Ecole Normale Supérieure

Frontiers in Psychologie

“Extend your left hand as if it were a jellyfish”: The role of metaphors in piano teaching is explained by their activation of experiential states

Roberto Casati*, Institut Nicod, CNRS-ENS-EHESS, Paris, France
Robert Kaddouch, Ecoles Kaddouch and Music, Paris,

Figure 1
Les métaphores pédagogiques de Robert Kaddouch

Les métaphores de Robert Kaddouch
Les métaphores de Robert Kaddouch

Teaching a music instrument essentially involves linguistic communication. The use of metaphorical,
imaged language is widespread in teaching contexts. Directive metaphors abound, and are perceived as
useful by both teachers and pupils. This papers proposes that the function of directive metaphor is to
activate analogical experiential states that control the gestual and sound output. We classify 150
metaphors used by one of the authors (a pianist and music pedagogist) in teaching the piano, and analyse
in detail some of them. In the light of the features of the metaphors analysed, alternate accounts of
metaphor (cognitive linguistic approaches, relevance theory approaches) are criticized, as is the
hypothesis that other types of bodily competence, described in the sources of the metaphors used, are
directly recycled in musical teaching and learning.
Music teaching, music learning, teaching, metaphors, cognitive linguistics, relevance theory, piano
i. Introduction
“We teachers inevitably and constantly use metaphors to define the various ways of producing tone on
the piano. We speak of the fingers fusing with the keyboard, of “growing into the keyboard”
(Rachmaninov’s expression) as if the keyboard were resilient and one coud “sink” into it at will, etc. All
these extremely approximate definitions are without a doubt useful, they arouse the pupil’s imagination
and when accompanied by illustrations at the piano they help to develop his ear and his motor-sensory
mechanism, what we call the “touch””. (Neuhaus 1973: 62)
The main purpose of our paper is to investigate linguistic communication in piano teaching. Playing the
piano is an exceedingly complex activity, involving sensorimotor, intellectual and emotional dimensions;
an activity that takes years of intense study to learn to execute proficiently. Accordingly, similarly
complex is teaching to play the piano. Pianists deploy a variety of skills and specializations – one may be a competent jazz improviser and be unable to read a musical score; beginners differ from trained pianists; professionals from amateurs. In discussing piano teaching we set for a rather generous, disjunctive definition, including cases of average pianists, improvisers, professional performers.
There are physical and technological aspects in the background. The piano is a percussive instrument
which exploits a complex path to transmit the energy generated by body, arm, hand and finger
movements to a set of pre-tuned strings. The interface is a keyboard which associates keys to strings in a
one-to-one correspondence. This sets the piano apart from other instruments using strings, such as the
violin, insofar the tunings of the strings are pre-determined and need not be generated on the fly.
Moreover, more than one key can be depressed at any given moment, thereby generating chords. This
“digital” aspect of the piano is, however, only one half of the whole story. On the one hand, the gesture’s
force can be modulated, thereby modulating the volume of the sound produced by the strings. More
importantly, strings are by default silenced and are only free to resonate in specific occasions, i.e. when
thwacked by the hammer (at which point their dedicated damper is released) or when the tonal pedal is
depressed (all damper released). Depressing a key thus creates a sound that not only depends on the
corresponding string, but on whichever other string is not silenced (Rosen 1999), as those generate
partials (overtones) of the thwacked string. Mastering sound production means interiorizing sensorimotor
knowledge of the sound generated in correspondence to patterns of pressed keys, i.e. patterns of hand-tokeyboard
interaction. The distribution of tones and half-tones on the keyboard further conditions motor
possibilities and facilitates or hinders learning and execution. Some scales (C, D, Eflat, E) are composed
by two identical sequences of black and white keys (translational simmetry), others display more
complex patterns. The C scale is easy to remember (all and only the white keys), but it does not afford
tactile beacons for breaking down movements, precisely as it is composed of white keys only.
The possibility shaping one’s hand in time so as to create dynamic patterns of pressed keys is constrained
by a number of anatomical, motor and cognitive factors. All movements of the relevant body limbs
(torso, arm, hand, finger) are rotational, centered on the overarching articulation. Keys, on the other hand,
are linearly arranged, and are reached and depressed by translations. Thus a certain amount of learning is
devoted to making the appropriate reaching and pressing movements, mapping sequences of hierarchised
rotational movements into translations.
A number of cognitive factors constrain piano playing. These include – among others – memory and the
ability to memorize, motor control, processing of auditory feedback, the ability to imitate, the ability to
extract motor information from listening, the ability to read music, as well as the very general ability to
understand spoken language in instruction.
Among the prominent challenges for a pianist figure technical challenges such as the production of
arpeggios and scales, regularity of beat, thumb passage, extension of the hand, hand synchronization,
different tempos on different hands, differences in hand specialization, highlighting thematic lines within
chords, and speed, among others. Interpretive challenges are subtle and involve an interface with the
conceptual system. The pianistic literature is extremely wide and some specialization is commonplace (it
is normal for a pianist to be mostly expert in a given repertoire: the romantic period, contemporary music,
baroque music, cool jazz, etc.) to the detriment of other repertoires. The ability to improvise (requested in
some styles, such as jazz) calls for a further set of competences and constraints. Pianists may or may not
be able to easily transpose form one key to another. Finally, music notation is hard to decipher and
requires a specific training, akin to leaning to read written language (the main difference being that music
notation elicits motor representations of the limbs and hands, as opposed to the elicitation of some
phonetic representation in reading witten langage), and with the added complexity of presenting separate
directions for each hand in parallel (each hand generates a set of sounds). Music notation itself is a
complex representational artifact, well suited to certain kinds of music (baroque, classical, romantic), less
suited to the representation of jazz (Casati, Ms.).
This (partial) view of the mechanical and physical constraints, of the cognitive contraints, and of the main
challenges facing pianists, provides an idea of the formidable complexities of learning to play the piano.
These map onto complexities in teaching. Let us consider learning first.
The main purposes of learning a musical instrument may vary. The training required to become a
professional pianist is different from that required for becoming an average amateur pianist. Within the
set of professional pianists, different skills acquire saliency (accompanying pianists are trained to rapid
first-sight reading, for instance.) Motor and developmental factors, some of which are specific, constrain
the phasing of learning. Children learning is different from adult learning. Some of the difficulties in
adult learning may be similar to those encountered by adult learners of a second language, although there
are cases of successful late learners (Marcus 2012). Children present specific issues, such as motivation,
the ability to understand instructions, planning, and bodily development constraints – these may be partly
general, partly specific to the instrument. (Full-size violins are used by children at about age 10 to 12, but
a full size piano keyboard is used from the very beginning of piano learning.) Self-teaching is possible
and although its results may be non standard (as shown by the cases of Thelonius Monk and Martial
Solal), it does deliver an ability to play the piano. Further issues in learning/teaching concern the fact that
a certain amount of undoing is necessary: stereotypes, or the preference for certain chords and scales,
create attractors that may block further learning.
Another remarkable issue is the organization of study, considering that puplis have to work on their own
anyway. There is a tendency to repeat things one has already learned, with further reinforcement, creating
an ever widening gap with the execution of difficult passages one did not properly learn.
There are no shortcuts to piano learning and teaching.1 As the old adage says, the only way to get to
Carnegie Hall is “exercise, exercise, exercise”, and his holds for all levels of practice. Repetition and
focus on difficult passages are key factors. Why is this so? Piano teaching is teaching in an artificial,
highly cultural setting, in the sense in which teaching to read is, and as opposed to “teaching” to speak a
first language (and possibly to teaching to sing), for which we have solid biological bases to dig into, as
well as a biological “plan” that unfurls in time, dictating the steps of language acquisition (and, possibly,
spontaneous music performance). Indeed, first languages, and other languages acquired during the
developmental window, are not generally considered to be the object of formal teaching and learning;
they appear to be simply acquired, or develop, instead.2 This implies that in the case of learning to play an
instrument some sort of forced “recycling” of other biological bases or previously acquired competencies
is requested.
It follows from difference that one cannot simply elicit piano playing by exposing the pupil to instances
of piano playing. There is no biological plan for piano acquisition as there is for language acquisition.
(Trained pianists can retrieve a lot of information from mere exposure, e.g. to an audio recording, but this
presupposes a higher level of competence). If mere exposure does not work, other “direct” methods do
not fare any better, and in point of fact are not used. There is no such thing as “Disklavier teaching”,3 for
instance. One cannot simply put the hands of the pupil on a Disklavier keyboard and playback a piece on
the instrument, hoping to acquaint the pupil with a pattern of movements on the keyboard. Indeed, in this
case the pupil’s fingers would be moved, but will not move. And as the pupil is not generating any action,
she would not receive the appropriate feedback, and eventually cannot learn from this one-sided
interaction. Another unlikely teaching shortcut is direct gesture imitation. The teacher cannot simply
1 A lively review of music learning from a cognitive perspective is provided by Marcus 2012.
2 Pinker 1994 for an overview.
3 A Disklavier is a piano whose keyboard can be controlled by an electromagnetic device capable of depressing keys
according to some pre-set pattern.
show a pianistic gesture and ask the pupil to do the same, because there is an exceedingly large number
of aspects of the performed gesture that she would have to choose from. (Once more, trained musicians
can make a lot of the request to imitate a gesture, because they select from a limited set of alternatives.
Linguistically assisted imitation is on the other hand used in normal learning situations.) Neither would
do the imitation of the result (the produced sound), for much the same reasons, i.e. the redundancy of
aspects to choose from.
As a matter of fact there is no piano teaching, and in general no instrument teaching, that goes without
linguistic assistance. All piano lessons involve instructions formulated in a language that is mastered by
both teacher and pupil. A pupil attending a piano lesson offered in a language she does not understand
experiences all too predictable difficulties.
Where is the linguistic challenge for both teacher and pupil? It is first of all in the lexical limitation of
language relative to the large number of aspects of the action involved. Consider an analogy with
painting. Suppose teacher and pupil shared a vocabulary of million lexical entries for colors. Whenever
necessary, the teacher would simply tell the pupil to add a bit of color X in a certain area of the painting.
But the actual lexical palette is much more limited. Likewise, if the piano teacher and the piano learner
had a million gesture descriptors to choose from, the teacher would simply tell the pupil to perform
gesture Y in a particular context. But, as for color, the lexical palette for gestures is limited. Now, in the
case of color, painters can resort to compositionality (“reddish yellow”, “pale green”), to an extent; or
they can take a ride on other areas of lexical competence (“frog green”). These strategies are available to
teachers of matters that involve movement, to some extent. Some gesture descriptors can be created by
composition (“twisted elongation”), and some from other areas of lexical competence (“screwdriving”).
But the analogy with colors breaks down at interesting points. Colors have a relatively simple
mathematical structure that can be exploited in generating descriptions. The structure of gestures is much
more complex, multidimensional, and in the specific case of piano teaching it further involves the
interaction with the keyboard and above all the production of the desired sounds and the highlighting of
the relevant musical aspects. Not only there is a limited stock of basic gestural descriptors; the reach of
compositionality is much more limited.
The possibility of taking a free ride on other areas of competence (“move it like a crane’s arm”) may
appear more promising at first sight. As a matter of fact, there are obvious limits to this strategy as well.
Some of the movements lexicalized in other areas can be, to some extent, “recycled” directly; but most
movements one would like to teach (e.g., a salvo of fastly repeated notes using three fingers in rapid
succession) are specific to the piano. This brings us to our working hypothesis. In teaching,
communication oftentimes makes appeal to metaphors, analogies, similes and other imagistic language in
order to overcome the limitations of the lexicon for action. The interest of tapping into different areas of
lexical competence lies in the particular cognitive strategy it enforces. The effectiveness of metaphors,
similes and imagistic language in teaching is explained by the particular features of metaphoric
To recap, the two main premises of our argument are, first, that no non-linguistic teaching methods are
available for learning to play the piano (mere exposure to the music, Disklavier teaching, imitation) and,
second, that the lexicon does not have enough expressive power to represent all the relevant gestural
aspects. Now, if metaphors are an effective way to overcome the limits of linguistic communication, of
the linguistic expression of directions in music teaching, how do they bring about their results?
A series of research questions present itself immediately. Are metaphors normally used? Which
metaphors? How are they introduced and used? Are they effective? And, once it is recognized that
metaphors are used because they are effective, the main issue is to explain why metaphors are effective.
Observe, thirst, that language is used in different aspects of music practice, over and above teaching.
Metaphors and analogies are for instance used in the description of music and of musical performances.
To quote but one example, this is what Robert Schumann wrote about Domenico Scarlatti: “One has
difficulty in always following his music, because he quickly knots and then unties again his musical
threads… his style is brief, pleasing, and piquant… how clumsy is the form, how rudimentary the melody,
how limited the modulation!” (Schumann 1946: 87, our italics).
Furthermore, a great deal of metaphors are also “frozen” in the technical language used for dynamics and
expressivity. About 250 Italian terms are traditionally used in music, mostly for expressing tempo and
dynamics. At least one quarter of those terms have recognizably metaphorical content (e.g. “sospirando”/
sigh, “leggermente”/light, “con dolcezza”/sweet, “mesto”/sad.)4
Conversely, we should also note that there are other documented uses of metaphors in teaching, outside
music. For instance, wine appreciation heavily relies on metaphorical or analogical descriptions (Suárez-
Tosté 2007). The case is however different because of the absence of a clear directive component. Tasters
ought to be able to make correct/uncorrect judgments, but the teaching is mostly used for enriching their
experience, so that they can make ever subtler distinctions. In the case of piano teaching, the directive
component addresses the ability to perform the appropriate gestures, in order to produce the relevant
On top of being used as music descriptors, metaphors and figurative speech are implicitly or explicitly
recognized as an essential part of music training. Barten (1992, 1998) has documented several uses based
on her observation of music classes. She noted the pre-eminence of what she called “motor-affective
properties”, and she tentatively subdivided that domain in three sub-classes: (1) Movement and Actions,
(“the notes jump all over the place”, “this wants to go up there”), further subdivided in Type of
Movement, Speed; Specific Actions; Kinesthetic Experience, (2) Attitudes and Tendencies (“have some
courage”, “wistful”), further subdivided in Bearing, Emotion, Sense of tendecy; and (3) Human and
Nonhuman Acts (“keep it airborne”, “play it like a baseball”), further subdivided in Human Activities,
Animal Actions, Mechanical Actions and Movements in nature. Barten proposed three explanations for
the use of metaphors: their irreplaceability in labeling expressive or aesthetic properties; the discovery,
on instructors’ part, of their effectiveness; and their ability to integrate formal musical knowledge. We
should note that these three explanations are not mutually exculsive. For instance, metaphors may be
effective precisely because they are irreplaceable. Moreover, irreplaceability is just a redescription of the
centrality of metaphorical uses in the teaching setting. Why metaphors are irreplaceable (if they are)
remains an open question.
Metaphors have acquired center stage in cognition thanks to the seminal work of cognitive linguists, in
particular Lakoff and Johnson (1980), who claimed that they are central to categorizing and reasoning,
mostly by providing conceptual enrichment. They argued that metaphors are not a purely linguistic
phenomenon, but they directly operate at the level of, and affect, concepts (mental representations).
Metaphorical understanding is construed as a projection from a source domain, i.e. the set of concepts
that one understands and that provides the metaphorical vehicle: “youth is the spring of life”, to a target
domain one seeks to understand: “youth is the spring of life”. The sources are in most cases well
understood core “image schemas”, concerning core aspects of our bodily, spatial and perceptual
experience (Hofstadter and Sander 2013). In general, thus, projection from an experientially available
concrete source provides understanding of an abstract, or of a more complex, target.
Other influential accounts of metaphor, such as the account provided by the Relevance theory of
4 Our count, based on a list available at
communication, deny that metaphor is an essential property of thought, and propose instead that it is a
mostly a linguistic phenomenon (Sperber & Wislon, 1995; 2008). Moreover, according to Relevance
Theory, metaphor is not a sui generis communicative device, let alone a core cognitive ability, but rather
a form of loose talk, in smooth continuity with other forms of loose talk such as approximation and
Recently, various authors have suggested that cognitive linguistic and Relevance theory have to be
hybridized in some way to provide a full picture of metaphor (Ritchie, 2009; Tendahl & Gibbs, 2008;
Wilson, 2011). Relevance theory may in particular be essential to describe the heuristics used by the
receiver to understand a metaphor, something that cognitive linguistic seems unable to do.
According to Relevance theory, a receiver understands a communicative act using the tacit principle of
relevance: she takes it for granted that the producer tries to be relevant, i.e. tries to transmit an
information that will allow her to derive a set of cognitive benefits that justify the effort needed to derive
them. More precisely, she takes it for granted that the producer tries to be optimally relevant, i.e. tries to
allow her to gain a set of cognitive benefits while deploying the smallest possible cognitive effort. For
Relevance theory, an hybridation with cognitive linguistic or, more generally, with theories pertaining to
the trend of “grounded cognition” (Barsalou, 2008), seems essential in order to answer a key question: :
why are metaphors relevant at all? Why choose a metaphorical path instead of a more explicit form of
expression which, arguably, would require less cognitive effort to be deciphered? Why, for instance, say
that “my surgeon is a butcher” instead of enumerating explicitly the various implications of this
metaphor: “My surgeon is brutal, incompetent, you should not trust him, etc.”?
A recent answer is that metaphors are optimal vehicles to express “experience-like” mental states
(Pignocchi, submitted). As opposed to “language-like” mental states, experience-like mental states are
grounded in the sensory, motor, emotional and introspective systems. Experience-like mental states are
encoded in a continuous format, more akin to that of perception, emotion, images, than to the format of
words and sentences. As opposed to classical accounts of mental images (Fodor & Pylyshyn, 1988),
experience-like mental states are not inert images but combinable mental states that can participate in
sophisticated inferential chains and represent abstract concepts (Barsalou, 1999; Prinz 2002).
According to the proposed account of metaphors, claiming that a metaphor is relevant because it allows a
particularly efficient exteriorization of experience-like mental states, implies that “my surgeon is a
butcher” does not allow the receiver to derive a set of language-like implications (“my surgeon is
incompetent,” etc.) but rather a set of experience-like representations (an image of my surgeon with a
butcher knife in his hand and blood stains on his apron, etc.) that could not have been expressed using a
more explicit form of communication.
This account of metaphor casts light on the use of metaphors in music teaching. Arguably, the motor
plans involved in playing the piano – the subject-matter of our study – are encoded in an experience-like
format, where any slight modification can be important, as it can significantly, even dramatically affect
the sound outcome. Moreover, the relevant motor plans must be deeply entrenched with other experiencelike
representations, such as representation of the auditory output, sensations, impressions and emotions
associated to it, or some element of imagery, etc. (Hommel et al., 2001). Thus, it is arguably easiest both
for the teacher and his pupil to communicate about the relevant motor plans using a communicative tool
efficient for transmitting experience-like representations – i.e. metaphor, according to our account. Thus,
the challenge for the teacher is to produce a communicative act that exteriorizes an experience-like
mental state, in a way that allows the pupil to reconstruct, at the end of the process of understanding, an
experience-like mental state as close as possible to the original one; and a state that can guide action.
Since both the origin and the end point of the act of communication are experience-like representations, a
metaphor, according to our account, ought be the most efficient vehicle, indeed in most cases the only
one available for successful communication.
ii. Materials and methods
Would a qualitative analysis of instructional language cast light on the mechanisms of metaphorical,
imagistic communication, when language is used to direct and shape complex behavior?
The anecdotal origin of the study was indeed related to the puzzlement generated by an instructional
setting. Two of the authors of this text are an established pianist and pedagogist (Robert Kaddouch) and
an average pianist (Roberto Casati), respectively, who found themselves in a teacher-pupil situation at
some point (harmony and improvisation class). When analysing some of the pupil’s strength and
weaknesses in improvisation, the teacher mentioned the occurrence of “risées” (wind gusts) in the
improvised conduct of the pupil, suggesting that that particular pattern should be further encouraged. The
pupil understood that a certain set of semi-randomly generated sound patterns was intendend by “risée”,
that it could not constitute the general case of improvisation, but that it was important to use it from time
to time to add spice to the improvisation (“adding spice” being itself a metaphor, of course.)
In order to substantiate our hypothesis, we sought to collect comprehensive data from pupils and former
pupils of Robert Kaddouch.
31 pupils were contacted in one of the following ways: by e-mail, by phone call, and during lessons. 17
of them were still taking lessons at the time of the study, the others were former pupils. (Mean age: 50,6;
oldest pupil: 77; youngest: 16; M=17; F=14.) Five of them were music professionals; two were high
performing climbers (something that will have some importance in what follows); ten were teachers,
either of music or of other disciplines.
They were requested to recollect and indicate metaphors or figurative expressions that were used by R.
Kaddouch during his classes as instrumental to the learning process. We collected and analysed the
responses, that were often relatively long texts, with anecdotal elements added, by first extracting short
sentences or series of terms that could unambiguosly identify the metaphor and convey its intended
meaning. Upon analysis of the responses, we thus listed 200 items; the lot was further reduced to 157
because of redundancies, unclear responses, or non analogicy related character of the responses. We
analyzed the metaphors in two independent ways. Casati and Pignocchi), assisted by three coders, would
code metaphors based on their apparent content, without informing (BLINDED Kaddouch) of the norms
used; we called this “external analysis”; the pedagogist author would annotate them based on the problem
they were meant to solve (“internal analysis”). (See Appendix I for the set of responses, with
translations.) The rationale for this double approach is that we needed some context for the use of the
metaphorical device, in order to understand the logic presiding to its adoption.
External analysis was performed by clustering the responses according to what appeared to be the main
features of the metaphorical vehicle. We recruited three coders (postgraduate students at our home
institution) to assist in the classification of the analogical vehicle and reduce researcher bias. The brief for
coders was simple and easily understood: given a metaphor like “youth is the spring of life”, the
appropriate coding category was “seasons” (i.e., coders were supposed to focus on the vehicle or source,
not on the content or target.) We then analysed the answers by coders and found agreement on 6 macro
categories (mental activities, space, navigation and dynamics, human activities, nature, sound and music,
other), further subdivided in 19 categories (methodology, rhetorics, language / dynamics, space,
navigation, perspective / sport, arts and crafts, art, social worlds and daily life / body / natural
phenomena, animals / music instruments, sounds; gender, objects.)
The occurrence of each metaphorical/analogical vehicle metaphors in the various clusters is presented in
Fig. 1.
We organized fig. 1 so that it presents a spectrum of vehicles, ranging from mental activities to physical
phenomena. Fig 2 further presents the weight of each subcategory, organized by number of occurrences
of each analogical vehicle.
Most expressions are in a group that includes sports, dynamics and arts and crafts. In the light of the
account proposed, this is somewhat to be expected, given that what is taught is oftenitimes a complex,
hard-to-describe movement, or a strategy to accomplish a given result, and that movement notions are
already lexicalized in the vehicle areas (e.g. sculpting for arts and crafts.) Indeed, most of the expression
used draw analogically from other gestural and executive competences.
As an example, consider an expression like:
[Sport: surf] “Accepting the wave in order to orient one’s trajectory”.
The intended meaning indicates that one should not contrast one’s own gesture’s strength with rigidity.
The expression activates experience-like representations (imagined or acquired) of the visual,
proprioceptive, motor qualities of the surfer’s interaction with the wave.
We point out the experiential character of the representations activated by the metaphor. Our claim is that
these representations are put in the immediate service of the activation of motor plans that would be hard
or impossible to detail in a language-like format. The key idea behind an ideomotor approach to action
(Hommel et al. 2001) is that the planning of a pianist’s action is predominantly based on a distal
representation of the expected result (sounds, and then sensations and impressions that are associated to
the sound). The metaphor used by the teacher targets this very complex of representations. The pupils is
expected to execute a movement whose perceptual (proprioceptive, auditory, emotional) result will share
some abstract features with those evoked by an image of surfing.
In our set or responses, [Arts and crafts] is significantly represented. But with 44 occurrences, [Sports]
indeed takes up more than a fourth of the total of expressions used. When looking into further
subdivisions of the category, we find that skiing (n=9) mountaineering (n=9) and cycling (n=5) take the
lion’s share. This reflects environmental circumstances: the music pedagogist worked for a number of
years in a location close to the Pyrenees mountains, where this type of outdoor activities is widespread.
He could thus count on some shared ground to deliver the analogy. By contrast, the 7 analogies based on
aviation depart from the common ground schema. The music pedagogist is indeed an amateur pilot, but
the same should not be assumed for his pupils. It is interesting that those relatively technical analogies
appeared to be easily understood (as per subjective reports and pedagogist’s own assessment).
As we said, the “free ride” hypothesis is not expected to reach too far. The shapes a hand should form in
order to play a certain chord with a certain intensity may have very little to do with any other shape it
should form, e.g. for picking up pebbles. It is not an existing vocabulary of shape representations that is
Let us at present examine in some detail a few of the metaphors used in the light of the problem they
were supposed to address (these are examples of what we termed ‘internal analysis’). In square brackets
figures the assigned category. The original French texts, and a translation, follow.
[Animals] “Etend ta main gauche comme une méduse/Extend your left hand like a jellyfish”
Given the existing piano literature, and the intrinsic features of keyboards, each hand faces different
technical/interpretive challenges. While the right hand is prevalently used for melody, the left hand is
prevalently used for the harmony, and must oftentimes play chords (simultaneously pressing several
notes, which can be far apart from each other). Reaching out for each note in a chord requires extending
the hand in a supple way. How supple? There are indefinitely many ways of moving a hand so that its
movement could be considered “supple”. The jellyfish image gives a hint. A jellyfish displays a
particular pattern of movement; in order to move forward, it contracts its bell and then relaxes it. As they
happen in water, contractions and expansions are gentle. This kind of plastic, soft movement is
researched by the teacher. To that effect, some images are conjured up by the request to move one’s hand
as a jellyfish. It is not important that either teacher or pupil know precisely the anatomy of the jellyfish;
and in any event, the hand is not a bell and does not have coronal muscles to contract. What matters is
that when seeing one’s hand as a jellyfish the pupil forms a set of experience-like representations (mixing
visual, proprioceptive and motor properties) that she can immediately use to plan and control her actions.
For instance, jellyfish are round and move in that particular jellyfish-like way, hence the hand is expected
to form a round shape and imitate a jellyfish-like movement. Pulsations of the jellyfish’ bell are separated
by quiet intervals, hence a certain relaxation is induced.
One further use of the metaphor is its memorability. Long-term memory is more efficient if the
information to be retrieved has been treated in depth, so as to be associated with more traces. An
imagistic treatment – here the somewhat surprising association with an aquatic invertebrate – offers extra
handles for memory retrieval.
[Arts and crafts] “Laisse tes mains sur le dernier accord, car un capitaine ne quitte son navire que quand il
est arrêté / Leave your hands on the last chord, for a captain only abandons ship when it has come to a
Keeping the hands on the keys “opens up” the strings, thereby letting them resonate for longer. The
captain’s image conveys a sense of responsibility towards the sound, a sense of being in charge. Actions
are not just mechanical performances, they are controlled by intentions and attitudes, and many of the
actions we perform are involved in ethical patterns.
[Arts and crafts] “Le blues, c’est comme un steak tartare, pas un chevreuil au Grand Veneur / Blues is like
a steak tartare, not like game meat at a posh restaurant”
On this occasion the teacher takes serendipitous advantage of the fact that the pupil (a 75 year old late
comer to piano learning) had worked in a restaurant before retiring. The distinction between fast food and
haute cuisine is not meant to devaluate blues. It just indicates that there are codes appropriate to each
situation. The pupil was playing a blues from a score, as if it was a minuet by Mozart. The image brings
interpretation down to earth, inserting it in a pattern of social codes. (It is worth mentioning that this
particular image was pivotal in reorganizing the pupil’s attitudes towards learning in general.)
[Dynamics] “N’essaie pas de répéter tes notes, envoies une impulsion comme un ricochet / Do not try to
repeat your notes, send an impulse, like a ricochet”
With quickly repeated notes (sometimes hitting the speed of 10 notes per second) it is not possible to
form individual intentions about the execution of each single one of them. At the same time there is no
simple way to form a “collective” intention on the whole set of repeated notes. One should use deviant
strategies, involving the movement of the forearm and of the wrist. “Ricochet” is a way to describe this
dynamic request. The actual natural ricocheting of fingers is insufficient to create the desired effect –
finger are not sufficiently elastic. “Ricochet” is here a mental representation that orients the initial fall of
the hand, before it strikes the keyboard.
[Sports: tennis] “Le coup droit au tennis/The forehand at tennis”
In order to strike a good forehand, the arm must come back closer to the body in a single, uninterrupted
gesture. One should learn to continuously recycle forces, without ever stopping, and without closing the
movement too early. The starting point of a movement already depends upon the shape one wants to
impress upon the final part – after one has hit the ball. One should conceive of a gesture in its dynamic
wholeness in order to better control each of its phases. This image is in a certain sense complementary to
the previous one; but in both cases the request is to insert an action in a wider gestural context.
[Dynamics] “Ce n’est pas vrai que cela n’avance pas, c’est un mouvement stationnaire, un bouillonnement
/ It is not true that this is not moving forward, it’s a stationary movement, a boiling”
This image was used in interpreting repetitive musical phrases. (In a variant, it is said that the helicopter
that is in stillstand uses more energy than it needs to move forward in air.) The image conveys the idea
that even in the presence of repetitions, even in the absence of a constant flow of musical events, the
musical discourse moves forward. “Stationary mouvement” is an oxymoron – similar to the one used to
convey the idea that for playing softly one needs more energy than for playing strongly.
[Natural phenomena: matter] “La pédale droite est comme un diluant pour peinture / the right pedal is
like a thinner for paint”
The use of the right pedal is an essential technique for manipulating sounds. Depressing the right pedal
releases all the dampers on the strings. All the harmonics of the whole set of strings are then available.
The effect is a dilution of sound. The right pedal is to musical treatment what a diluter is to paint. It
enhances certains sounds, combines others among them; and one can to some extent control the degree of
dilution. Known that it is to be seen as a diluter induces an immediate attention to modulation, and the
request to control the resulting sound in terms of losing sight of the single components.
[Sport, climbing] La lecture de la parois / reading the wall
Over and above each single gesture there is a dynamic environment, related to the need to recycle energy
to produce another, ever changing gesture, because the latter is preceded or followed by yet another
move. This activity generates an understanding of fingering, i.e. the ability to predict a path on the
keyboard in terms of succession of fingers. We have only 10 fingers for 88 keys, which means that some
robust planning ahead is necessary. A climber “reads the wall” in the way a pianist prepares her fingering,
knowing that in order to go further, some design work is anyway necessary, as everything, from
positioning of support points, extensions, breathing, to hand and forearm gestures will affect the sound.
Likewise, the mature climber will read her wall to recycle forces from a point to another, in an manner
appropriate to the style of climbing and in connection with the physical requirements of the wall.
Debussy claimed that “fingering is a second technique”. Fingering is to the pianist what breathing is to
singers. In order not to be taken by surprise, one should run a predictive reading of the score. And by
analyzing a pupil’s fingering one has an idea of what the pupil understood of the score.
The interaction that provides the context to this metaphor took place between the pedagogist and a
renowned climber. The pedagogist did not expect the pupil to invest his body spontaneously and
accurately in playing the piano, based on his awareness of a common bias concerning the
conceptualization of the piano. It seems at first sight obvious that one should play by performing “down”
movements “falling” on the keyboard (the reasoning being that since the keys are pressed down, our
movements will need to move in this direction, as it happens for a a computer keyboard.) In point of fact,
the muscular force is mostly deployed upwardly. In this way, a push on a key is a force resulting from a
movement of the arm upward, whereby both power and precision are obtained. Thus, fingers “land” but
are not “pushed” – in the same way that an airplane lands in “refusing the ground”, or a bird flaps its
wings in small beats down to sit on the branch.
The image used when the pupil tried to force his way on the keyboard was: “Is climbing clinging to or
moving around from hold to hold, like a squirrel ?” The body does not hang from a hold, it uses holds in
the search of an equilibrium. When the student understood the image, his body became reportedly more
active in the management of the movement at the keyboard. His upward movements are deployed by
spreading its points of support, distilling, considering the note played as a parcel of the gesture, which in
itself is a dynamic image of the sound to be produced, of the musical phrase to build.
The target is not to play the note, but to give a meaning to the note played, a meaning of which the body
is the dynamic image. One should further consider that the movement of the finger on the key is not an
autonomous item, as it is deployed in interaction with the keyboard. The same holds for a climber’s
concerns; in climbing, each single decision requires considering both the body and the wall.
[Arts and crafts] “Ton travail ne doit pas apparaître dans ton jeu, tout comme on ne doit pas voir les
pièces qui constituent la table ronde / One should not see the pieces that constitute a round table”
A carpenter who reaches underneath a round table looking for the constitutive elements knows that no
piece of wood is round but may be unable to realize how the carpenter could create this perfect
roundness. The result of an interpretive study, of the analysis of a work or of an improvisation should
conceal the elements that constituted its preparatory phase. One should not hear a beat or a scale, but a
flow that runs over the tempo. The image of roundness evokes perfection and precision, which put strong
constraints on the shape of the pieces to be assembled.
The starting point of our study is the widespread habit of using metaphors in music teaching. Given that
this fact is acknowledged in the literature on music teaching, we focused on a qualitative analysis of a
corpus in order to substantiate our hypothesis, that the use of metaphors is justified by their efficiency in
conveying experience-like representations which properties can immediately participate in the creation of
complex motor plans. Metaphors are more efficient than an explicit description of the relevant movement
since they use more efficiently the resources of experiential representation, minimizing the loss of
information induced by a conversion in a language-like format. Moreover, metaphor directly taps into
abstract experience-like representation that can immediately be put to work in the formation of complex
motor plans. An analysis of the large number of metaphorical vehicles gathered lends support to the
hypothesis. Metaphors, analogies and imaged language are used predominantly to teach complex
movements and conducts for which there is no spoken word available. True enough, there are not enough
words (lexicalized concepts) for fine-coding movements in communication, and the metaphors in our
corpus were overwhelmingly related to the necessity of communicating about the structure of gestures.
According to our hypothesis, however, it is not lack of words (lexicalized concepts) per se that elicits the
use of metaphorical or imaged language. Metaphors are used because they activate experience-like states
(imaginative states, for instance). They would be needed even if there were many more lexicalized
concepts. Our speculation is that this is because experience-like states, as opposed to language-like states,
are directly interfaced with fine-grained motor plans. Metaphorical vehicles capitalize on previous
knowledge of complex gestures in other domains, but as such knowledge cannot be recycled as it is, the
activation of experiential states is the key element in our account. Searching for the relevance of a given
metaphor, the pupil activates experience-like mental states that have abstract qualities (visual, motor,
proprioceptive) that can immediately participate, without conversion in a language-like format, in the
creation of complex motor plans.
The effect of the communication of experiential concepts is not linear or mechanical. The teacher knows
that activation of experiential states has an effect on the pupil, but she cannot access the exact mechanics
that makes the effect possible. Redescribing one’s left hand as a jellyfish makes many other aspects and
views of the hand possible and active in controlling the hand’s actions. These views and aspects are
Of course, processing a metaphor, thus ending up in a certain experience-like state, is no guarantee of
correct execution of the target action on the part of the receiver. Practice, which generates increasingly
subtler feedback loops and correction, has not substitute in learning music.
The research was conducted in the absence of any commercial or financial relationships that could be
construed as a potential conflict of interest.
Barsalou, L. W. (2008). Grounded cognition. Annu. Rev. Psychol., 59, 617-645.
Barsalou, L. W. (1999). Perceptual symbol systems. Behavioral and brain sciences, 22(04). 577-660.
Barten, S.S. (1992). The Language of Musical Instruction. Journal of Aesthetic Education, 32, 2, 53-61.
Barten, S.S. (1998). Speaking of Music: The Use of Motor-Affective Metaphors in Music Instruction,
Journal of Aesthetic Education, 26, 2, 89-97
Fodor, J. A. & Pylyshyn, Z. W. (1988) Connectionism and cognitive architecture: A critical analysis.
Cognition, 28(1). 3 –71.
Hofstadter, D., Sander, E. (2013, L’analogie. Paris: Odile Jacob.
Hommel, B., Müsseler, J., Aschersleben, G., & Prinz, W. (2001). The theory of event coding (TEC): A
framework for perception and action planning. Behavioral and brain sciences, 24(05). 849-878.
Lakoff, G., Johnson, M. (1980). Metaphors We Live By. Chicago: University of Chicago Press.
Marcus, G. (2012). Guitar Zero. The New Musician and the Science of Learning. New York: The
Penguin Press.
Neuhaus, H. (1973). The Art of Piano Playing. New York: Praeger.
Pinker, S. (1994). The Language Instinct. New York: W. Morrow and Co.
Prinz, J. (2002). Furnishing the Mind: Concepts and Their Perceptual Basis. Cambridge, Mass.: The
MIT Press.
Ritchie, L. D. (2009). Relevance and simulation in metaphor. Metaphor and Symbol, 24(4). 249-262.
Rosen, C. (1999). On playing the piano. The New York Review of Books, October 21.
Schumann, R. (1946). On Music and Musicians. Berkeley and Los Angeles: University of California
Sperber, D., & Wilson, D. (2008). A deflationary account of metaphors. The Cambridge handbook of
metaphor and thought, 84-105.
Suárez-Tosté, E. (2007). Metaphor inside the wine cellar: On the Ubiquity of Personification Schemas in
Winespeak. Metaphorik, 12, 53.
Tendahl, M., & Gibbs Jr, R. W. (2008). Complementary perspectives on metaphor: Cognitive linguistics
and relevance theory. Journal of Pragmatics, 40(11). 1823-1864.
Wilson, D. (2011). Parallels and differences in the treatment of metaphor in relevance theory and
cognitive linguistics. Studia Linguistica Universitatis Iagellonicae Cracoviensis, (128). 195-213.
Fig. 1. Number of metaphorical vehicles per cluster.
Fig. 2. The 19 types of analogical vehicles.

Citation: Casati R and Kaddouch R
Received: 27 Mar 2013; Accepted: .
Copyright: © Casati and Kaddouch. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.