothesis that
language evolved from manual gestures, and was gradually
transformed in speech by means of a system of dual motor
commands to hand and mouth (Gentilucci and Corballis,
Neurosci Biobehav, Rev 30:949–960, 2006). The present
study aimed to verify whether this system of integration
between gestures (and transitive actions) and speech is
involved also in the language development of infants.
Vocalizations of infants aged between 11 and 13 months
were recorded during both manipulation of objects of
diVerent size and request arm gestures towards the same
objects presented by the experimenter. Frequency in voice
spectra increased when the infants manipulated or gestured
to large objects in comparison with the same activities
directed to small objects. These data suggest that intrinsic
properties of an object evoking commands of manual interaction are used to identify that object, and to communicate.
P. Bernardis · A. Bello · P. Pettenati · S. Stefanini ·
M. Gentilucci (&)
Dipartimento di Neuroscienze, Università di Parma,
Via Volturno 39, 43100 Parma, Italy
e-mail: [email protected]
A. Bello · P. Pettenati · S. Stefanini
Centro Studi sullo Sviluppo Motorio e Linguistico del Bambino,
Università di Parma, Via Volturno 39, 43100 Parma, Italy
Keywords Hand–arm action · Gesture · Speech · Voice
formants · Infants · Language development
Introduction
The origin of human language continues to pose a major
challenge to contemporary neuroscience. In one view, spoken language might have evolved from a communication
system employing manual gestures (Corballis 2002; Gentilucci and Corballis 2006; Hewes 1973; Rizzolatti and Arbib
1998). This hypothesis is primarily supported by the discovery of two classes of neurons in monkey premotor cortex. The Wrst class discharges when the animal executes a
manual action as well as it observes the same action (i.e. the
grasp) performed by another individual (Gallese et al.
1996). During evolution this mirror system might have
acquired the function of interpreting and sending messages
with the arm (Rizzolatti and Arbib 1998). The second class
discharges when the same action (i.e. the grasp) is performed with the arm or the mouth (Rizzolatti et al. 1988).
This system established initially in the context of the acts of
grasp and ingestive movements, adapted later for communication. In other words it may form the basis for the transfer
of a communication system from hand to mouth.
By means of the mirror system, humans can understand
action meaning by activating the motor representation of
the observed action and might have used this gesture repertoire to exchange information. Using the dual hand-mouth
command system this arm gesture repertoire could be
shared with and transferred to the mouth (Gentilucci and
Corballis 2006). EVects of these two systems have been
observed in modern humans. Indeed, it was found that the
execution and the observation of the grasp of objects
inXuence the simultaneous pronunciation of syllables
123
600
(Gentilucci 2003; Gentilucci et al. 2001; Gentilucci et al.
2004b). SpeciWcally, when executing or observing the
grasp of large as compared to small objects the lip opening
and formant 1 (F1) in the voice spectra increased, corresponding to variation in Wnger shaping during these grasp
movements (Gentilucci 2003; Gentilucci et al. 2001; Gentilucci et al. 2004b). These eVects were not restricted to the
grasp action. Indeed, also the execution/observation of the
action of bringing large and small fruits to the mouth
aVected the simultaneous pronunciation of syllables. The
execution/observation of the action guided by large fruits
induced an increase in formant 2 (F2) in comparison with
the execution/observation of the same action guided by
small fruits (Gentilucci et al. 2004a,b).
Gentilucci and colleagues (Gentilucci et al. 2004b) discovered that the eVects of observation/execution of transitive actions on speech were greater in children (6–7 years
aged) than in adults. On the basis of this result, we hypothesized that the activity of the mirror and the dual hand–
mouth command systems could be involved in the Wrst
stage of language development. This is in accordance with
the evidence that a strict relationship exists between early
speech development in children and several aspects of manual activity. For example, canonical babbling in infants
aged 6–8 months is generally accompanied by rhythmic
hand movements (Iverson and Thelen 1999). In addition,
spoken language and manual gesture develop in parallel
during the following stages of communicative and linguistic acquisition. Between 8 and 11 months, typically developing infants start to show systematic evidence of word
comprehension (Caselli et al. 1995). This phenomenon is
correlated with increasing capacity to execute explorative
and manipulative play with objects and with the emergence
of deictic gestures (request, giving, showing, declarative
pointing), that precede and accompany the appearance of
the Wrst words (Volterra et al. 2005; Bates and Dick 2002).
The deictic gestures are used to draw attention to objects,
locations or events (Capirci et al. 2005). The infant’s communicative intent to request is conveyed by request (hand
with opened Wngers) or pointing gestures. Frequently,
request gestures are accompanied by vocalizations (Bates
1976): Pizzuto et al. (2005) found that these vocalizations
are temporally coordinated with gesture execution. However, up to now the functional relationships between these
vocalizations and gestures have not been investigated.
Summing up, language development seems to be more
strictly related to manual activity mainly in the age approximately between 8 and 13 months, i.e. before and at the
beginning of verbal production. At this age infants become
able to use appropriately the objects and to execute deictic
gestures in order to communicate their requests.
The hypothesis that manual interactions with objects contribute to language development predicts that vocalizations of
123
Exp Brain Res (2008) 184:599–603
infants aged between 8 and 13 months are aVected by the size
of simultaneously manipulated objects. In other words, the
results of the manual explorative activity, i.e. intrinsic properties related to interactions with objects, can aVect vocalization and used to identify an object. In addition, if these
properties are object of communication signal, they can aVect
also vocalizations, which accompany request arm gestures.
Methods
Nine typically developing infants (four females and Wve
males) recruited from nursery schools in Parma participated
in the experiment. The Ethics Committee of the Medical
Faculty of the University of Parma approved the study.
Data obtained by the Italian version of the short Infant
form of MacArthur-Bates Communicative Development
Inventory (CDI) questionnaire Wlled in by the infants’ parents (Caselli et al. 2006) are reported in Table 1. Data were
used to establish the stage of children language acquisition:
according to the number of produced words two infants
were in preverbal phase, whereas seven infants were in
emergent lexicon phase (Clark 2003). All infants were able
to execute complex actions with objects. Request gestures
were observed in all the infants, whereas declarative gestures (i.e. pointing gestures by which infants share attention
and interest with adults to objects and events; Liszkowski
et al. 2004) were observed in seven infants.
The infants were tested at their homes. During the experimental session the infant sat on a child-chair in front of a
table where the stimuli were presented; the experimenter and
her/his mother sat behind the table in front of the infant. The
stimuli were wooden objects (cube, cylinder, ball, cup and
car) of two sizes: small object (side of 2 cm) and large object
(side of 4 cm). The procedure was the following: the experimenter showed the object to the infant and manipulated it
(mean duration 1.8 s, range 1–2 s). Only one object was presented at time, whereas the others were kept hidden in a box.
Objects were presented in random sequence. The object
attracted the attention of the infant, who was curious and
urged to get in contact with that by executing a request arm
gesture towards it. Then, the object was placed on the table
plane at reachable distance, and the infant after reachinggrasping, manipulated it with his/her hand(s) (mean duration
3.6 s, range 1–7 s). Then, a new object was presented.
A digital camcorder and a professional microphone were
used to videotape the experimental session. They were
placed at a Wxed distance of 2 m from the infant. The videotapes were transferred on a laptop and coded by two judges
in separate sessions. The judges marked each vocalization
produced by the infant while he/she manipulated or executed request arm gesture towards the object. Vocalizations
were not grouped into categories. Agreement between
Exp Brain Res (2008) 184:599–603
601
Table 1 Communicative and linguistic abilities of the infants tested in the present study
Name
Age
(months)
Comprehended
words
Produced
words
Produced actions
and gestures
Canonical
babbling
Post-canonical
babbling
Request
gestures
Declarative
gestures
1
I.D.
12
22
1
7
Yes
Yes
Yes
Yes
2
A.P.
13
20
2
11
Yes
Yes
Yes
Yes
3
N.M.
12
37
5
12
Yes
Yes
Yes
No
4
A.C.
12
18
8
7
Yes
Yes
Yes
Yes
5
S.P.
13
12
0
4
Yes
No
Yes
Yes
6
L.D.
12
49
2
10
Yes
No
Yes
Yes
7
A.F.
12
33
0
12
Yes
Yes
Yes
No
8
P.T.
11
35
5
7
Yes
Yes
Yes
Yes
9
V.M.
13
45
11
10
Yes
Yes
Yes
Yes
coders was about 98% in classifying the vocalizations and
95% in classifying the gestures. Only trials selected by both
the judges were included in the successive analyses.
In summary, we analyzed one set of vocalizations during
manipulation and one set during request arm gesture for
each of the ten diVerent stimuli (Wve objects £ 2 dimensions). The total vocalizations were 747: the pronounced
syllables were 75, which were not included in the analyses.
Indeed, the syllables are incomparable with simple vocalizations, since formant transition modiWes mean formant
values. The total number of vocalizations during the two
manual activities was 261: 31% were discarded because
they were emitted during request arm gestures directed to
unidentiWed objects inside the box. The mean number of
vocalization per infant selected in order to be included in
the successive analyses was 20, median: 18, range: 14–26
(vocalization number in the large object condition, mean
9.3, median 8, range 7–14; vocalization number in the
small object condition, mean 10.7, median 9, range 7–17).
From the audio track, we analyzed the voice spectrograms
of the vocalizations using the Praat software (www.praat.org).
We computed the following voice parameters: Intensity,
Pitch, F1 and F2. It is well known that voice formants are the
frequencies that exactly deWne a vowel from an acoustical
point of view (Leoni and Maturi 2002) and can be distinguished already in spectrograms recorded during spontaneous
infants’ vocalizations (Gilbert et al. 1997). We calculated the
Standard Deviation (SD) of each parameter. The voice
parameters and their SDs during manipulation and request
gesture were submitted to Analyses of Variance (ANOVAs)
whose within subjects factors were object size (small vs.
large) and activity (manipulation vs request arm gesture).
Results
Figure 1 shows typical examples of manipulation (Fig. 1a)
and of request arm gesture (Fig. 1b) of an infant who vocal-
ized during the two activities. The infants manipulated the
objects using one (51% out of the manipulations) or two
hands (49% out of the manipulations). About 53% out of
the manipulations of the small object were executed with
two hands (47% with one hand), and 46% out of the manipulations of the large object were executed with two hands
(54% with one hand). In other words, the infants used more
frequently two hands to manipulate small than large
objects. This indicates that the eVects on vocalization (see
below) were not due to commands to one and two hands
when manipulating small and large objects, respectively.
Mouthing, Wngering, shifting the object from hand to hand,
rotating, activities typical of infants aged between the 6–
12 months were also observed. However, vocalizations during these activities were not found. The infants performed
request arm gestures by extending only one arm towards
the object in most (91%) of the cases. All Wngers or only
the index Wnger (while the others were Xexed) were
extended. In both the two activities the infant while vocalizing directed always her/his gaze towards the object. However, the infants looked at the experimenter or at the mother
before or after vocalization in 26% out of the manipulation
activities, and in 17% out of the request arm gesture activities. In the remaining cases they Wxated the object till the
experimenter gave it to them.
Since the distribution of the F1 values was skewed
(1.07) we performed a logarithmic transformation on the
data, which reduced the skeweness to 0.4. The skewness of
the other analyzed parameters was less than 0.32. Figure 1
shows the main results of the statistical analyses performed
on voice spectra parameters. The size of the object aVected
F1 (Fig. 1c) during manipulation and gesturing
(F(1,8) = 11.4, P < 0.01, 2 = 0.135, f = 0.39). F1 was
higher when the activity was directed to large than to small
objects (mean 797 vs. 750 Hz, median 766 vs. 724 Hz,
range 458–1288 vs. 479–1244 Hz). In addition, F1
(F(1,8) = 4.7, P = 0.06, 2 = 0.095, f = 0.32 Fig. 1c) and
SD of F1 (F(1,8) = 5.2, P = 0.053, 2 = 0.036, f = 0.19,
123
602
Exp Brain Res (2008) 184:599–603
Fig. 1 Upper panels: examples
of manipulation (a) and request
arm gesture (b) of an infant tested in the present study, who
vocalized during the manual
activities. The object guiding the
motor activity was a large solid.
Lower panels: variation in F1 (c)
and Standard Deviations (SD) of
F1 (d) as a function of object
size when the infants vocalized
during manipulation of and request arm gesture towards small
and large objects. Note that logarithms of F1 are reported in c.
Error bars are standard errors
Fig. 1d) showed a trend to increase in request arm gesture
than in manipulation. The eVect on SD of F1 was signiWcant
for the large object (interaction between size and activity,
F(1,8) = 5.8, P < 0.05, 2 = 0.102, f = 0.34, Fig. 1d). Finally,
SD of the voice intensity was higher in request arm gesture
than in manipulation (F(1,8) = 9.1, P = 0.01, 2 = 0.096,
f = 0.32, 5.7 vs. 3.9 db). No signiWcance was found in the
ANOVAs performed on the other voice parameters.
Discussion
The vocalizations of the infants were modulated by the size
of the object during the activities of manipulation and gesturing. SpeciWcally, F1 was higher when the object was large
than when it was small. This eVect might be interpreted as
consequent only to the visual analysis of the object since the
infant Wxated the stimulus during the vocalizations. In other
words, it could depend on arousal (i.e., larger objects are
more interesting and therefore more arousing for infants).
However, if this were true, we should Wnd an eVect on all the
voice parameters and especially on voice intensity. This was
not observed. In addition, the results concerning SD of F1
indicate that voice modulation as a function of object size
was less variable during manipulation than during gesture.
This suggests a stricter relation between hand and mouth
when the hand directly interacted with the object. To explain
this relation, we propose that intrinsic object properties elic-
123
ited manipulation commands, which were sent to the hands
as well as to the mouth (Gentilucci et al. 2001). Moreover,
F1 is related to internal mouth aperture: that is, higher F1 is
associated to larger mouth aperture (Leoni and Maturi
2002). Consequently, the command to open the Wngers by a
larger amount (i.e. during manipulation of the large object)
was sent also to the mouth whose aperture was further
enlarged producing higher vocalization. Moreover, we
hypothesize that when an object held in the experimenter’s
hand was visually presented to the infants, manipulation
commands to both the hand and the mouth were activated by
a mirror system. They induced the same eVects on the voice
of infants gesturing to the object as they did on the voice of
infants manipulating the same object. However, the commands were not precise because the object was not held in
the infants’ hands and no somaesthesic information on
object size and/or no vision of their manipulating hand(s)
were provided. Thus, variability of mouth aperture and voice
increased (see results concerning SD of F1).
A problem concerning the data of the present study is
whether the infants’ arm movements directed towards the
objects were request gestures. According to Butcher and
Goldin-Meadow (2000), request gesture is deWned as the
infant’s arm movement during which he points to the object
and alternatively directs the gaze to the object and to the
adult who is able to satisfy his implicit request of possession.
In the present experiment, the gaze was only occasionally
directed to the experimenter or to the mother. However, the
Exp Brain Res (2008) 184:599–603
fact that the infant vocalized indicates that he/she attracted
the attention of the adult on the object pointed to and speciWed by voice parameters. In other words, from a functional
point of view vocalization might substitute gaze. In accordance with this hypothesis, Pizzuto and colleagues (Pizzuto
et al. 2005, see also Iverson et al. 1994) proposed that the
vocalization reinforces deictic gesture making the request
of attention more salient than when it is conveyed by gaze.
The fact that manipulation commands can be associated
to request gestures towards an object can be explained by
the proposal of Vygotskij (1934) that request gestures
derive from reaching to grasp movements, i.e. from actions
guided by object intrinsic properties like the manipulation
activity is (Lock 1978). The process of transformation from
reaching-grasping to request gesture can be brieXy
explained as follows: usually, when the infant fails to reach
and grasp an object, the adult automatically grasps and oVer
it to the infant. This, in turn, induces the infant to execute
the reaching-to-grasp gesture even towards unreachable
objects in the expectancy to receive them from the adult.
Through time, these gestures lose the feature of transitive
action and acquire a communicative intent.
The results of the present experiment are in agreement
with data of previous studies carried out on adults in which
the observation/execution of transitive actions (i.e. guided
by an object) aVected the voice of syllables pronounced
simultaneously with action execution (Gentilucci 2003;
Gentilucci et al. 2001; Gentilucci 2004a,b). Interestingly,
the observation of action pantomimes aVected the voicing
of syllables as the transitive action did (Gentilucci et al.
2004a). This suggested that the type of interaction of the
hand with the object, rather than the object per se, could
inXuence speech. Consequently, even if this is not testable
in infants, commands of manual interactions (i.e. the motor
representations of manual interactions) might be used as a
primitive form of object identiWcation. Once transformed in
vocalization they might be used to communicate with other
individuals (request arm gesture activity).
Acknowledgements We wish to thank the personnel of the nursery
schools (Educational Services of Comune di Parma) for their help in
contacting the families. We especially thank the infants who participated in the study and their parents. We are very grateful to Virginia
Volterra for discussion of the data and comments on an early version
of the paper. We thank Aaron Shield for the comments on the manuscript. The work was supported by MIUR (Ministero dell’Istruzione,
dell’Università e della Ricerca) and by Fondazione Monte Parma.
References
Bates E (1976) Language and context. The acquisition of pragmatics.
Academic, New York
Bates E, Dick F (2002) Language, gesture, and the developing brain.
Dev Psychobiol 40(3):293–310
603
Bernardis P, Gentilucci M (2006) Speech and gesture share the same
communication system. Neuropsychologia 44:178–190
Butcher C, Goldin-Meadow S (2000) Gesture and the transition from
one- to two-word speech: when hand and mouth come together.
In: McNeill D (ed) Language and gesture. Cambridge University
Press, Cambridge, pp 235–257
Capirci O, Contaldo AM, Caselli MC, Volterra V (2005) From action
to language through gesture: a longitudinal perspective. Gestures
5:155–178
Caselli MC, Bates E, Casadio P, Fenson J, Fenson L, Sanderl L, Weir
J (1995) A cross-linguistic study of early lexical development.
Cogn Dev 10:159–201
Caselli MC, Stefanini S, Pasqualetti P (2006) The MacArthur-Bates
CDI in Italy: developmental trends, variability and clinical application. First European Network meeting on the Communicative
Developmental Inventories, Dubrovnik, Croatia, pp 51–61
Clark EV (2003) First language acquisition. University Press, Cambridge, MA
Corballis MC (2002) From hand to mouth: the origins of language.
Princeton University Press, Princeton, NJ
Gallese V, Fadiga L, Fogassi L, Rizzolatti G (1996) Action recognition
in the premotor cortex. Brain 119:593–609
Gentilucci M (2003) Grasp observation inXuences speech production.
Eur J Neurosci 17:179–184
Gentilucci M, Corballis MC (2006) From manual gesture to speech: a
gradual transition. Neurosci Biobehav Rev 30:949–960
Gentilucci M, Benuzzi F, Gangitano M, Grimaldi S (2001) Grasp with
hand and mouth: a kinematic study on healthy subjects. J Neurophysiol 86:1685–1699
Gentilucci M, Bernardis P, Crisi G, Dalla Volta R (2006) Repetitive
transcranial magnetic stimulation of Broca’s area aVects verbal
responses to gesture observation. J Cogn Neurosci 18:1059–1074
Gentilucci M, Santunione P, Roy AC, Stefanini S (2004a) Execution
and observation of bringing a fruit to the mouth aVect syllable
pronunciation. Eur J Neurosci 19:190–202
Gentilucci M, Stefanini S, Roy AC, Santunione P (2004b) Action
observation and speech production: study on children and adults.
Neuropsychologia 42:1554–1567
Gilbert RG, Robb MP, Chen Y (1997) Formant frequency development: 15 to 36 months. J Voice 3:260–266
Hewes GW (1973) Primate communication and the gestural origins of
language. Curr Anthropol 14:5–24
Iverson JM, Capirci O, Caselli MC (1994) From communication to
language in two modalities. Cognit Dev 9:23–43
Iverson JM, Thelen E (1999) Hand, mouth, and brain: the dynamic
emergence of speech and gesture. J Conscious Stud 6:19–40
Leoni FA, Maturi P (2002) Manuale di fonetica. Carocci, Roma
Lock (1978) Action, gesture and symbol. The emergence of language.
Academic, London
Liszkowski U, Carpenter M, Henning A, Striano T, Tomasello M
(2004) Twelve-months-olds point to share attention and interest.
Dev Sci 7:293–307
Pizzuto E, Capobianco M, Devescovi A (2005) Gestural-vocal deixis
and representational skills in early language development. Interact Stud 6(2):223–252
Rizzolatti G, Arbib MA (1998) Language within our grasp. Trends
Neurosci 21(5):188–194
Rizzolatti G, Camarda R, Fogassi L, Gentilucci M, Luppino G, Matelli
M (1988) Functional organization of inferior area 6 in the
macaque monkey. II. Area F5 and the control of distal movements. Exp Brain Res 71:491–507
Volterra V, Caselli MC, Capirci O, Pizzuto E (2005) Gesture and the
emergence and development of language. In: Tomasello M, Slobin D (eds) Elizabeth bates: a festschrift. Lawrence Erlbaum
Associates, Mahwah, NJ
Vygotskij LS (1934) Thought and language. MIT Press, Cambridge, MA
123