74 TEACHERs COLLEGE COLUmbIA UNIvERsITy
visuomotor Atoms of Copy-Drawing
However, this “inverse” approach cannot be just
taken for granted, because of two reasons at least.
First, the perceptual experience that artists have of
their own artwork is undoubtedly biased by their
experience of the whole process that led to the
result. Against this objection it could be argued that
at least the so-called Early visual analysis, which is
thought to be cognitively impenetrable (Pylyshyn
2000), should be common to the artist and the per-
ceiver. It is well known however that even the per-
ception of basic image features such as contrast and
orientation, which are processed in the early steps
of the visual system, can be strongly modulated by
top-down factors including visual attention and task
performance (Reynolds & Chelazzi 2004). Further-
more, perception of a visual scene is known to be
determined also by how overt attention is deployed,
i.e. via eye movements, and it is plausible that the
attention of the artist would be directed to regions
of the image that are critical during the creative act,
but that are not necessarily the same that perceivers
will attend to. The second, more general reason is
that there is no evidence, at present, that the percep-
tual process and the creative process share the same
neural mechanisms. At least two different positions
could be taken here: a) according to recent formula-
tions of the dual vision theory (Milner & Goodale
1995), two separate pathways of visual processing
exist in the human brain, one implementing per-
ceptual functions (e.g. object recognition) and ulti-
mately delivering to us a coherent visual experience
of the external world, and the second one subserv-
ing the control of motor actions. According to this
view, it could be argued that mainly the Vision for
Perception pathway is involved during artwork per-
ception, while the Vision for Action stream is the
one upon which the visual creative process relies.
b) following the sensorimotor approach to percep-
tual experience (O’Regan & Nøe 2001), and even
more profoundly in view of the existence of mirror
neurons (Metta, G., et al. 2006), it could be argued
that the perception of a visual artwork involves an
internal simulation of the actions that produced
that specific image, which would provide a common
ground to artwork creation and artwork perception.
In this work we do not commit to any of the
above-mentioned positions; we propose instead
that a grounded, “direct” analysis of the creative
process itself is a much-needed approach in the
scientific debate on visual creativity. Creative pro-
cesses can be regarded, from the vantage point of
cognitive science, as a goal-directed activity involv-
ing several human skills and abilities: sensorimotor
coordination, evaluation and decision, memory and
emotion. In this perspective, we surmise that the
analysis of the creative process by scientific means
can prove itself a powerful methodology for the
understanding of human capabilities such as those
mentioned above, at least as much as the analysis
of visual artwork perception has proven fruitful for
the understanding of human vision.
In order to narrow down our field of analysis, we
focused on sensorimotor coordination, namely the
problem of how sensory and motor resources are
integrated to give rise to efficient behaviors for the
solution of specific tasks. In particular, as explained
below, our analysis concentrated on eye–hand
coordination in the task of performing an accu-
rate drawing from observation, namely copying
an original image on an initially blank canvas. The
issue of eye-hand coordination in drawing has been
addressed by a number of authors (Viviani & Flash,
1995; Cohen, 2005; Land 2006; Gowen & Miall,
2006; Coen-Cagli et al. 2009; Tchalenko 2009). On a
global behavioral level, a consistent feature of repre-
sentational drawing strategies is the following exe-
cution cycle: fixation on the original image; saccade;
fixation(s) on the canvas; saccade; fixation on the
original image. The specific kind of visual process-
ing that takes place when fixating on the original is
still unclear in general, but two main positions have
been outlined: 1) fixations on the original serve to
encode image features to visual working memory,
and such mental image is later recalled and con-
verted to a motor plan (Tchalenko et al. 2003); or,
2) the visuo-motor mapping from image features
to hand motor activity takes place during fixations
on the original image, without the need to invoke
working memory (Coen-Cagli et al. 2007).
This last view is consistent with results from our
eye tracking experiments, which explored how eye
movements of human subjects involved in drawing,
differ from those obtained in free viewing control
experiments. The results can be summarized by
the observation that not only are eye movements
in drawing strongly biased by the task, but also a
precise dependency can be established between
the peculiar motor constraints of the task and the
gazing behavior. In particular, we show that the
observed eye movements represent a precise strat-
egy to meet the hand motor constraint of graphical
continuity.