Content uploaded by Katharine Parton
Author content
All content in this area was uploaded by Katharine Parton on Feb 20, 2014
Content may be subject to copyright.
The Second International Conference on Music Communication Science, 3-4 December 2009, Sydney, Australia
http://marcs.uws.edu.au/links/ICoMusic09/index.html
Features of conductor gesture: Towards a framework for analysis within
interaction
Katharine Parton and Guy Edwards
School of Languages & Linguistics
The University of Melbourne
ABSTRACT
This paper presents the results of a study of conductor
gesture and examines the effect of that gesture on
musician action within a tertiary music rehearsal and
pedagogical context. This study follows experimental
work on conductor gesture and attempts to complement
the results of experimental studies by using a corpus of
naturally-occurring video recorded data of conducting
classes. The aim is to determine whether the physical
actions (gestures) of conductors can be seen to be
systematically related to musician action (playing of
instruments) in such naturally occurring data. In addition,
we explore how quantitative analytical techniques, more
familiar to experimental approaches, can be used as the
basis for a framework for examining conductor gesture in
data collected using ethnographic research approaches.
This study concludes that conductor gesture has a
significant effect on musician action, and suggests some
salient features of conductor gesture.
1. INTRODUCTION
In contemporary Western art music tradition, the
conductor, and the practice of conducting, has attained
unparalleled levels of esteem and prestige. Conductors are
essential to the successful performance of much Western
art music, and serve as a critical locus of temporal and
artistic coordination and direction (Berlioz [1948] 1965,
Stokowski [1943] 1965). Conductors perform these
functions through complex physical actions, involving
(most saliently) the hands, arms, gaze, facial expressions
and (in many cases) the baton, but also extending, at times,
to whole body movements.
Conducting can therefore be said to be a practice,
achieved within a particular social and cultural context,
whereby specific communicative functions are performed
through physical action (gesture). Furthermore,
conducting is explicitly intended to produce a desired
change or effect in the orchestra (or equally, other musical
ensemble). This paper aims to use a quantitative approach
to evaluate conducting as a communicative practice, and
to propose some of the salient features of conductors’
gestures.
Critically, this paper argues that conducting should be
examined through ethnographically valid data, alongside
experimental studies of musician-conductor
communication. Effectively, therefore, this study is a
corpus-based approach to studying conductor gesture as it
occurs in actual rehearsal contexts.
2. BACKGROUND
The examination of conductor gesture, and other musician
gesture, has been approached from a number of different
perspectives and methodologies. There are two main
traditions which inform this study; firstly, the examination
of conductor gesture from a psycho-linguistic and
psychological perspective, and secondly the application of
ethnographic and anthropological approaches to musician
communication. In addition, this study is critically
situated within the study of gesture as a communicative
act. It is to this aspect, gesture as situated action, that we
turn now, prior to a discussion of recent developments in
the study of musician and conductor gesture.
Gesture and physical actions are performed by individuals
to achieve particular goals and outcomes. In social
situations – that is, situations in which the mutual
monitoring of physical and verbal actions is possible
(Goffman 1997) – gesture and physical action co-occur
with speech, forming a cohesive system (McNeill 1985,
1992, 2000). Physical actions can be seen to be situated
and dependent on the environments in which they occur
(Goodwin 2003). These ‘situated actions’ (Goodwin 2000,
2003) are performed by participants relative to the
environment in which they occur; gestures are constituted
in terms of the physical world, and also in terms of the
shared social understandings held by participants that
create social context. The social construction of embodied
and situated actions is demonstrated, for example, by
Ochs, Gonzales & Jacoby (1997) describing the social
interactions of physicists that include tools and physical
objects such as diagrams, whiteboards and pens, and
Suchman (1996), who argues that airport control rooms
are shared workplaces created by the physical and social
interaction of individuals.
Conductor gesture, and the practice of conducting, has
been the subject of extensive study by musicians and
musicological theorists; this includes conducting manuals
and textbooks (such as Green 1996, Rudolf 1994, among
many others), and more abstract discussions of the
function and purpose of conductors (such as Berlioz [1948]
1965). Typically, this work is focused on instructional and
educational outcomes, and is of limited use in evaluating
how conducting functions as a communicative practice.
Some work has been proposed towards ‘pragmatics’ and
‘semiotics’ of conducting (such as that cited in Luck &
Nte 2008). A study by Boyes-Braem & Bräm (2000)
examined ‘expressive’ gestures made by conductors,
using an expert informant to code such gestures for
meaning and comparing the form of conductor gesture to
standardized sign-language. Similarly, Poggi (2002)
The Second International Conference on Music Communication Science, 3-4 December 2009, Sydney, Australia
http://marcs.uws.edu.au/links/ICoMusic09/index.html
undertook a study of conductors’ facial expressions, with
the aim of developing a lexicon, again based on a single
informant. Parton (2007) undertook a study of conductor
gestures according to McNeill’s continua (McNeill 2005),
suggesting that conductor gestures show some consonance
with gestures produced alongside speech, and with
normative human gestures. These studies (excluding Luck
& Nte 2008, which we discuss further, below) are
generally limited to qualitative micro-analysis of
synchronic examples of conductor gesture, or to particular
features across a set of gestures. These studies, however,
clearly situate conductor gesture within a broader field of
the study of interaction and communicative practices.
A study by Haviland (2007) of the multi-modal
interactional resources used by musicians in the context of
a chamber music master-class between a professional and
a student string quartet concluded that musicians deploy
multi-modal (gestural) communicative resources in, as
Haviland (2007) suggests, complex and situated ways. He
argues that musician multimodal gesture practices suggest
that musicians are creating meaning dynamically, and that,
for musicians, the distinction between cognition and
embodiment is (at least) awkward, if not entirely
problematic (Haviland 2007). Critically, this study, along
with similar studies of musician interaction (such as
Weeks 1996), are based on ethnographic methodologies
of examining naturally-occurring data, rather than
isolating particular variables within experimental or
otherwise manufactured conditions.
Experimental and more controlled studies of conductor
gesture and conducting practices have focused on
examining the effects of conducting gesture, including
both expressive gestures, and temporal gestures (typically
made with the right hand to provide the ‘beat’ and
temporal coordination across an ensemble). The
interaction of participants with novel musical technologies
(Less, World & Borchers 2005, Marrin & Picard 1998)
has provided a basis for examining ‘conductor-like’
interactions mediated by particular technologies. Similarly,
Kabisch, Williams & Dourish (2005) examined human
computer interaction (HCI) with ‘sonic environments’ and
other participants. Experimental approaches include
studies of temporal gestures and the effect of participant
education (Kelly 1997), expressive gesture effects
(Skadsen 1997) and motion-capture studies (Clayton
1986). Luck & Nte (2008) conducted an experiment
designed to test the effects of conductor temporal gestures,
using a complex motion-capture and response
measurement system. They found a significant effect on
participant synchronization with conductor gestures for
previous experience as a musician or a conductor, but no
significant effect for variation in gesture size.
There is a clear, and not entirely unexpected, divide
between studies of conductor gesture which use
experimental, robust approaches to measuring conductor
gesture in non-natural environments, and studies which
are based on qualitative analysis of naturally occurring
data. As noted by Luck & Nte (2008), experimental
approaches to the ‘applied phenomenon’ of conducting
are limited in that they exclude the interactive and situated
nature of the activity. Conversely, qualitative studies are
limited in the extent to which they can produce robust and
verifiable results. The intention of the present study is to
attempt to bridge this gap in the examination of conductor
gesture.
3. METHODOLOGY
As highlighted in the preceding section, this study is
(partially) an attempt to investigate and validate
experimental findings through analysis of
naturally-occurring data. The methodology, therefore, is
fundamentally oriented to the analysis of a corpus of
naturally-occurring conductor gesture; more detail on data
used in the study is given in the following section.
The aim of this study is to measure the effect of conductor
action on subsequent actions by musicians. The intention
furthermore is to examine this effect in naturally
occurring data; to determine what effect (if any)
conducting has on the in-situ behavior of musicians.
Further detail on conductor action and musician action
examined in this study is given in the Coding section,
below.
It is important to note that, as a fundamentally
corpus-based study, some research techniques commonly
used in experimental studies – notably the use of control
group(s) – cannot be applied. In coding and analysis, our
intention has been to explore a framework for quantitative
approaches to natural data in conductor gesture research.
Data
Data for this study were collected during tertiary-level
conducting classes. The classes are delivered by a
professional conductor; student participants in the class
conduct short excerpts of orchestral repertoire whilst
receiving instruction and observing demonstrations of
conducting technique. The students conduct small mixed
ensembles of between 4 and 10 players (varying
depending on the class); the players were of either senior
tertiary or professional musicians. In total, data for 30
conductors (29 intermediate student and 1 professional /
instructor) were collected. Data were collected using a
single video-camera positioned approximately five metres
directly in front of the conductor.
Coding
Data coding and measurement was conducted using a
specialized gesture analysis software program (Elan). The
data were separated into two sections.
Section A – Musician response to occurrence of
conductor gesture
Data for this section uses a particular section of music
conducted by the participants (Haydn Symphony No. 104,
bars 3-5 and 9-10). This section features 5 repeated
musical phrases, and the score carries no dynamic
markings. Participants could therefore choose to gesture
(using the left hand in an ‘expressive’ gesture) to indicate
a desired change in volume (dynamic) of ensemble
playing. Conductor left-hand gestures were coded as
either gesture (travelling or static) or no gesture. Musician
response was measured using the sound (waveform)
represented in ELAN, and coded as either change, or no
change. Each instance (N=92) is therefore described in
terms of a conductor gesture and a musician action. No
The Second International Conference on Music Communication Science, 3-4 December 2009, Sydney, Australia
http://marcs.uws.edu.au/links/ICoMusic09/index.html
coding was conducted of relative timing of any change in
volume to conductor gesture; musician response was
considered if it occurred within the musical phrase.
Section B – Varying conductor gesture and musician
response
This section focuses on musician responses to varying
conductor gesture at the start of playing. The conductors
used gesture to ‘cue’ or start the ensemble, using the
right-hand, normatively responsible for ‘beating’ and
temporal coordination. Conductor right-hand gestures
were coded for salient variations, as follows:
1. Size (SMALL, MID, LARGE)
2. Handshape (BATON, POINT, GRIP /
Other)
3. Body orientation (FRONT,
FRONT-LEFT, LEFT)
4. Position of gesture in relation to torso
(INNER, OUTER)
5. Accompaniment of gesture by left-hand
movement (NO ACCOMP, LEFT
HAND)
6. Arm extension (CLOSE, MID, FULL)
Musician response was measured by the lag from the
completion of the gesture stroke to the beginning of
audible sound, measured in milliseconds. 27 instances
were coded for this section.
4. RESULTS
Section A – Musician response to occurrence of
conductor gesture
The results of this analysis showed that conductor gesture
has a significant effect on musician responses.
Table 1: Musician response to gesture (N=92)
Conductor Gesture
Musician
response
No
musician
response
No gesture
2
36
Gesture (travelling or static)
43
11
Total
45
47
There is a systematic difference between musician
responses (no change compared to change) between
conductor gestures (Chi Square 49.401, p < 0.001).
Analysis was conducted to the level of whether a
conductor gesture was present. The results of this analysis
support the contention that the presence of a conductor
gesture was associated with a response from the
musicians.
Section B – Musician response to varying conductor
gesture
The results for this section are less robust, given a lower
number of instances (27) included in the data. This may
have contributed to non-significant results for variation in
lag times for size of gesture, hand-shape and arm
extension (close or far from the body). The co-presence or
absence of an accompanying left-hand gesture also did not
have a significant effect on musician start times.
Table 2: Gesture position and lag times (N=27)
Position of gesture
in relation to torso
N
Mean
lag
Std Dev
Inner
14
207.86
44.752
Outer
13
118.46
95.118
Significant results (using a t-test) were found for the
position of the conductor’s right hand in relation to the
torso (t = 3.163, p < 0.005) (Table 2). Right-hand gestures
made with the hand ‘within’ the torso were associated
with a longer average lag time (207.9ms) compared to
right-hand gestures made with the hand held out to the
side of the torso (average lag 118.5ms).
An unrelated ANOVA analysis found a significant
variation (F = 6.53, p = 0.005) in lag time for body
orientation. Body orientation was coded for front-facing,
front-left (45 degree left) and left (90 degrees left).
Further post-hoc analysis of this variation showed that
front-left orientation varied significantly (p < 0.05) from
front-facing and left-facing, but that there was no
significant variation between front and left (90 degree)
body orientations. No significant variation in lag time
from conductor gesture to musician action was observed
for changes in hand-shape, arm extension or
accompaniment of the right-hand gesture by a
simultaneous left-hand movement.
5. DISCUSSION
The results from this study show the expected result; that
there is a systematic relationship between conductor
gesture and the responses of musicians being conducted.
More importantly, however, the results of section A show
that the quantitative analysis of naturally occurring data
shows a similar (and expected) relationship between
conductor gesture and musician action as seen in
experimental studies of conductor expressive gesture.
The results of the second section are less conclusive; the
intention was to propose salient features of conductor
gesture that could be associated with a measurable
variable of musician response. The results suggest that
there are some features of conductor gesture that are
particularly salient; namely body orientation and gesture
position relative to the torso. The results for body
orientation are, to some degree, problematic, and show
unexpected differences in variation between front and left
orientations. The relative positioning of the gesture and
lag time, examined using a t-test, are more robust, and
suggest that variation in conductor gesture has an effect
on subsequent musician action.
As stated in the aim, the aim of this study was to explore a
framework whereby natural data – that is, a corpus of
‘real-life’ conductor gestures and musician actions –
could be used to validate the conclusions of similar
experimental studies of conductor gesture. The findings of
this study are, in many respects, expected; conductor
gesture is associated with responses from musicians, and
spatial orientation of conductor gesture affects how
musicians respond to that gesture. This paper has
attempted to mirror the results of experimental studies,
such as Luck & Nte (2008), using naturally occurring data
The Second International Conference on Music Communication Science, 3-4 December 2009, Sydney, Australia
http://marcs.uws.edu.au/links/ICoMusic09/index.html
to overcome the limitations of experimental research
design. Whilst the results of the analysis may be (to some
degree) problematic, it is clear that there is significant
potential to apply robust quantitative techniques to natural
data.
6. CONCLUSION AND FUTURE
DIRECTIONS
The framework used to approach conductor gesture and
musician action in this research is deliberately focused on
clearly observable physical change. We have not sought
to impose an aesthetic framework – for example,
examining tone quality or sonority of playing – in coding
musician responses, nor to categorise conductor gesture
within paradigms or lexicons of conducting technique.
The intention is to establish from natural data similar
conclusions about action and response as may be observed
in controlled conditions, and to begin to suggest some of
the salient features of conductor gesture that can be
identified within data constrained by an observational and
ethnographic research method.
There are clear future directions suggested as outcomes
from this research; the relationship between musician
dynamic change and conductor gesture is one which these
results clearly indicate as a possibly productive path. In
addition, the development of more accurate measurement
techniques for video-recorded (naturally-occuring,
non-experimental) motion data will enable more accurate
quantitative analysis of conductor and musician action.
Larger, and more complex, ensembles may also provide
richer data for similar research.
7. REFERENCES
1. Berlioz, H. (1965). On conducting [1948]. In The
conductor's art. Columbia: Columbia University
Press.
2. Boyes Braem, P., & Bräm, T. (2000). A pilot
study of the expressive gestures used by classical
orchestra conductors. In K. Emmorey, & H. Lane
(Eds.), The signs of language revisited. Mahwah,
New Jersey: Lawrence Erlbaum & Associates.
3. Goffman, E. (1997b). The neglected situation. In
C. Lemert, & A. Branaman (Eds.), The Goffman
Reader (pp. 229-233). Malden, MA: Blackwell.
4. Goodwin, C. (2000). Action and embodiment
within situated human activity. Journal of
Pragmatics , 32, 1489-1522.
5. Goodwin, C. (2003). The body in action. In J.
Coupland, & R. Gwyn (Eds.), Discourse, the body
and identity (pp. 19-42). New York:
Palgrave/MacMillan.
6. Green, E. (1996). The modern conductor. New
Jersey: Prentice Hall.
7. Haviland, J. (2007). Master speakers, master
gesturers. In S. Duncan, J. Cassell, & E. Levy
(Eds.), Gesture and the dynamic dimenstion of
langauge (pp. 147-172). Amsterdam: John
Benjamins.
8. Kabisch, E., Williams, A., & Dourish, P. (2005).
Symbolic objects in a networked gestural sound
interface. ACM Conference on Human factors in
computing systems.
9. Lee, E., Wolf, M., & Borchers, J. (2005).
Improving orchestral conducting systems in public
spaces. ACM Conference on human factors in
computing systems.
10. Luck, G., & Nte, S. (2008). An investigation of
conductors' temporal gestures and
conductor-musician synchronisation, and a first
experiment. Psychology of music , 36 (1), 81-100.
11. McNeill, D. (2005). Gesture and thought. Chicago:
The University of Chicago Press.
12. McNeill, D. (1992). Hand and mind. Chicago: The
University of Chicago Press.
13. McNeill, D. (Ed.). (2000). Language and Gesture.
Chicago: The University of Chicago Press.
14. McNeill, D. (1985). So you think gestures are
nonverbal? Psychological Review , 92, 350-371.
15. Parton, K. (2007) Conductor gesture and
McNeill’s continua Paper presented at ICoMCS,
University of New South Wales.
16. Poggi, I. (2002). The lexicon of the conductor's
face. In P. McKevitt, S. Ó Nualláin, & C. Mulvhill
(Eds.), Language, vision and music (pp. 271-284).
Amsterdam: John Benjamins.
17. Rudolf, M. (1994). The grammar of conducting
(3rd Edition ed.). Toronto: Schirmer Books.
18. Suchman, L. (1996). Constituing shared
workplaces. In D. Middleton, & Y. Engestrom
(Eds.), Cognition and communication at work (pp.
15-34). Cambridge: Cambridge University Press.
19. Weeks, P. (1996). A rehearsal of a Beethoven
passage: An analysis of correction talk. Research
on Language and Social Interaction , 29 (3),
247-290.
