Cognitive Semiotics

Home , Cognitive semiotics, Last Supper (El Greco), Madonna della Vittoria, Max Wertheimer

Multidisciplinary Journal on Meaning and Mind

Issue 5 . Fall 2009

PETER LANG Bern · Berlin ' Bruxelles ' Frankfurt am Main · New York · Oxford · Wien COGNITIVE SEMIOTICS

EDITORS-IN-CHIEF Per Aage Brandt and Todd Oakley

CO-EOrTORS Ana Margarida Abrantes, Tim Adamson, Une Brandt, Riccardo Fusaroli, and Jes Vang

EDITORIAL ASSISTANT (official address and address for unsolicited submissions) Larimee Cortnik Department of Cognitive Science Center for Cognition and Culture Case Western Reserve University College of Arts & Sciences Crawford Hall, 612D Cleveland, Ohio, 44106-7179 USA Phone: (+1) 216 368-6538 · Fax: (+1) 216 368-3821 [email protected]

COORDINATING EDITOR (general address for solicited submissions and editorial contact) Jes Vang · [email protected]

EDITORIAL ADVISORY BOARD Liliana Albertazzi, Bernard Baars, Enrique Bernárdez, Peer Bundgaard, Roberto Casati, Christopher Collins, Seana Coulson, Ian Cross, Terrence Deacon, Merlin Donald, Shaun Gallagher, Barend van Heusden, Robert Innis, Jana M. Iverson, Mark Johnson, Torben Fledelius Knap, Kalevi Kull, Ronald Langacker, Michael Leyton, Ricardo Maldonado, Juana Isabel Mann-Arrese, Erik Myin, Frederic Nef, Pierre Ouellet, Jean-Luc Petit, Jean Petitot, Martina Plümacher, Roberto Poli, Ernst Pöppel, Andreas Roepstorff, Bent Rosenbaum, Maxine Sheets-Johnstone, Chris Sinha, Linda B. Smith, Göran Sonesson, Frederik Stjernfelt, Eve Sweetser, Leonard Talmv, Evan Thompson, Colwyn Trevarthen, Reuven Tsur, Mark Turner, Patrizia Violi, Wolfgang Wildgen, Dan Zahavi, Lawrence Zbikowski, Jordan Zlatev, and Svend Ostergaard.

MANUSCRIPT SUBMISSIONS For style guide and other directions for authors go to the journal's website: www.cogniavesemiotics.com

PUBUSHING DETAILS © by Verlag Peter Lang AG, Hochfeldstrasse 32, CH-3012 Bern

Tel. +41 31 306 17 17; Fax +41 31 306 17 27; E-Mail: [email protected]; Internet: www.peterlang.com

All rights reserved. The journal and its parts are protected by copyright. Any utilisation outside the strict limits of the copyright law, without the permission of the editor, is forbidden and liable to prosecution. ISSN 1662-1425

Printed in Switzerland

COVER DESIGN AND GRAPHICAL LAYOUT Jes Vang COGNITIVE SEMIOTICS

ISSUE 5 · FALL 2009

AESTHETIC COGNITION

EDITED BY PEER BUNDGAARD &. JEAN PETITOT

4 Editorial Preface

7 Jean Petitot Non-Generic Viewpoints as a Method of Composition in Renaissance Paintings

42 Peer F. Bundgaard Toward a Cognitive Semiotics of the Visual Artwork - Elements of a grammar of intuition

66 Wolfgang Wildgen Geometry and Dynamics in the Art of Leonardo da Vinci

93 Ivan Darrault-Harris Non-Genericity as an Invariant of the Readability of Pictures

103 Zoï Kapoula, Qing Yang, Marine Vernet &. Maria-Pia Bucci Eye Movements and Pictorial Space Perception: Studies of paintings from Francis Bacon and Piero della Francesca

122 Alessandro Pignocchi What Is Art? A methodological framework for a pluridisciplinary investigation

136 Ellen Dissanayake The Artification Hypothesis and Its Relevance to Cognitive Science, Evolutionary Aesthetics, and Neuroaesthetics

OFF-ΤΉΕΜΕ ARTICLE 159 Michael Kimmel Analyzing Image Schemas in Literature Editorial Preface

The present issue of Cognitive Semiotics is devoted to the study of meaning making in art. Artworks and aesthetic experience have long since been privileged objects of investigation within the humanities and philosophy. There are very good explanations for this. Perhaps the main reason being that aesthetic (and proto-aesthetic) activities reflect the most important examples of human semiotic behavior. Through millennia, and across cultures, people have pervasively constructed meaning in aesthetic matter and by aesthetic means. Thus, nothing seems more natural than to systematically explore this domain. However, the fact that aesthetic objects, aesthetic experience, or aesthetic communication are obviously different from normal objects, everyday experience, or plain communication, scholars often consider aesthetic meaning construction as something essentially different from other types of meaning making: either because of its content (a type of insight that only artists have access to), or because of its vehicle (the use of expressive means that only characterize aesthetic communication). However vast, the continent of aesthetic meaning making has by and large been considered a continent apart. In the heading we have chosen for this volume, the word "cognitive" suggests a somewhat different approach to artistic objects, their meaning and the experience of them. Indeed, the contributors to this volume display, in different degrees, what could be considered a "naturalizing" approach to aesthetic meaning construction (both from the producer's and the receiver's perspective). Naturalizing methodologies, of course, come in many different and sometimes mutually exclusive variants, but a common trait between them correlating aesthetic meaning making and aesthetic cognition, does not constitute a genus of its own (sustained by some autonomous aesthetic module in the mind), but should be considered, characterized, and if possible, explained in the light of what we know about (the evolution of) everyday cognition and human semiotic behavior in general. Such an approach is probably so much more promising when the task consists in defining the way artists actually shape their meaning intentions, and why these means are particularly efficient. Painters can convey meaning to the beholder thanks to their motifs and the iconographical (commonplace) values attached to the latter. However, paintings are visual media, and if we are to take seriously the existence of some sort of visual or aesthetic semiotics, then it must be possible to track down the ways in which something in a painting can take on a perceptual (and not simply a conceptual or iconographical) signification. Otherwise, painters can depict or reproduce already established meanings and values, but not form them on their own, or even from scratch. In regard to this there is nothing new under the sun: the craftsmanship of painting consists, as Rudolf Arnheim stated it back in 1950, in conveying to the eye what the motif represents to the mind. EDITORIAL PREFACE | 5

In other words, the focus in this issue of Cognitive Semiotics is on aesthetic semiosis; the construction of perceptual signification in the artwork. Therefore, a natural way of proceeding for many of the contributors would first be to ask: what are the features, relations, and qualitative properties of the phenomenal world to which the visuo-cognitive system is geared to respond to in specific ways? And then to move on with the next question: how can such features, relations, and so forth, be exploited by the artist as a tool for perceptual meaning construction? If this relation between natural signification in everyday perception and aesthetic meaning construction can be convincingly laid down, it becomes possible to establish the general devices by means of which painters encode meaning in shape, colors, strokes, and relations between the latter (composition). In his paper, Jean Petitot considers composition as a way of constructing perceptual meaning properly. If you want to construct morphological or broadly spatial counterparts to symbolically or iconographically significant entities in a painting, you must be capable of selecting certain spatial relations as outstanding and thereby endow them with a given semantic function. In a detailed analysis of Mantegna's The Madonna of Victory and Raphael's St. George, Petitot shows how painters semiotize spatial relations by adopting non-generic, unique and critical vantage points, or by constructing such non-generic configurations or morphological correlations. Such a procedure is a key element of compositional meaning making to the extent that it exploits humans' inbuilt sensitivity to rare or critical phenomena. In the same vein, Peer F. Bundgaard attempts to lay down some of the principles which rule meaning making in visual artworks. In his paper, he shows how artists exploit features which are intrinsically significant for the visuo- cognitive system in plain, everyday perception, thereby transforming the automatisms of perception into a rhetoric of aesthetic intuition. The central concern is to demonstrate how purely spatial relations can become significant, or, in other words, how conceptual meaning can be anchored in perception. Wolfgang Wildgen makes a case for Leonardo da Vinci's contribution to a "semiotics of art". In an analysis of the Last Supper, Wildgen first shows how the concepts of "geometry" and "dynamics" can be used to lay bare the compositional meaning of the painting. In the second section of his paper, he aims to characterize the configurational and gaze-dynamical structures in some of Leonardo's paintings (Virgin in the rocks, St. Anne) as morphological counterparts to valence patterns (case frames, scenes) in sentences or short narratives. Finally, these patterns are related to models from dynamic systems theory. Iran Darrault-Hams' article is a case study of the use of non-genericity as a meaning generator in a concrete artwork, here a diptych by Cranach the Elder picturing Lucrece committing suicide and Judith after the murder of Holo- phem. 6 I P. BUNDGAARD & J. ΡΕΠΤΟΤ

Zoi Kapoula et al. present an empirical eye-tracking study of aesthetic experience, with particular focus on the difference between eye movement patterns in observers trained in experiencing art and observers who are not. With their study the authors reveal, for example, that trained subjects are much more sensitive to elements of the painting which seem to play an important compositional role, whereas observers without art training tend to focus on the key objects of the motif. In his contribution, Alessandro Pignoccbi critically assesses the import of disciplines such as cognitive psychology, evolutionary psychology, and the neurosciences on traditional approaches within aesthetics. He aims to define the foundations of a methodological framework which distinguishes between four basic topics: (1) the investigation of the cognitive phenomena elicited by the experience of things called 'artworks', (2) the investigation of the psychological structures determining the intuitive categorization of something as art, (3) the study of the intuitions and arguments used to build justifications about whether something is a work of art, and (4) the elaboration of a normative discourse about what ought to be called 'art'. Finally, he makes a claim to the effect that each of these topics can benefit from the cognitive approaches, provided that some specific methodological recommendations are respected. In her paper Ellen Dissanayake leads back crucial formal features of "artification" (such as formalization, repetition, exaggeration, elaboration, and manipulation of expectation) to mother-infant bonding, or to those features of maternal communication that infants universally respond to, features that Dissanayake thus proposes to consider as "proto-aesthetic". She advances a hypothesis to the effect that these capacities and receptivities, which evolved from about 1.7 million years ago to enable mother-infant bonding, became a reservoir of affective mechanisms that could be used subsequently by ancestral humans when they first began to "artify" - that is, when they invented the "arts" as vehicles of ceremonial religious experience. Her hypothesis emphasizes the importance of preverbal, presymbolic, pancultural, cross-modal, supra- modal, participative, temporally organized, affective, and affinitive aspects of aesthetic cognition and behavior.

We are extremely grateful that Cognitive Semiotics has offered us the opportunity to present a discussion about crucial aspects of aesthetic cognition. We wish to thank the contributors for their valuable discussion about the foundations and devices for aesthetic meaning making. We are also indebted beyond imagination to the editors, their patience, their thorough readings and constructive critique, as well as to the anonymous peer reviewers who improved all the above papers.

Peer F. Bundgaard & Jean Petitot

For more information on upcoming issues, calls for submissions, etc., please visit our website at http:/ / www.cogrtitivesemiotics.com. Jean Petitot

Non-Generic Viewpoints as a Method of Composition in Renaissance Paintings1

In order to spatially represent semiotic relations in a visual work of art, one needs to consider more than simple spatial relations since, in any group of figures, the latter are necessarily spatially related to each other whatever their semiotic relations can be. Thus, in order to monitor semiotization one must resort to some kind of "second order" relations, that is, singular spatial relations, which can be identified as playing a functional role. Such a constraint pertains to the techniques of compodtion. There are in fact very few means to select and singularize spatial relations. Here, I will consider only Renaissance paintings, where the unitary Brunelleschian geometrical space acts as a framework. Much scholarly work is available about techniques of composition, which consist, e.g., in placing the vanishing point at a strategic place in the picture or in introducing geometrical schémas (circles, triangles, etc.) intended to regulate the distribution of figures. There are, however, other methods of composition that have never been thoroughly investigated. Among the latter, one of the most powerful consists in using non- generic viewpoints. As a method, non-genericity is a very efficient way to pictorially represent semiotic relations whenever point of view plays a constitutive role in a given artwork. In several previous works (1986, 2004, 2009a,b), I have studied the use of non-generic viewpoints by certain Renaissance painters, particularly Piero della Francesca and Nicolas Poussin. In the present paper, I will focus on two other examples, one by Raphael, and the other by Mantegna. My approach will be stricdy methodological. the task consists in extracting, on the basis of purely formal criteria, immanent, intrinsically significant, morphological and non-conceptual relations which can legitimately bear meanings, be semiotized in the interpretation and thus ensure that the interpretation is not merely the speak- er's subjective construction.

CORRESPONDENCE: Jean Petitot. School for Advanced Studies in Social Sciences (EH ESS), Paris, France. EMAIL· [email protected]

Full color reproductions of all images and paintings can be found online at http://www.cognitivesemiotics.com (click on "issues overview" in the sidebar and scroll down to #5 to access image files for individual contributions).

Cognitive Semiotics, Issue 5 (Fall 2009), pp. 7-41 8 I J. ΡΕΠΤΟΤ

1. Compositional spatial relations and the concept of non-genericity In order to better define the problem, I would like to take over Omar Cala- brese's way of addressing the issue concerning the specifically pictorial tools for representing narrative structures. In his analysis of Duccio's and Simone Martini's Maestà, Calabrese remarks that a painting can be considered a "text":

For readers accustomed to the semiologists' definition of this notion: any object manifested in whatever substance of expression endowed with some sort of closure which delimits it as a unit per se, and which displays internal relations between its parts (Calabrese 2002: 5).

Now, what may "internar relations mean in the case of perceptive objects that can only stand in "externar relations to each other? This is the reason why the definition of spatial relations that are also compositional constitutes a fundamental theoretical problem. For visual artworks, the substance of expression (in Hjelmslev's sense) is space, and the form of expression is thus constituted by spatial relations. However, all perceived objects are perceived as being located in one and the same space, and whatever their positions may be, they will always be spatially related to each other. If relations in the structural and semiotic sense of the notion are to appear, they must consequendy make out a subset of very particular, even exceptional relations. And the whole difficulty consists in defining such a subset on the grounds of immanent and purely spatial criteria. Here is where the concept of non-genericity comes into play. The concepts ofgenericity and non-genericity are crucial geometrical ones. They go back at least to the geometer-painters from the Renaissance (Piero della Francesca, Mantegna, and others), but they have never been appropriately theorized until Hassler Whitney and René Thom developed their theory in the middle of the 20th century on the grounds of the notion of "posizione generale" previously formulated by 19th century Italian algebraic geometers. Let us proceed by considering a form, or a configuration or a structure F, which can vary smoothly under the control of some external parameters w. A state F„. of F is called generic if its qualitative type does not change when w varies slighdy, in other words, when it resists small deformations. In a plane, for example, the fact for two lines to be parallel or orthogonal is a non-generic property. The same goes for the property of being aligned for two segments or of being equilateral for a triangle. A typical case is the situation where w is a viewpoint and F„ the apparent contour of a shape seen from w. Non-genericity can then induce remarkable perceptual effects. It is, for instance, well known that the apparent two- dimensional contour of a transparent cube seen in (axonometric) perspective NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSITION IN RENAISSANCE PAINTINGS | 9

from a generic point of view is spontaneously interpreted as a three- dimeisional object by the visual system. We even know that there exists a percotual bimodality ("Necker cube" phenomenon) which triggers an aliénation between two perspectival interpretations depending on whether the two dagonal summits A and B, placed along the depth axis, are interpreted as "A ir front" and "B behind" or "A behind" and "B in front" (Figure 1).

Figurtl. The so-called "Necker cube" phenomenon. The apparent contour of a transparent cube seen from ι generic point of view (in the middle) may be interpreted in two contrary ways as a three- dimenional object (to the left and to the right).

Yet, in the non-generic case of a hexagonal and maximally symmetrical appaent contour (point of view d placed along one of the four diagonals of the cube, the visual system no longer reconstructs the third dimension. The latter vanisies, and the figure is interpreted as a hexagon (Figure 2). This 3D —> 2D collajse was well known in the Renaissance.

Ftgur 2. The apparent contour of a cube in a non-generic position is interpreted by the visual system as a puriy two-dimensional figure (a hexagon). 10 I J.PETITOT

This example can be used to give a more precise explanation of non-genericity. Figure 3 represents 12 viewpoints distributed along a small cone around the diagonal axis d. As can be seen from the figure, the unstable sextuple point in the center of the hexagon bifurcates into two triple points, and there exist further 6 non-generic situations in which one of the branches of one of the triple points is aligned with one of the branches of the other. These 6 unstable cases, albeit to a lesser degree than the hexagonal case, make out the transition between the 6 possible generic viewpoints on the cube. In other words, there exist 6 generic viewpoints separated by 6 non-generic cases with instability degree 1 (this degree is called "codimension"), and these 12 cases are organized by a non-generic viewpoint of codimension 2.

Figure 3. 12 generic and non-generic viewpoints of codimension 1 that totally or partially stabilise the non-generic viewpoint of codimension 2 in Figure 2. The non-generic viewpoints of codimennon 1 correspond to the columns 1, 3, 5 (the last figure of each row is repeated in the beginning of the following row, and the last row runs on into the first one). The generic viewpoints correspond to columns 2,4.

One further typical example of non-genericity is that of alignment. In fact, alignment yields a typical example of non-genericity. If you embed η aligned segments in an environment Β of randomly oriented segments, then two very different cases can be observed: a. either Β is sufficiently dense for there to be some probability for finding η aligned segments; in this case, the visual system does not remark anything; b. or B is sufficiently sparse for there to be quite no chance for finding η aligned segments; in this case, the alignment "pops out" as in columns 1, 3, 5 of figure 3 (Figure 4). NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSITION IN RENAISSANCE PAINTINGS | 11

Figure 4. A random distribution of oriented segments. The eye is particularly sensitive to alignments.

Since the visual system is a (most likely Bayesian) probabilistic neuronal machine, which learns to extract statistical regularities from the environment (see e.g. Yuille & Kersten 2006), it is also very good at detecting ran events, which it treats as intrinsically significant exacdy because they are rare. This is what certain specialists - in particular Jean-Michel Morel (see Desolneux, Moisan & Morel 2003) - call the Heimholtç principle. Since non-genericity is rare, it is perceptually salient and provides an immanent, purely perceptual, criterion for defining the difference between perceptual structure and artistic composition, at least for those artworks in which point of view plays a fundamental role. It is precisely because their probability vanishes that non-generic configurations can express a semiotic intentionality and what Kant called in the third Critique a "subjective formal finality". The notion of non-genericity is rather subtle. It is not just a matter of statistically rare events. In fact, in a set of equiprobable events any event chosen as a reference is as rare as any other. In a card game, for example, a four ace hand is not rarer than the quadruplet C {the jack of clubs, the 4 of diamonds, the 2 of clubs, the 10 of hearts}, and one could indeed imagine a game where C would be the top configuration that any player would try to obtain. The non- generic elements are rare, but they are so for structural reasons. They are indeed defined in terms of their instability relative to small variations w of the control parameters, and it is in fact their instability that makes them rare. Behind this affirmation, there are deep mathematical results, in particular Thorn's transver- sality theorem. 12 I J. ΡΕΠΤΟΤ

In the visual artworks, the concept of non-genericity is intimately associated with the concept of structun (in the technical structuralist sense of the term). As Goethe was the first to explain in his Laokoon (1798), inspired by his long and patient studies of biological morphogenesis, observable aesthetic structures are parallelisms between directions, orthogonalities, symmetries, unstable moments, etc., in a nutshell epitomes of non-generic elements (see Petitot 2004). This present view of non-genericity has nothing to do with the idea that, since any unexpected encounter is improbable, it is for that very reason intrinsically significant. Such a technique, profusely exploited by surrealist artists, rests on an opposite logic. Here, what counts is, as we just said, the instability of certain spatial relations, and it is their very instability that transforms their external relations into internal ones.

2. A first example by Raphael I have already dealt with this topic in a certain number of texts where I have analyzed some examples, in particular by Piero della Francesca, Poussin and by contemporary artists (Felice Varini and Georges Rousse) (see Petitot 1986, 2004, 2009a,b). Here, I will look at two different examples, one by Raphael, and the other by Mantegna.

2.1. A triple point in Raphael's Saint George The first example concerns one of the two representations of Saint George painted by Raphael (Washington National Gallery of Art, 1505). The preparato- ry study, conserved at the "Gabinetto dei Disegni e delle Stampe" of the "Galleria dei Uffizzi" in Florence (Figure 5), already shows a well structured representation of the narrative structure. NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSITION IN RENAISSANCE PAINTINGS | 13

Figur, 5. Raphael, Saint George and the Dragon, Gabinetto dei Disegni e delle Stampe of the Gallina dei Uffici.

One might notice the parallelism between the horse and the dragon (their bodies are translated along the axis of the lance, and the coiled tail of the dragon is exactly tangent to the belly of the horse, or rather to the girth), the symmetrical torsion of their heads, the horse looking toward the observer, the fusion of Saint George's head and the horse's head, on the one hand, and on the other hand the fusion of the lower princess' body and the horse's tail. This construction is remarkable insofar as, at the narrative level, the legend combines three actors taking on three precise actantial roles in Greimas' sense (see Gremas, Courtès, 1993): Saint George is the Subject, the dragon is the Anti- sübtet and the princess is the Object of value. However, in the painting, it is the horse which embodies their narrative relations by means of positional fusons: head-head, belly-back, tail-lower body. One should furthermore notice a particularly interesting non-generic element: the point of tangency between the extremity of the lance and the bonier of the cape floating in the wind (Figure 6). 14 I J. PETTTOT

Figure 6. Non-generic point of tangency between the extremity of the lance and the border of the cape.

The scenery is also particularly interesting. It consists of two groups of each three trees with perfectly parallel trunks: a 3-type group (i.e., a truly ternary group) exactly above the head of the princess, and a (l+2)-type group (a fully exposed tree and two trees partially occluded by the bushes) exactly at the extremity of the cape. Other important elements are, of course, the cave to the left surmounted by three trees, as well as the central church tower. Their meaning is canonical: the dragon is a chthonian monster, and Saint George is the knight of the Church. In the final painting (Figure 7), the figurative representation of the narrative structure is kept unchanged. NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSITION IN RENAISSANCE PAINTINGS | 15

Figure 7. Raphael, Saint George and the Dragon (1505, Washington National Gallery of Art).

However, Raphael has added a sword, which is symmetric to the lance relative to the axis of the foot (together they form a sort of compass). At one extremity, on the hilt, it has a small ball tangent to the elbow, and at the other extremity, at the end of the blade, it cuts the extremity of the dragon's tail. The appearance of the sword is important, since, in the iconography of Saint George, there is a progressive transition, as to the weapon that kills the dragon, from the lance to the sword (in the latter case, the lance is represented as broken).2 The transition is epitomized by Carpaccio's Saint George (1502—1507, Scuola di San Giorgio degli Schiavoni, Venice), where the knight and the dragon confront

2 See Petitot 1979 for a detailed analysis. 16 I J. ΡΕΠΤΟΤ

each other in a very outstretched horizontal structure with the lance as axis, and where the painter has represented the very moment when the lance breaks. The importance of this lance/sword duality is relevant for Raphael to the extent that he painted two completely different Saint George, the second of which (Louvre, 1504) being organized around the arm rising the sword and the fragments of the broken lance scattered all over the ground (Figure 8).

Figure 8. Raphael, Saint George and the Dragon (1504, Louvre).

But it is at the level of the scenery that Raphael elaborates the most his study by creating an astonishing structure. He introduces in the setting two extremely non-generic elements (Figure 9): a. a triple point formed by the crossing of the back extremity of the lance with the crossing of the two main branches of a tree in the background; b. a tangency not at one simple point (minimal degree non-genericity), but along a whole segment of the lance and the border of the cape (infinite degree of non-genericity). NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSITION IN RENAISSANCE PAINTINGS | 17

Figure 9. The non-generic triple point lance-crossing of branches and the hypertangency lance/cape.

The triple point is non-generic, and its extreme singularization opens the possibility of a structural interpretation (i.e. an interpretation without content, purely structural, different from any hermeneutic projection). In fact, the cape covers a "trinity" of branches (Figure 10): first, a trunk bifurcates into two main branches (which is an intrinsic catastrophe of the tree); then one of these branches (the rightmost one) bifurcates in its turn, and since the angle of bifurcation is superior, it ends up crossing the other branch in an apparent crossing (an extrinsic catastrophe due to point of view). The latter crossing is generic (structurally stable), since it is qualitatively robust relative to small changes in point of view.

Figure 10. The two bifurcation catastrophes and the crosnng catastrophe in the three parts of the tree. 18 I J. ΡΕΠΤΟΤ

What obtains here is an interesting formal graph structure of the Trinitarian tree (Figure 11).

Figure 11. The graph of the Trinitarian tree.

Now, the cape occludes the "intrinsic" (ontological) bifurcations: only the apparent "extrinsic" generic (phenomenal) crossing addressed to the observer subsists. But a Trinitarian structure is restored, which is obtained by transforming this crossing in a non-generic triple point adding the lance as a supplemen- tary third branch (Figure 12).

Figure 12. The triple non-generic point adds the lance to the generic crossing of two branches, thereby reconstituting a Trinitarian structure which is relative only to point of view and therefore purely "extrinsic" (phenomenal). NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSITION IN RENAISSANCE PAINTINGS | 19

For this reason, it is structurally relevant to analyze the configuration of the cape floating in the wind. The rightmost part of the cape touches 4 trees with rigorously parallel bifurcations, in turn a highly non-generic structure due to the fact that the 4 trees are replica of one and the same model (the bifurcations are thus all at the same height) and aligned along a perspective axis. What is more: this group of 4 trees is symmetric to another group of (4 + 1) trees that are parallel, but without bifurcation (Figure 13).

Figure 13. The rightmost part of the cape: there are two groups of parallel trees; the group of 4 is in perspective, the group of 5 (4+1) is more frontal

By virtue of its extraordinary morphology, the cape reproduces a ternary structure. It contains a dominant bifurcation forming a horizontal Y, a large shaded fold that starts bifurcating in the inner black part of the cape. This shaded part covers the Trinitarian tree. Then, in turn, the inner rightmost (black) part of the cape bifurcates into yet another Y that, arguably, corresponds to the two groups of trees (Figure 13). This means that, just as the horse embodies the actantial relations between the three semantic roles of the narratives, the cape embodies the "theological" relations between the three groups of trees. This example from Raphael shows how the theological interpretation of the work is encoded in the setting. The non-generic geometrical relations lay bare a system of formal relations which works like a system of equations. And just like the attribution of a numeric value to one of the variables of a system of equations makes it possible to compute the numeric value of the other 20 I J. ΡΕΠΤΟΤ

variables, here the attribution of a meaning to one of the terms - e.g., the theological interpretation of the "Trinitarian" tree — leads step by step to a global semiotic interpretation. The most relevant interpretation seems to be "theological": Saint George substitutes an ontological (divine) trinity with a phenomenal (human) trinity and establishes a link between the latter and a collective structure (the Church?, Humanity?).

2.2. Addendum: Bellini's Madonna degli alberetti Recendy (November 2009), while we were visiting once again the Accademia Gallery in Venice, my wife pointed out to me the trees in the Madonna degli alberetti painted by Giovanni Bellini in 1487 (Figure 14).

Figure 14. Giovanni Bellini, Madonna degli alberetti, 1487, Gallerie dell'Accademia, Venera. NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSITION IN RENAISSANCE PAINTINGS | 21

The configuration of the tree at the viewer's right (i.e. at the Virgin's left) is quite exactly the same as in Raphael's Saint George. First, a long part of the trunk is perfectly tangent to the Virgin's left sleeve. Second, the bifurcations of the main branches are exactly the same (up to a symmetry and a slight deforma- tion) as those in Raphael (Figure 15).

Figure 15. The two ternary trees: left, Beltini and right, Raphael.

3. A second example by Mantegna 3.1. The Madonna of Victory The second example concerns The Madonna (or the Virgin) of Victory by Andrea Mantegna (1495-96, Louvre). When Mantegna (1430-1506) painted this masterpiece, he was already 65 years old and enjoyed huge prestige. Since 1453 he had been the leading court painter at the Gonzaga's court in Mantua where to Louis III had called him. In 1490, he had returned from a stay in Rome during which he had become a great admirer of antiquity, and from 1492, Isabella d'Este (Francesco II's wife) employed him at her Studiolo of the Castello di San Giorgio. The painting had been ordered by the marquess Francesco II to celebrate his victory at the battle of Fornova (close to Parma) on July 6th 1495. At this time, the political situation was dominated by the battle of the anti-French league (Venice, the duke of Milan, the pope Alexander VI, the emperor Maximilian I of Habsburg, the king of Spain Fernando of Aragon), commanded by the Gonzagas, against Charles VIII (1470-1498, son of Louis XI), who had taken Naples on February 22nd 1495, after Florence and Rome, and had taken the title of King of Naples and Jerusalem (Figure 16). 22 I J. PETTTOT

Figure 16. Andrea Mantegna, The Madonna of Victory (1495—96, Louvre).

The work was financed by a Jewish banker, Daniele Norsa, who had been punished for having destroyed an image of a miraculous virgin in a house he had just bought. The painting was inaugurated with great pomp and indescriba- ble enthusiasm on the day of the first anniversary of the batde. Considered as an "opera eccellentissima" and object of intense devotion, it was installed in a chapel designed by Mantegna himself. The work was unanimously appraised. Passing through Mantua in 1793, the great archeologist and art historian Luigi Lanzi dubbed it "the threshold of modern painting", "the last step before it attains the perfection inherited from Leonardo". In 1790, Goethe, also passing through Mantua, made a similar judgment. This is the reason why The Madonna of Victory was selected by NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSITION IN RENAISSANCE PAINTINGS | 23

Bonaparte's commissioners Thion and Vicart in 1797—1798 during the Italian campaign to be part of the convoy of masterpieces triumphandy transported back to Paris. It is still considered an apotheosis of the Brunelleschian space, an ideal organization of the visible world by means of a centralized, unifying perspective; an ideal space which Mantegna by the way often transgressed. The painting is a "sacred conversation", that is to say a "Virgin with Child" representing the backer surrounded by saints. To the right we see Saint George with his broken lance, holding the mande of the Virgin, and, behind him, Saint Longinus with his red helmet and the red lance that pierced the right side of Jesus on the cross. According to the legend, after his conversion, Longinus died in Mantua; moreover, after the battle of Fornova, Francesco gave his brother Sigismondi his lance comparing it with Longinus' lance. Located symmetrically to the marquess, Saint Elisabeth (or Saint Anne) is holding a rosary. Symmetrically to the left, Saint Michael is also holding the Virgin's mande, his sword wrapped with a white to red ribbon, and behind him Saint Andrew. As regards the Virgin's position, the painting is in fact a blend of a "Virgin with Child" sitting on her throne, and a "Madonna of Mercy" where the Virgin, standing up, protects certain Chosen ones under her open mande. Mantegna had promised Sigismundi a Madonna of Mercy sheltering Francesco, his brothers, and Isabella under her mande. Next, he changed the configuration, while still respecting Sigismundi's wish that Saint George and Saint Michael, "victorious, one by his body, the other by his mind", should open the Virgin's mande. Furthermore, he translated the theme of the protection/benediction by Virgin's right hand — a gesture often compared to the one found in Leonardo's Virgin of the Rocks (1483, Louvre, Paris, and National Gallery, London, Figure 17). 24 I J. ΡΕΠΤΟΤ

Figure 17. Leonardo da λ/inà, The Virgin of the Rocks (1483, two versions: Loutre, Paris and National Gallery, London).

The Virgin is represented in a gold shimmering red robe (red and blue are conventional colors for the Virgin) and a green mantle with golden embroiders (gold is a holy color). She is sitting on her throne under a vegetal dome which is geometrically and architecturally analogous to the archetype often represented in Venice by Giovanni Bellini, e.g. in the Frari (triptych of Madonna with Child, 1488) or in San Zaccaria ( Virgin with Child and musical angel\ 1505): Bellini was by the way Mantegna's brother-in-law (Andrea had married his sister Nicolosia in 1453). (Figures 18 and 19). NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSITION IN RENAISSANCE PAINTINGS |

Figure 19. Giovanni 'bellini, Virgin with Child and musical angel (Venice, San Zaccaria, 1505). 26 I J.PETTTOT

The whole painting is filled with a wealth of elements and precious decorations, most probably referring to the esoteric environment of the Mantua court: rich moiré fabrics, precious stones, rock crystals, corals (a branch of corals hanging from two festoons of red pearls gathered in groups of six separated by quartz spheres), citrus fruits, feathers, birds (parrots), wood with inlaid mother-of- pearl, bronze, gold, etc. Gleams and transparencies animate matter. The steps of the throne show a representation of the creation of Adam, the original sin, and Adam and Eve chased away from Paradise. On the stool, on which the Virgin's feet are resting, one can read the inscription REGINA CELI LET(ARE) ALLELVIA, which is the beginning of the antiphony sung to Mary during the Easter liturgy. Several crosses can be seen, in particular the one on John the Baptist's small flag carrying the inscription ECCE/AGNUS/DEI/ECCE/Q(VI) TOLL/IT P(ECCATA) M(VNDI), the one held by Saint Andrew, and the one on Saint George's breastplate. The back of the throne is crowned by a magnificent rosette reproducing a pattern from the altar piece of San Zeno in Verona (1457): a disk of jewels with a 16-rays sun in its middle, a ruby (Mars' stone, a military symbol) in its center, a crown of knots circumscribed by 32 precious stones. This crown is formed of two braids of two intertwined strands (Figures 20, 21, 22).

Andrea Mantegna, The Madonna of Victory (1495-96, Louvre). Detail (see Fig. 16). NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSITION IN RENAISSANCE PAINTINGS |

Figure 20. The rosette crowning the back of the Virgn's throne.

Figure 21. Andrea Mantegna, the altar piece in San Zeno (Verona, 1457). 28 I J. ΡΕΠΤΟΤ

Figure 22. Andrea Mantegna, the altarpiece in San Zeno, detail.

3.2. Non-generic elements: parallelisms and alignments Let us now turn to the detection of non-generic elements sharing strong structuring import. Parallelism plays an important role in this masterpiece as a good way of establishing correlations. We can note: a. The parallelism between Saint George's and Saint Longinus' lances: one broken, the other intact and particularly sharp, one white, the other red. This fact makes it legitimate to compare the two saints (Longinus looking at Saint George who is himself looking at Francesco Gonzaga). For example their headgear make out an interesting geometrical motif: one sole feather tangent to the lance, in Saint George's case, a cusp-formed panache in Longinus' case (Figure 23a). NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSITION IN RENAISSANCE PAINTINGS | 29

b. The extremities of the two lances function as indexes pointing at the same floral motif in the vegetal dome (Figure 23a). c. Saint George's lance is prolonged by Saint John the Baptist's cross (Fig- ure 23b).

Figure 23. (a) Parallelism Saint George / / Saint Longnus. (b) The alignment lance — cross.

d. This cross is perfectly parallel to the one hold by Saint Andrew. This justifies a comparison of their backgrounds: the first cross stands out against the Virgin's mantle, while the second one stands out against the grey and cloudy sky. Hence a correlation between mande and sky, which is confirmed by the fact that in the Madonna of Mercy the mande is identified with cosmos (Figure 24). 30 J J. ΡΕΠΤ0Τ

Figure 24. The two parallel crosses and their respective backgrounds: the interior of the mantle and the

e. Now, still by means of parallelism, Saint Andrew's cross elicits Saint Michael's sword (whose hilt is of course also a cross), which is wrapped in a vertical ribbon progressively changing from white to red. This gives the color red a particular import because it is simultaneously the color of the space above (coral and festoons), the Madonna's robe, the crown- rosette, the space below, Longinus' helmet and lance, and Saint Elisa- beth's arms (Figure 25). NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSmON IN RENAISSANCE PAINTINGS |

Figure 25. Saint Andrew's cross and Saint Michael's sword.

f. As regards Saint Michael, a remarkable compositional feature connects him in the lower leftmost part of the painting to Francesco II: the tangency of knee and doublet, tangency of feet, the astonishing motifs of the spurs and one of the folds in the fabric (Figure 26). 32 I J. ΡΕΠΤΟΤ

Figure 26. Symmetry Saint Michael / Francesco II.

g. We may finally observe that Saint John the Baptist's cross perfectly connects two hands: Saint George's hand which holds the lance, and Saint Elisabeth's hand which holds a rosary that lies in the exact condnu- ity of the cross. We thus obtain a maximally non-generic alignment lance-cross-rosary, with an astonishing separation thigh/face created by a double tangency of the shaft of the cross (Figure 27). NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSITION IN RENAISSANCE PAINTINGS | 33

Figure 27. Saint Johtt the Baptist's cross connects Saint George's hand holding the lance with Saint Etisabeth's hand holding a rosary. Alignment of lance — cross — rosary.

Thus, the simple organization of non-generic parallelisms makes the observer's eye circulate through the whole painting and induces the emergence of several immanent (intrinsic) correlations ready for semiotization and hermeneutical interpretation. It is not our purpose to propose such an interpretation. Our aim is purely methodological and we want only show how rich the non-genericity method can be.

3.3. Jesus' right hand and the cosmos-mantle There exist yet other essential non-generic elements and in particular two elements incarnated by both Jesus' hands. 34 I J. ΡΕΠΤΟΤ

As regards the right hand— which sketches a blessing—, it is perfecdy superposed (non-genericity) on the broach that clips together the two tails of the mande (Figure 28).

Figure 28. The blessing Jesus' right hand is superposed on the broach of the mantle.

Let us now take therefore a closer look at this open mande, which is unusual for seated Virgins with Child. Its edge describes a cusp, a very peculiar geometrical motif that is perfecdy replicated by the red festoons above (Figure 29). NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSITION IN RENAISSANCE PAINTINGS | 35

Figure 29. The cusp of the mantle and the parallel cusp of the festoons.

As I have already remarked, this theme is characteristic of the Madonnas of Mercy, a sublime example of which is the version painted by Piero della Francesca (1445-1462) in San Sepolcro (Figure 30). 36 I J. ΡΕΠΤΟΤ

Figure 30. Piero della Francesca, Madonna of Mercy (1445-1462, San Sepolcro).

It is also present in another famous painting by Piero (unfortunately in a very bad state after its chapel was destroyed by an earthquake in 1785), Madonna del Parto (about 1460) in Monterchi (Figure 31). NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSmON IN RENAISSANCE PAINTINGS |

Figure 31. Piero della Francesca, La Madonna del Parto (c. 1460, Monterchi).

Thomas Mattone (1980), one of the main specialists of Piero, has written a brilliant study of this masterpiece with its two vanishing points, a standard horizontal one and second, vertical one determined by the grid pattern in the interior of the tent. He particularly emphasizes the existence of two parallel cusps, one described by the edge of the curtain and the other by the top of the tent. He even adds a virtual, compositional cusp constituted by the angels' arms (Figure 32). I J.PETTTUT

Figure 32. The three parallel cusps in the Madonna del Parto according to Thomas Mortone (1980).

Mantegna was very well acquainted with both Piero and the theological signification of the tent. It represents the cosmos in which appears the Son from Heaven and houses the Tabernacle and the Arch of Alliance. Figure 33 shows two of Thomas Martone's examples of this configuration: one by Benozzo Gozzoli (1420-1497) in Pisa, another from a Renaissance relief. NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSITION IN RENAISSANCE PAINTINGS | 39

Figure 33. Thomas Mortone's two examples of a cosmos-tent housing a tabernacle, (a) Beno^p Gongoli (1420-1497), Pisa, (b) relief from the Renaissance.

One might wonder - even though it is perhaps a bit farfetched - if the hairstyle of the Virgin, which reproduces the cusp motif, is not a relevant detail in this context (Figure 34).

Figure 34. The cusp-shaped hairstyle of the Virgin. 40 I J. ΡΕΠΤΟΤ

3.4. Jesus' left hand: the carnations and the flowers of the throne In what concerns now Jesus' left hand, it is holding two carnations (a conventional symbol of redemption and Passion, the death on the cross), of which one perfectly covers (non-genericity) a motif on the back of the throne, a ideologi- cal superposition which introduces a prophetic time (Figure 35). We see, therefore, that the two metaphysical symbols: the mantle (cosmos) and the throne (glory), which together glorify the Virgin's body, are unified by the Child's hands.

Figure 35. Superposition carnation-motif of the throne.

Conclusion We have seen in both Raphael and Mantegna that the non-generic elements of a painting operate as a kind of "system of equations". They provide a set of pointers profiling formal and immanent structural correlations that are intrinsically significant because of their non-genericity and that precede any interpretation. If you attribute a semiotic value to one of these correlations by choosing an isotopy (for example a theological one), then step by step, like in the numerical resolution of equations, you semiotize the whole of the work. Another analogy could be suggested here, not to arithmetic and algebra (variables-equations-values), but to biology. The non-generic makeup selects certain elements of a structure, at a preconceptual level, and the hermeneutic interpretation defines the functional signification of these elements. However that may be, we see how such purely phenomenal structures can be spontaneously converted into semiotic structures if they present a selective property such as non-genericity. NON-GENERIC VIEWPOINTS AS A METHOD OF COMPOSITION IN RENAISSANCE PAINTINGS |

References Calabrese, O. (2002). Ducäo e Simone Martini. La Maestà come manifesto politico. Turin: Edizioni Silvana. Desolneux, Α., Moisan, L., & Morel, J-M. (2003). Maximal meaningful events and applications to image analysis, Ann. Statist. 31 (6): 1822—1851. Goethe, J. W. (1798). Über Laokoon. Ästhetische Schriften 1771-1805 (ed. by F. Apel): 489-500. Frankfurt M., 1998. Greimas, A. J. & Courtés, J. (1993). Sémiotique. Dictionnaire raisonné de la théorie du langage. Paris: Hachette. Mantegna 1431—1506. Musée du Louvre, Hazan, Paris, 2008. Martone, T. (1990). Convegno Intemazionale sulla 'Madonna del Parto" di Piero della Francesca. Monterchi. Petitot, J. (1979). Saint-Georges: Remarques sur l'Espace Pictural In J. Petitot, Sémiotique de l'Espace φρ. 95-153). Paris: Denoël-Gonthier. Petitot, J. (1986). Osservazioni in margine alle relazioni di Thomas Martone e Louis Marin. In O. Calabrese (Ed.), Piero teorico dell'arte (pp. 207-210). Rome: Gangemi. Petitot, J. (2004). Morphologie et Esthétique. Paris: Maisonneuve et Larose. Petitot, J. (2009a). Morphology and structural aesthetics: from Goethe to Lévi-Strauss. In Β. Wiseman (Ed.), The Cambridge Companion to Lévi-Strauss (pp. 275—295). Cam- bridge: Cambridge University Press. Petitot, J. (2009b). La non-genericità come metodo di composizione. Testure. Scritti seriosi e schifi scherzosi per Omar Calabrese, pp. 196—206. Tuscany: Protagon Editori. Yuille, A. & Kersten, K. (2006). Vision as Bayesian inference: analysis by synthesis? Trends in Cognitive Sciences 10 (7): 301—308. Peer F. Bundgaard

Toward a Cognitive Semiotics of the Visual Artwork — Elements of a grammar of intuition1

In this paper I introduce some of the principles which rule meaning making in visual artworks — both in regard to construction and interpretation of meaning. The approach is naturalistic in that aesthetic meaning construction is claimed to rest on features and structures which are intrinsically significant for the visuo-cognitive system in plain, everyday perception: the artist exploits such features, and so doing he transforms the automatisms of perception into a rhetoric of aesthetic intuition. Sec- tion 1 of the paper consists of a phenomenological characterization of the aesthetic object as such (in its difference from everyday objects). In the core of the present work, section 2, I give concrete examples (with a general import) of the meaning making tools used in visual art and their roots in everyday perception. The central concern here is to show how purely spatial relations can become significant, or, in other words, how conceptual meaning can be anchored in perception, or, finally, how artists can encode meaning in shape. The final section discusses the general principles ruling the semiotics of the visual artwork and addresses a couple of methodological issues.

CORRESPONDENCE: Peer F. Bundgaard. Center for Semiotics, Aartius University, Denmark. EMAIL [email protected]

0. Introduction As addressed here, the semiotics of visual artworks has a limited scope. Its concern is how meaning making takes place in paintings (or whatever visual art work) and if this meaning making is sufficiendy principled to sustain or justify a theory aiming at defining the laws or principles ruling meaning making in art. The semiotics of the artwork is therefore not (directly) concerned by value, nor by beauty (in whatever conception), nor by, say, the empathy system (the emotional response patterns) activated in the experience of an artwork, even though this may actually play a crucial role in aesthetic experience. In fact, the

1 Full color reproductions of all images and paintings can be found online at htφ://www.cogrütivesemiotics.com (click on "issues overview" in the sidebar and scroll down to #5 to access image files for individual contributions).

Cognitive Semiotics, Issue 5 (Fall 2009), pp. 42-65 TOWARD A COGNmVE SEMIOTICS OF THE VISUAL ARTWORK | 43 genuine object of the present study is not aesthetic experience proper, but the possibility of pictorially constructing meaning in experience.2 Restricting the scope of an investigation tends, in general, to ease things. In the present case, however, this is far from self-evident Indeed, if the claim is that meaning making really does take place in art - i.e., that the artist expresses himself in this and that way with the purpose of obtaining this and that meaning effect — then, obviously, a semiotics of the visual artwork cannot content itself with talking in general terms of artworks and aesthetic experience, it is also, and critically, committed to capturing meaning making such as it takes place in a particular piece of art. The difficulty then consists in developing a general characterization of meaning construction in visual artworks which rests on principles that can be applied to the concrete analysis of a particular painting. The aim of this paper is to present a couple of these principles for meaning making. It will do so within a naturalistic conception of aesthetic experience; i.e., one according to which aesthetic experience is not a mental activity involving modes of cognitive processing foreign to everyday cognition: on the contrary, most of the principles ruling the shaping of meaning in the artwork and triggering the correlative meaning effect in the beholder are rooted in or exploit plain properties of the visuo-cognitive system.3 Yet, the fact that all processing of visual stimuli, aesthetic or not, is taken in charge by one and the same cognitive system or subsystems does, of course, not imply that aesthetic objects are objects just like any other objects (similarly, aesthetic experience is not, as regards its relation to the object, just like any other experience). The first section of this paper is therefore devoted to a brief phenomenological description of the aesthetic object. The second, and most substantial section, will present certain principles of meaning making, defined in terms of the exploitation of the automatisms of perceptiom each of these will be illustrated by concrete analyses, some of which will show how the artist's meaning intentions can be read off the painting as such. In the third section I will explicitly state the methodological principles governing the present approach and tentatively touch upon the relation between everyday and aesthetic cognition as regards the relation between signification and value.

1. The aesthetic object: aesthetic intentionality Aesthetic objects are intentionally shifted object (as the Polish phenomenologist Roman Ingarden (1985) has remarked). Our attitude toward (or "intentional

2 The exact phenomenological account for aesthetic experience is, of course, most valuable. It is just not the central topic of the present paper. I have dealt with the issue, in an Ingardian vein, in Bundgaard 2002 and 2004, cf. also Ingarden (1969, 1985, 1989) - and I shall briefly touch upon it below. 3 In this respect and in others, which soon will be made clear, the present approach is deeply indebted to Rudolf Arnheim's psychology of art (Arnheim 1954,1969). I P. F. BUNDGAARD

set", as phenomenologist use to put it) is not the same as towards everyday objects. In natural or normal experience, i.e., in the ordinary way of intending an object in perception, whatever qualities of the object we perceive will be perceived as aspects specifying the object as a thing, as the physical kind of thing appearing in my perception such as it appears. All color qualities belong to the thing, all qualities pertaining to texture and shape determine that thing: they are indexes of its reality. In an aesthetic attitude, this no longer holds true. Something blue on a painting, or some textural property of it, is not perceived as belonging to the canvas. It is not the canvas which is blue, it is, for example, the sky painted on the canvas. Qualities of entities perceived in an aesthetic mode of perception do not determine a thing, they participate in the presentation of an object (a concrete or an abstract motif). In short, in one set, qualities are accidents on a 2-dimensional plane, in another set, they may be properties of objects displayed in a 3-dimensional represented space (if the painting is figurative). It goes without saying that no matter what artwork and no matter what qualitative property of that artwork can be intended within both the natural and the aesthetic intentional set, and be so by one and the same person. Thus, a conservator can in one moment be rather concerned by a partially detached or otherwise deteriorated quantity of pigment, and in the next moment take one step back and appraise the object presentified4 by that very amount of pigment (thus not seeing pigment, but, say, a hand or a face). The different ways of being intentionally related to the object give rise to Ingarden's distinction between the artwork as a physical thing (Gemälde), and the artwork as a pure intentional object or picture (Bild)} Here is not the place to address the intricate issue concerning the relation between the pure intentional, and therefore mind-dependent aesthetic object (which is relative to a certain type of intentional act), or the physical thing on which it is ontologically founded. Rather, the above distinction serves the purpose of introducing yet another distinction. Obviously, the intentional picture-side of the artwork can be divided into two phenomena or, as we shall see, into two intentional objects: one, which is the motif or the object represented in the painting; another which is the painting as a representing, depicting object.

Edmund Husserl (1980) introduced this distinction in his phenomenology of picture consciousness (refining beforehand, as it were, Ingarden's analysis). He

4 The term "presentify" has roots in Edmund Husserl's phenomenology {vergegenwärtigen) and means here: "make present" (display before the eyes). 5 The differences between the two exceed what I have just mentioned above. For example, the painung, the physical thing, can be accessed and its reality attested through several senses, the picture only through the eyes. The painting can be carried around, the picture cannot. The painung can be seen and recognized as such from any point of view, the picture cannot. Any partial perception of the painting, for example partially covered by a cloth, gives rise to a genuine experience of it; not so for the picture which requires complete perception. TOWARD A COGNITIVE SEMIOTICS OFTHE VISUAL ARTWORK | 45 points to the existence of three objects (that is, three levels at which one and the same thing can be constituted as an object of consciousness):

When we distinguish between object and picture, we soon remark that the concept of picture is here two-fold. The depicted object has indeed a double counterpart: 1) the picture as a physical thing, as this painted and framed piece of canvas, this printed piece of paper, etc. [...]; 2) the picture as the pictorial object [Bildobjek^ which appears through this particular bundle of colors and shapes. Hereby we do not mean the depicted object, the pictorial sujet [Bildsujet], but rather the exact analogon to the fantasy picture, that is to say the appearing object, which is the representative of the pictural sujet (Husserl 1980:19).

We have three objects: 1) the physical picture, the thing made out of canvas, marble, etc.; 2) the representing or depicting object; and 3) the represented or depicted object (ibid).6

We recognize a painting as a Picasso (or a Rembrandt, a Mondrian, etc.) even though we have never seen that specific item. In that case, we of course recognize a general manner of representing, depicting or presentifying things. In short, we recognize a kind of Bildobjekt. This relative autonomy of the presentational layer of the artwork relative to its motif is key to the present approach for several reasons:

- The difference between Bildobjekt and Bildsujet is what immediately makes the object appear as an eminently intentional object, that is: as expressing a meaning intention: the motif is not simply given, but given in a specific way, with this and that shape (partially distorted as in El Greco or in Ba- con), through these colors, in this configuration, by means of these strokes, framed in this and that way, etc. The fact that this difference constitutes the artwork as a semiotic object whose meaning is not coex- tensive with the motif, but also relative to its way of representing it, may explain why we do not spontaneously reject paintings with disgusting motifs (as Aristotle observed it),7 or, inversely, why we may find it excit- ing to contemplate artworks representing dead dull things (Cézanne's apples, Monet's ponds, etc.).

6 Also quoted by Stjernfelt 2007, chapter 14. This chapter is an enlightening critical discussion of Husserl's picture theory. Cf. also Bundgaard 2002 and 2004. 7 This point is not dealt with by Freedberg and Gallese (2007) in their plea for a mirror-neuron system support (or an embodied empathy support) for aesthetic experience. Aesthetic experience may - perhaps should - also be characterized in terms of the emotional arousal triggered by, for example, the objects represented in a painting. This arousal can arguably be said to be underpinned or reinforced by the mirror neuron system. But why then doesn't our brain instruct us to take off like a bat out of hell when we look at mutilations of bodies, putrefying pigs, battle field scenes, cannibalism or whatever disturbing images painters through the ages have found it interesting to depict and museums pleasing to expose? 46 I P. F. BUNDGAARD

- The difference between Bildobjekt and Bildsujet allows the painter to encode all sorts of signification, to shape all sorts of meaning intentions or to create all sorts of rhetorical effects in the painting. Let me just mention one well-known example (Figure 1): Gombrich (1950) and Arnheim (1974) have remarked, if, in Griinewald's crucifixion, Jesus appears as unnaturally tall, this probably does not reflect Griinewald's per- sonal assumptions as regards the real size of the represented person, but rather the importance he assigned to him. Similarly, it is probably not a coincidence that the five pillars of Christianity are all painted in white, and thus pictorially and perceptually grouped even before they are conceptually assembled (Arnheim 1954: 89 - Mary: virginity; the inscription on the cross: kingship; the Bible: revelation; the lamb: sacrifice; the loin cloth: chastity). In section 2, we shall go through a number of examples which illustrate important ways in which painters obtain meaning effects by correlating design features of the presentational, depicting level of the Bildobjekt with conceptual motif-meaning represented at the level of the Bildsujet.

Figure 1. Grünewald. 1506—1515. Rhetorical effect obtained by amplification at the depicting level TOWARD A COGNITIVE SEMIOTICS OFTHEVISUAL ARTWORK |

- The difference between Bildobjekt and Bildsujet implies that the depicting layer of the painting can be perceived (and appraised) per se, relatively (and sometimes entirely) independent of the things it represents. This is what is at stake in the aforementioned type-inclusive case where we recognize a painting as an "Italian Renaissance painting", "a Renoir", "a Malevich", etc. It is also what happens when we judge one caricature to be better than another. However, the autonomy of the presentational, depicting layer is stronger than suggested by the above examples. There are, indeed, cases where we do not simply appreciate or recognize some characteristic presentational style, but where the depicting layer is perceived as such, with its own inherent Gestalt-structure which displays an organization and order that contradicts the one manifested by the motif, the Bildsujet. In such cases one and the same artwork may, as a Bildobjekt, present two autonomous, equally consistent and ordered, but incompatible objects (to the extent that they alternate in perception). A clear case in point is Monet's painting below.8 For reasons discussed in Section 2, an immediate, and probably unavoidable way of perceiving this painting is as a plain figure-ground structure with an abstract grid-like purple configuration on a somewhat orange background. However, the motif requires another way of parsing the perceptual matter. According to the represented structure, the upper part of the purple grid-structure is made out of trees (which, then, are not really purple), whereas the lower part of it represents the reflections of the trees in the pond, and so on. In short, the painting can, as one and the same Bildobjekt, presentify two Bildsujets·, an abstract, gestaltic, purely presentational one, and a concrete, representational one with a recognizable motif. Each of these obtain different, mutually exclusive structures, and the aesthetic experience can thus be described as one of a bi-stable object, flipping forth and back between presentational order and represented layout.

Monet, particularly the later Monet, epitomizes this fact. Cf. Bundgaard (2002) for other examples (Turner, Hammershai) and a more thorough discussion of the presentational and the representational strata in artworks. That analysis owed much to Per Aage Brandt's lectures on that issue. For a more recent development, cf. Brandt 2006. 48 I P. F. BUNDGAARD

Figure 2. Claude Monet, Four Poplars (Metropolitan Museum of Modem Art, New York). The double structure of the artwork. At a presentational level, we see a grid-like abstract Gestalt in a 2D space; at a representational level, we see four trees, a pond, the reflection of the trees, etc. in a 3D space. In Husserlian terms, the Bildobjektpresentifies two Bildsujets; an abstract and a concrete one.

- Finally, the difference between Bildobjekt and Bildsujet reveals an essential feature of the aesthetic object. Aesthetic objects are not only shifted intentional objects, in the sense that their qualities do not specify a thing, but rather presentify an object. There is indeed more to it: while qualities sensed in everyday perception pertain to one thing and give access to one thing, qualities of an aesthetic object (which make out the presentational Bildobjekt) may perfecdy well presentify and give access to two or more perfecdy consistent objects. This is what we have just seen in the Monet- case where the object of perception is either an abstract Gestalt or a natural scenery with each their structural basis. So, what characterizes aesthetic objects is not simply the physical thing {Gemälde)/aesthetic object (Bild) divide, nor the fact that paintings are amenable to different coherent perceptual interpretations, but also that one and the same set of qualities - contrary to everything we know about physical qualities and physical things — can presentify several, equally consistent, but different objects which are genuinely autonomous (they are not just aspects of the same object, or the same object seen from different points of view). This is obvious in one of the Eleven Configurations by Jean Arp (analyzed by Arnheim (1974: 235) (Figure 3). TOWARD A COGNITIVE SEMIOTICS OF THE VISUAL ARTWORK | 49

Figure 3. Jean Arp, woodcut from Eleven Configurations. © Jean Hans Arp / bilkdkunst.dk. As Arnheim has shown the object is amenable to different, all equally consistent representations of the spatial layout of the presented object.

As Arnheim remarks, this lithograph is amenable to several (5) interpretations according to which it shows so many different objects. It can, for example, be considered as showing a three-layered configuration with a large black shape below, a medium-sized white shape upon it and then a small black piece on top of the white one. It can also be interpreted as showing a two-layered configuration with a large black shape below (as in the first case), and then a white shape with a hole in it on top of the black one (in this case the small black piece in the center is part of the large black piece). Or, as another possibility, the object may present a one-layered configuration with a large black shape with a hole in it (through which appears the white ground) and then a small black piece as an island in that hole). Notice that this is not simply an issue about perceiving the same object in different ways as when we perceive or conceive of one and the same thing as, say, either a bowl or a cup. The following Gedankenexperiment shows this. If a thing can be intended in different ways, as a bowl or a cup, and if this thing fell to the ground, then only one thing would actually hit the ground (regardless of the different ways in which it had been intended). Now, if we imagine that the objects presented by Arp's woodcut were to fall to the ground, then very different things would hit it: in the first case we would have to pick up three fall pieces, in the second case one full black piece and one white with a hole in it, and in the third case, one large black piece with a hole in it and one small black piece. Therefore, aesthetic objects may presentify several objects in a non- ambiguous way.9

9 The presentation is non-ambiguous in the sense that it is not an either-ot affair, nor something that is supposed to be settled after more scrutiny. This, of course, makes a huge 50 I P. F. BUNDGAARD

To summarize: 1. Aesthetic objects are intentionally shifted objects: their qualities do not specify a thing (the canvas), but presentify a represented object. 2. There is an essential difference between the represented level of the aesthetic object and its presentational level; between what it shows, and the way it shows it, or between Bildobjekt and Bildsujet. Meaning effects and aesthetic effects can be encoded at the presentational level. 3. The aesthetic objects can be perceived and appraised both at the presentational and the represented level. Each level can possess its own structure which is perceivable and conceivable in its own right. 4. One and the same picture can presentify different objects in a consistent and non-ambiguous way.

I shall now corroborate these claims with a description and a semiotic analysis of a sample of pictures.

2. Means of meaning making: the exploitation of the automatisms of perception The principles governing the automatisms of perceptions are pre-conceptual and pre-reflective. They precede and further object recognition and higher order categorization of shapes as forms, but they are, or course, not themselves tokens of perceptual judgments: they organize the visual matter, thus facilitating the constitution of a perceptual judgment on the basis of a structured percept. What, then, qualifies as a principle ruling of the organization of the visual matter? Among the evident candidates we find Max Wertheimer's different Gestalt laws of form organization (Wertheimer 1923): the grouping of qualitative elements by virtue of their resemblance along a given qualitative dimension (e.g., resemblance in color, local shape, orientation) or grouping by virtue of proximity, symmetry, or construction of shape by virtue of the law of possible continuity. As is clear from the examples in Figure 4, these laws apply on visual matter as such pure shapes, independently of these shapes being or not being recognizable shapes.

difference with respect to everyday perception where simultaneous interpretations of one and the same set of perceptual data are known to be mutually exclusive. Ambiguity results from the incapacity of fulfilling a perceptual judgment on the basis of a given percept. In the aesthetic case, there is no problem in fulfilling perfectly consistent perceptual judgments on the basis of one and the same percept, and therefore no ambiguity. TOWARD A COGNITIVE SEMIOTICS OFTHE VISUAL ARTWORK | 51

Symmetry

(*) OOGOÛÛDOÛOO Similarity

(d) • ••••• Cl06ur·

Figure 4. Principles ruling the organisation of form perception according to Wertbeimer 1923. The simple point is that physical^ speaking it is possible to group line 2 and 3 in row (a), or the árele and the square in row (b) than line 1 and 2 or circles/squares in row (b). Now, we don't, and this points to the existence of higher order principles (Gestalt laws) constraining the construction of penepts on a precategorical basis. Such constraints can be exploited by artists. Thus, in the Monet painting above (cf. Figure 2), the experience of a pure grid-like Gestalt is furthered by the laws of symmetry, nmilarity, proximity, and continuity/ smoothness.

This obviously explains why the above Monet painting lends itself to a double perceptual experience; one representational in terms of trees, pond, reflection of trees, sky, etc., and another presentational, in terms of purple grid Gestalt on a more or less orange background. The latter obtains immediately from straightforward Gestalt processing and grouping of qualities in terms of resemblance of color, proximity and continuity (of the vertical purple lines) as well as symmetry. All in all, everything in Monet's pictorial set-up prompts the eye to consistendy organize the visual scene in such pure Gestalt terms, which of course runs counter to the objective space represented in the picture.10

10 See also that the 3D => 2D collapse is further, and remarkably, prompted by the non-generic alignment of the tree shapes and their reflection in the water. Non-genericity is meant as highly unstable and statistically improbable configurations ("suspicious coincidences") of 52 I P. F. BUNDGAARD

Notice that the tension between the presentational and the representational organization of the perceptual matter is not one that in any satisfactory way can be characterized in terms of '(pleasing) recognition of objects in perceptually noisy environment.' The situation is not a one layered one with certain recognizable objects plus some amount of visual noise." It is rather a genuinely two-layered one with two equally well structured organizations of the visual matter. Hence the fundamental claim to the effect that paintings may, occasio- nally, presentify two different objects in a non-ambiguous way.

2.1. Grouping as a means of encoding meaning in shape and color: constructing semantic relations by means of perceptual correlations The most arduous task of a semiotics of the visual artwork possibly consists in developing conceptual tools to capture the encoding of meaning in the artwork. This problem is of critical importance for figurative paintings with rich motifs. Rich motifs are simply those which represents scenes or characters which are naturally carry meaning for humans, either for biological or emotional reasons (scenes of death, joy, love, sexual arousal, conflict, distress, despair, anxiety, and the like) or for cultural reasons (scenes representing founding religious or historical events, etc.). Within this framework, 'encoding meaning in the artwork' implies constructing a visual or pictorial counterpart to the conceptual issue addressed (or the emotion displayed, the valorization effectuated) by the painting at the representational level. To use Rudolf Arnheim's admirable expression, the problem consists in showing that "an abstract pattern organizes the visual matter in such a way that the intended expression is directly conveyed to the eyes" (Arnheim 1954: 152). In short, what is told to the mind should also

figures or lines in space. Such non-generic configurations often only obtain from one extremely singular point of view and do not resist the slightest variation. Non-generic configurations are intrinsically significant for the perceptual system (they are rare and therefore perceptually salient); this is, indeed, why artists make use of them as means of meaning making (cf. below, and Petitot 2004, 2009). In the above Monet painting, the law of continuity of course forces, the eye to group these elements, but applies only because such an alignment has been produced thanks to a non-generic point of view. As shown in Petitot (2009, the present volume), this 3D => 2D collapse is morphologically analogous to the one experienced in Necker cubes represented from a non-generic point of view. 11 This is what Ramachandran and Hirstein (1999) seem to suggest in their naturalization of aesthetic experience. I am sympathetic with their attempt to establish the close connection between everyday and aesthetic perception. 1 do think, however, that they go astray in trying to establish this link in terms of value, i.e., on their view, what is valuable in everyday perception transposes to art. Thus, elements disliked in everyday perception (asymmetry, non- genericity, etc.) are claimed to be avoided by artists. As regards their claims about symmetry and non-genericity, this is plainly wrong. The idea I marshal here is rather that the naturalization of aesthetic experience should be established in terms of significance (not value). Thus, what is intrinsically significant in everyday perception transposes to art (where it can be exploited to trigger certain meaning effects). TOWARD A COGNITIVE SEMIOTICS OF THE VISUAL ARTWORK | 53 be shown to the eyes, and this is what should be developed theoretically. Notice that this idea bears consequences for the interpretation of artworks at large (while being the condition sine qua non for a genuine semiotics of the visual artwork): it is only by pointing at the formal procedures by means of which an artist concretely organizes the visual matter that it can be claimed that his meaning intention can be read off a painting. A major methodological caveat should be made in that the theoretical tools intended to capture the correlation between organization of visual matter and conceptual representations do not make out some sort of semiological code book assigning this and that pre-established signification to the use of this and that symbolic forming tool. The grammar of aesthetic intuition aims at establishing the means thanks to which meaning can be shaped and encoded in the artwork. It does not make any claim to the effect that a specific meaning effect should follow from the use of a given meaning shaping means. The only claim is that those structures or features which are intrinsically significant to the eye can be used as a support for the representation of higher order semiotic objects and their correlations: by exploiting them in his creative work, the artist elevates the automatisms of perception into a rhetoric of aesthetic intuition. The easiest —and only - way to show this is by proceeding by means of examples. Since it is impossible here to even suggest the range of the possible correlations between perceptual and conceptual organization,12 I have chosen to concentrate on a couple of pictures with a religious motif that address the conceptual problem concerning the relation between contrary terms (for example, the divine vs. the human domain, the infinite vs. the finite, etc.). The point is then that the articulation, and even the transition, between the two contrary domains - a transition or correlation that can only be suggested conceptually - is exacdy the problem that the painting solves pictorially. Consider Vermeer's painting Woman Holding a balance (already analyzed in Bundgaard 2009) (Figure 5).

12 Cf. Petitot (2009), Wildgen (2009), both in the present volume, for more examples. 54 I P. F. BUNDGAARD

Figure 1. Jan Vermeer. Woman Holding a balance.

Arguably, the motif is a reprise of the Annunciation theme with a young pregnant woman gazing downwards while her womb is illuminated by the light from above. The religious motif is profiled by the fact that while she's weighing material goods in the foreground and the lower part of the picture, the painting in the background and die upper part of the picture represents Jesus weighing souls on Doomsday. If this is so, the set-up is the following: the lower part and the represented foreground of the painting constitute the "earthly" domain (epitomized by the action of weighing pearls), whereas the upper part and represented background of the painting represent the "heavenly" domain (epitomized by Jesus' weighing the souls). Moreover, as an Annunciation, the motif of the painting is evidendy the impossible, but nevertheless real continuous passage between the heavenly and the earthly domain (i.e. God's becoming flesh, etc.). Conceptually, this founding mystery of Christianity is easily done with: the case is stated in a book, and then it is up to you to believe in it or not. But how do you "paint" the state, as it were, how is the impossible TOWARD A COGNITIVE SEMIOTICS OF THE VISUAL ARTWORK | 55

transition or articulation constructed in vision? What is the pictorial counterpart to the conceptual issue in this case? Granted, in the organization of the painting the problem to be solved consists in articulating the lower and the upper part (or foreground and background of the picture). Vermeer effectuates this in three ways: (a) chromatically, (b) morphologically, and (c) compositionally.

(a) Chromatically. The robe covering the pregnant womb and the curtains through which the light from above shines (a conventional symbol in many Annunciations) are colored in exactly the same red-golden hue (which again by convention is considered the divine color). The key point here is (just as in the Grünewald example mentioned above) that by virtue of the principle of grouping, these elements are visually correlated thus suggesting an essential correlation between womb (below) and curtain-light (above). Hence the principle, what fires (qualitatively) together, wires (conceptually) together.

Correlation by color identity

Figure 2. Jan Vermeer, Woman Holding a Balance. Correlation of top and bottom, background and foreground: curtain above and robe below have exactly the same color. They are, thus, correlated in vision by virtue of the principle of grouping (similar qualities).

(b) Morphologically. The young woman's weighing is not only conceptually reproduced by the motif of the painting in the background. Also, the very shape of the visible part of the painting reproduces the shape of the balance in the foreground. The correlation here is constructed in vision by means of morphological resemblance. 56 I P. F. BUNDGAARD

Figure 3. Jan Vermeer, Woman Holding a Balance. Morphological correlation: the shape of the balance is reproduced in the visible part of the painting in the background. Foreground and background are correlated in vision by means of grouping by shape resemblance.

(c) Composition. Finally, the painter has succeeded in concretely fusing fore- and background in vision. By choosing a non-generic, structurally unstable, and therefore visually significant point of view, the handle of the balance in the foreground is perfecdy aligned with the ledge of the frame in the background—indeed to the extent that handle and ledge are locally merged and perceptually inseparable from the each other. The mystery of the continuous passage from the heavenly to the earthly is constructed in vision by means of continuous qualitative passage between fore- and background (cf. Arnheim 1969 for a similar conclusion). TOWARD A COGNmVE SEMIOTICS OF THE VISUAL ARTWORK | 57

Figure 4. Jan Vermeer, Woman Holding a Baiarne. Non-generic alignment of handle in foreground and ledge of the frame in the background. The non-generic alignment creates a localfusion offore- and background, a visualpathway between the two.

It is worthwhile noticing that such morpho-qualitative correlations seem to be standard (but not for that matter trivial) means of pictorial meaning making. In Mantegna's ha Madonna della Vittoria (analyzed in detail by Petitot 2009, present volume) the same triad of chromatic, morphological, and compositional correlations can be observed: (a) morphological duplication of the cusp-like shape of Mary's coat below and the festoon above (cf. Petitot 2009: 35); (b) chromatic duplication of the golden-red color of Mary's coat below and the festoon above; (c) compositional correlation by means of the complex (and in a sense contradictory) mediating shape, namely Longinus' lance, which is both connected to the regicide and to the very master plan: in dying, Christ is indeed supposed to enable the very transition between earth and heaven which is the underlying theme of the painting. This complex mediating role of the lance is 58 J P. F. BUNDGAARD

further emphasized by the fact that it shares exactly the same (divine) chromatic value as Mary's coat and the festoon.

(a) chromatic correlation

(b) morphological correlation

Figure 5. Andrea Mantegna, La Madonna della Vittoria (1496). Three morpho-qualitative correlations between semantically speàfied spatial domains (bottom: earthly domain; top: heavenly domain). The procedure is the same Vermeer resorts to.

The above are examples of how semantic relations are constructed in vision as perceptual correlations. As already suggested, they do not exhaust the list of formal procedures by way of which painters organize visual matter with the purpose of obtaining either a pictorial counterpart structure to the represented motif (as Grünewald, Vermeer and Mantegna) or an independent Gestalt- configuration which can be appraised independently of the represented scenery (as in Monet). To mention one other procedure, it may operate solely on the level of shape and encode given semantic value at this level. Hoffman (1998; Hoffman and Singh 1997) has remarked (as to the saliency of parts and the TOWARD A COGNITIVE SEMIOTICS OFTHE VISUAL ARTWORK | 59 reconstruction of depth in vision) that convex points on 2D-contours are computed as saddle points (with two directions of curvature) on 3D volumes. Convex points thus encode depth and volume, and the sharper, more cusp-like they are, the more volume do they encode (and the more salient or protruding are the represented parts). Arguably, this is what Picasso exploits when expressing physical exhaustion and existential hardship (in his paintings from the "blue" period) by diminishing the saliency of body parts and thereby proposing a distorted representation of the human body from which the original layout is difficult to retrieve (as in figure 10). Or, on the contrary, when he exalts the vitality and sexuality of the female body by maxi- and magnifying the saliency of its parts (as in figure 11).13

Figure 6. Pablo Picasso, Woman Ironing, and The Old Guitarist. © Succession Picasso / billedkunst.dk. Deletion of part saliency obtains through reduction of convex points, smoothening of concavities (both of which encode volume) and creation of unnatural salient body segments (e.g. the section leading from upper left shoulder to the neck).

13 Cf. also Leyton (2006) for an original theory of the encoding of meaning in shape. 60 I P. F. BUNDGAARD

Figure 7. Pablo Picasso, Nude in a brown Armchair (1932). © Succession Picasso / billedkunst.dk. Maximisation of part saliency and body volume by means of sharpened concavities. It is worthwhile noticing that the volume engendering procedure outweighs, at first glance, the unnaturally flattened, compressed 2D means of depiction. The paintings in figure 10 and the one here are thus duals in a stronger sense: the pictural object (the manner of depiction) in figure 10 is realistic, but it conveys an unnaturally meager 3D body, whereas the pictural object in figure 11 is absolutely unnatural and unrealistic while nevertheless triggering a realisticfull-fledged 3D representation of the female body.

3. Concluding remarks: principles governing the semiotics of the visual artwork In conclusion, I shall briefly state some of the principles on which the present outline of a cognitive aesthetic semiotics rests.

1. Painting is an activity which consists of constructing meaning in vision, that is, in presenting shapes (and relations between shapes) with qualitative properties which are perceptually (pre-conceptually) significant. Meaning making in visual art is primarily fulfilled at the presentational level in the Bildobjekt as defined above, or to use Brandt's felicitous expression (Brandt 2006) in terms of what happens on the canvas, not simply in terms of what is represented by the painting. A major task in figurative art thus consists in creating perceptual counterparts to the conceptual relations or representations represented in the motif of the painting. TOWARD A COGNITIVE SEMIOTICS OFTHE VISUAL ARTWORK | 61

2. We do not have different visuo-cognitive systems, say, one for plain perception and another for aesthetic perception. If painters are to construct meaning in vision, then it follows that they can exploit only those features which are intrinsically significant for the visuo-cognitive system as such, or they can exploit only those principles of form organization which govern perceptual form organization at large (grouping). 3. The grammar of aesthetic perception consists of such features and principles of form organization which allow many devices for possible meaning making.14 In the present paper we have seen meaning effects which rest on the pre-conceptual Gestalt principles of form organization; the exploitation of the signification inherent in non-generic point of views or configuration; and the encoding of meaning in pure shape. 4. We have seen, following Husserl, that it is an ontologcal property of any visual sign (any icon in Peirce's sense) that it consists of a presentifying stratum (the Bildobjekt) and a reference or a representing stratum (the Bildsujet). Pictures represent something and do so in some specific manner. This holds true for all pictures of whatever sort. Now, even though meaning making in visual art essentially takes part on the level of the Bildobjekt (correlated with the Bildsujet, the motif), this does not obtain automatically, necessarily or for any, say, a priori reason. Pictures may perfectly well be flat representations: pure illustrations or just bad or uninteresting pictures. Only when painters give pictorial shape to their (aesthetic) meaning intentions (with tools as the above) does the presentational layer of the painting become semiotized, as it were. The painter's craftsmanship is in the semiotic shaping of the Bildobjekt, not (only) in the skilled drawing of the Bildsujet. 5. Similarly, the constitution in perception of the painting as a pure piece of presentation (irrespective of what is represented), and the ensuing alternation between 2D and 3D interpretations of it (as in Monet), is not something that obtains automatically. The depicting stratum of the painting, the Bildobjekt, must be organized in a principled way in order to give rise to an experience of a detached, autonomous Gestalt presentation. So the same chorus from above can be sung with respect to this: only when painters organize the depicting matter of the painting in certain ways (with tools as the above, and first and foremost "qualitative grouping" in terms of stroke, color, proximity as well as "non-genericity"), does an autonomous presentational Gestalt structure detach itself from the Bildob- jekt.

1» This, of course, does not imply that the semiotics of the visual artwork is reducible to such features and principles (or to the identification of the way they are used in a given painting). It may take a considerable amount of background knowledge (religious and cultural frames, etc.) to get the interpretation of a picture right; still the meaning intention of a painter is in general not encapsulated in the motif, but in his way of reactivating the motif. 62 I P. F. BUNDGAARD

6. Moreover, if it can be assumed that the semiotic (and aesthetic) impact of the painting depends on the way in which the painter organizes the presentational layer of the artwork, it is also evident that there is not just one type of interaction between presentation and representation in the work of art. A possible task for future research in this domain would consist in typologizing the prototypical relations between presentation and representation. In need of mention are two possible extremes of the typological continuum and one intermediary case. On the one side, we'll have the pure alternation between autonomous Gestalt presentation and representation (i.e., fully counter-structural exploitation of the automatisms of perception) as in Monet or Turner. Here, the meaning effect is probably purely aesthetic, pertaining to the ordered bi-stability of the object. At the other end of the continuum we will have paintings which exploit the two-dimensionality of the medium to establish significant spatial relations which pertain only in perception (i.e. relative only to what happens on the canvas), not in the space represented by the painting. This is obviously the case in Raphael's St. George (analyzed by Petitot 2009, this volume), in the Vermeer and in the Mantegna, both commented above (figures 5 and 9). In these three cases, as well as in coundess others, a spatial or chromatic relation, which exists only as a "phenogram" on the canvas (i.e., as a perceptual event which only obtains on the surface of the picture), not as an "ontogram",15 a real relation represented by the painting, figurativizes or gives pictorial shape to a conceptual correlation. This way of operating on the Bildobjekt in order to create significant spatial relations clearly does not fully detach the presentational layer of the painting as a Gestalt per se, and the 2D-3D alternation does not obtain. Without wanting to inflate the amount of concepts unnecessarily, one could nevertheless suggest that whenever we obtain a perceptual correlation between elements located on the surface which arguably supports or pictorially rearticulates the meaning conveyed by the motif, we have a phenogram that functions as a semiogram, i.e. a significant perceptual correlation. Part of the task of the semiotics of the visual artwork is to conceptually determine or motivate the conversion of phenograms into semiograms - both in general and in a concrete artwork.

In between the two extremes, different types of significant relations between the pictorial organization of Bildobjekt and the semiotic/conceptual import of the Bildsujet may be captured. A remarkable case is the Constable example below16 where the painter - from the preceding study to the actual painting —

15 The concepts of phenogram and ontogram arc due to the German Gestalt theoretician Edwin Rausch and quoted by Arnheim 1954: 419. 16 Example from Svend 0stergaard. TOWARD A COGNITIVE SEMIOTICS OF THE VISUAL ARTWORK | 63 changes the position of a tree with respect to a leaping horse in the represented landscape, arguably because the tree, as a shape and in virtue of its location right next to the leaping horse is automatically, preconceptually interpreted as a barrier, as something that blocks the élan of the horse. The case is of course complex because one element (the tree) is interpreted as a pure, schematic 2D Gestalt blocking the 3D represented leaping horse. In short, something happening on the canvas has import on the interpretation of what is represented by the painting; therefore Constable changed its location, and straightened, thus strengthening the contrast between the horizontal momen- tum of the figure and the verticality of the ground.

Figure 8. John Constable, study and orignal version of Leaping Horse. Constable chooses to displace the tree from the study and to place it behind the leaping horse in the original. Probably because the tree, 64 I P. F. BUNDGAARD

as a pure 2D Gestalt is automatically interpreted as blocking the movement of the horse. This is an example of a case where a pure phenogram (the position of a shape (the tree) with respect to another shape (the hone)) carries a meaning (in this case an unwanted one). Notice also how Constable in the final version chooses to straighten out the tree, thus produríng a salient contrast between the static verticalità of the tree shape and the dynamic horiapntatìty of leaping horse.

7. Finally, the present approach calls for one major clarification: what about the alleged grammar, what is its status? How is it to capture meaning making in artworks at large, not only in chosen samples? Well, in a sense it is not because its scope is restricted. The present approach, the elements of a grammar of aesthetic intuition in general introduced here, do not rule meaning making in visual art or delimit the domain of possible meaningfulness in art from the domain of senselessness or aesthetic un- grammaticality. They are introduced as pertaining to a weak, semiotic system: they are not presented as constituting a closed list of meaning making tools which are necessarily, fully or partially, instantiated in a given artwork. Rather, they are considered as pictorial symbolic forming devices (cf. Bundgaard 2009) at the disposal of any given art maker who may choose to make use of them or not.

The binding element of this weak semiotic system of meaning making devices is the fact that all elements of this type of grammar17 are elements rooted in everyday perception. To use Patrick Colm Hogan's expression (Hogan 2003), they are "maximized patternings" of features, relations that are intrinsically significant in everyday perception, or they are maximized exploitations of principles ruling the organization of the perceptual matter in everyday perception. As such, they are good tools when it comes to theoretically establishing how meaning can actually be articulated in the visual medium (and not simply referred to, allegorically, conventionally, symbolically, by the elements represented by the medium); i.e., they are good means to constitute the visual artwork as a genuine type of semiotic object with its own organizational principles. Now, the semiotic system is weak — it does not claim to be exhaustive — but it follows from the above that it defines strong constraints on interpretation, i.e., on any local semiotics of a work of art. The constraint is such that whenever an artwork is claimed to mean something, express some signification, trigger some meaning effect or have some purely aesthetic effect, it must be possible to show, with reference to publicly accessible features of the painting and knowledge about human perception, cognition, or culture, how such meaning effects are constructed in vision. Any element that meets this strong condition, by means of which this correlation is established, qualifies as an element of the

17 Nota bene: a grammar which does not exclude other structures or elements likely to trigger other types of meaning effects: if I paint the Pope as a punk, this is a straightforward thing to do from a purely visual point of view, not so from a cultural point of view. TOWARD A COGNITIVE SEMIOTICS OFTHE VISUAL ARTWORK | 65 weak semiotic system of the artwork. In this paper, I have focused on elements like "qualitative grouping", "non-genericity", "part saliency", and in general on the Bildobjekt-Bildsujet interaction.

References Arnheim, R. (1954). Art and Visual Perception: A psychology of the creative eye. Berkeley: University of California Press (special edition, 2004). Arnheim, R. (1969). Visual Thinking. Berkeley: University of California Press. Brandt, P. A. (2006). Form and meaning in art In M. Turner (Ed.), The Artful Mind (pp. 171-188). New York: Oxford University Press. Bundgaard, P. F. (2002). Presentation and representation in art - ontic and gestaltic constraints on aesthetic experience. Visto 7 (1-2): 187—203. Bundgaard, P. F. (2009). The grammar of aesthetic intuition. Synthese. 10.1007/sll229- 009-9631-8. Available at: , Retrieved 13/09/2010. Freedberg, D. & Gallese, V. (2007). Motion, emotion and empathy in esthetic experience. Trends iti Cognitive Science 11 (5): 197—203. Husserl, E. (1980). Phantasie, Bildbewusstsein, Erinnerung. Zur Phänomenologie der anschaulichen Vergegenwärtigungen. Husserliania XXIII. The Hague: Nijhoff. Gombrich, E. H. (1950). A History of Art. London: Phaidon Press. Hoffman, D. D. (1998). Visual Intelligence. How we create what we see. New York-London: Norton. Hoffman, D. D. & Singh, M. (1997). The saliency of parts. Cognition 63: 29-78. Hogan, P. C. (2003). The Mind and Its Stories. Narrative universals and human emotions. Cambridge: Cambridge University Press. Ingarden, R. (1969). Erlebnis, Kunstwerk und Wert. Tübingen: Max Niemayer Verlag. Ingarden, R. (1985). Selected Papers in Aesthetics. Munich: Philosophia Verlag. Ingarden, R. (1989). Ontology of the Work of Art. Athens, OH: Ohio University Press. Leyton, M. (2006). The Structure of Paintings. Wien-New York: Springer. Petitot, J. (2004). Morphologie et esthétique. Paris: Maisonneuve et Larose. Petitot, J. (2009). Non-genericity as a method of composition in Renaissance art, Cognitive Semiotics #5 (present issue). Stjernfelt, F. (2006). Diagrammatology — An investigation on the borderlines of phenomenology, ontology, and semiotics. Dordrecht: Springer. Wertheimer, M. (1923). Laws of Organization in Perceptual Forms. In W. Ellis (Ed.), A source book of Gestalt psychology (pp. 71-88). London: Routledge & Kegan Paul (1938). Wolfgang Wildgen

Geometry and Dynamics in the Art of Leonardo da Vinci1,2

After a short biographical sketch in chapter one, chapter two explains how Leonardo contributed to a "semiotics of art". Two central concepts of his analysis of art are "geometry" and "dynamics". The application of these two notions to the fresco-painting "Last Supper" and to sketches for it is the topic of chapter three. The geometrical solutions for the distribution of 13 persons are discussed in relation to the iconographie tradition and to the art of some followers of Leonardo (e.g., Dürer). Chap- ter four speciali2es on figurai composition and valence patterns in a series of works by Leonardo (Epiphany, "Virgin in the rocks" and finally "St. Anne"). In the analysis of bodily configurations, gestures of grasping, reaching, and gaze-directions a semantic structure comparable to valence-patterns (case-frames, scenes) in sentences or short narratives is uncovered. In chapter five these patterns are related to the models from dynamic systems theory. The final chapter asks if the geometrical and dynamical features discussed are also relevant in the analysis of abstract art.

CORRESPONDENCE: Wolfgang Wildgen. Languages and Literary Studies, University of Bremen, Germany. EMAIL [email protected]

0. Introduction The term "art of Leonardo da Vinci" refers first to his famous paintings, of which few are finished and well conserved, and to the sketches and drawings which prepared his paintings and documented his scientific studies. Secondly, his theoretical thoughts on art and aesthetics have been collected after his death by Francesco Melzi and gave rise to the "Trattato della pittura" (codex Urbinatus 1270).3 On the basis of the theoretical positions found in Leonardo's

1 Full color reproductions of all images and paintings can be found online at ht φ://www.cogniuvesemioucs.com (click on "issues overview" in the sidebar and scroll down to #5 to access image files for individual contributions). 2 The topic of this paper was treated in lectures at Aarhus (Denmark), Ijmoges (France) and Urbino (Italy), Kassel, and Frankfurt/Oder (Germany). The publications W ildgen (2004, 2005 and 2006) in French and German treat specific aspects of the topic. 3 Many of his original texts illustrated by sketches and drawings have been edited since; cf. MacCurdy (1977), Clayton (1996) and Codex Ixicester (2007). For specific themes we refer the reader to the critical edition of the "Trattato della pittura" by Pedretti (1995).

Cognitive Semiotics, Issue 5 (Fall 2009), pp. 66-92 GEOMETRY AND DYNAMICS IN THE ART OF LEONARDO DA VINCI | 67

"trattato" and in his notebooks (cf. MacCurdy 1977) I shall try to elaborate the semiotics of art, which is explicit and implicit in the work of Leonardo da Vinci.

1. What is "Semiotics of Art" in the context of Leonardo's work? 1.2. The semiotics of painting, music, poetry, and science In his 'Trattato della pittura" Leonardo relates the art of painting with poetry, music, and science as an art based on mathematics (mainly geometry). If Leonardo thinks that the painter is master of all types of art, i.e., that painting is not only a science but that it has aspects of poetry and music, he presupposes a common basis for all four of the following domains: painting, poetry, music, and science. This means that he refers to a universal symbolic activity of man. It is the universality of Leonardo which makes him a semiotician, and it is his life-long reflection on the principles and on the "science" of painting ("scienza della pittura", second part of his treatise) that makes him not only a theoretician of art but also of sign-usage. In the movement of Renaissance artists from a cultural practice, learned in the workshop of painters, musicians, and poets, to a reflection on the universal principles underlying this practice and the conse- quent development of new practices going beyond the "maniera" of their predecessors, Leonardo establishes the stage not only for the rapid further development of painting and art, but also for the rapid evolution of science. Although Copernicus' "Commentariolus" began to circulate only after 1515, Leonardo may stand for the new generation of intellectuals in the time of Copernicus and the "modern" civilization of art and science in the 16th and 17th century, which is the basis of contemporary science. In his treatise Leonardo states that the objective of painters is to represent mainly two things: man and his mind.4 The nature of man becomes visible, and therefore, accessible to the eye in the different movements, and in the proportions of his body parts. In order to represent man and his mind the artist must first create a pictorial space, which is the foundation for the topic of the painting. The basic technique rediscovered and further developed in Renais- sance time is called the "linear perspective"; i.e., the artist must be able to represent the third dimension with the means of a pictorial plane. Secondly s/he must consider light and shadow before placing objects in space. Finally, landscape, sky, objects, animals, and persons included in the painting (mosdy individuals or groups of individuals) must be arranged in space relative to light and shadow. The central concern is therefore the composition of the topic and the choice of those postures which are able to represent the motion and the mind of the central persons. In the context of cognitive semantics (cf. Wildgen 2008a for an overview) we could consider the following levels of semiotic analysis:

4 Cf. Pedretri(1995:§180,p. 219). 68 J W. WILDGEN

- The space (time) of the scene depicted (perspective, light/shadow, outfit of the scenario). - The thematic persons (the Virgin alone, together with Jesus, with Jesus and John, with Anne and Jesus). These thematic persons (or animals, e.g., a lamb) constitute a relational schema which is represented by their relative positions in space (e.g., the arrangement of aposdes at the table), the static relations (e.g., the Virgin sitting on the knee of St. Anne) and the movements (e.g., the relative movement of the head in relation to the trunk, the gestures of the hands, and the direction of the gaze).5 - A narrative content related to a known episode, e.g., the moment when Jesus just said that one of the aposdes will betray him in the context of the "Last Supper". The scene may be identified as one moment resulting from a series of prior events and having specific (known) consequences.

In the following I will analyze the two groups of paintings of Leonardo da Vinci in the order of these three levels.

1.2. Space, light/shadow and the outfit of the scenario in Leonardo's paintings The mastery of perspective is an inheritance of the following artists of the 15th century: Brunelleschi, Masaccio, Uccello, Mantegna, Bramante, and Piero della Francesca. Leonardo has left no special treatise on perspective (it was probably lost), but his remarks on light and color show that he presupposed a theory of perspective similar to that handled by Alberti and Piero della Francesca. Leonardo's basic aim is not so much the scientific reconstruction of our way of seeing but an optimal rendering of "relief' by the use of light and shadow.6 Rules in painting concern mainly the domains of restriction and not the domains of liberty; i.e., they concern the transition between a bad towards a good painting. Beauty asks for an adequate management of light and color for giving the proper significance/relief to the central topics (cf. Clark 1958: 76 f.). The relief cuts the face against a light background (e.g., a window) with parts of the face lost in the shadows (cf. Pedretti 1995: § 86f). In relation to early Renaissance the perspective is no more a mean to render realistically the "window of the eye"; the "relief' is rather a figure of significance, of relevance. Leonardo's art goes beyond mimesis of nature and discovers the contribution of space, light and shadows to the meaning-profile presented to the viewer. The beauty of a scene is this extra added to recognized conventional meanings."

5 Cf. for a detailed analysis of the gestures in Leonardo's work W'ildgen (2005). 6 Cf. Pedretti (1995: 199). His specific technique may be due to the fact that he had learned it in the workshop of the sculptor and painter Verrocchio (1435—1488). 7 The eminent role of limits and their gradual nature may be associated with catastrophe theory (a model of limits and transitions). In Petitot (2004: 55—57) the feature called "non- genericity" is understood as fundamental visual significance; i.e. if a figure is able to change GEOMETRY AND DYNAMICS IN THE ART OF LEONARDO DA VINCI | 69

Leonardo's aesthetics are nearer to Cézanne than to classicism because the creation of meaning dominates the representation (imitation) of the external world. Although in his writings Leonardo is scientific and argues mathematically using geometrical constructions, his painting is "impressionistic" in the sense given to this term by Cézanne and his followers (cf. Merleau-Ponty 1989). However, Leonardo does not "lose the objects" and for him the superiority of painting over poetry is founded in its immediate correlation with nature, which creates a degree of trueness and security not accessible by language.8 By highlighting movement and accident, Leonardo asks for an interpretative activity of the viewer based on his experience. Movement and action may be extrapolated from an instantaneous picture to a process which has caused it and which will bring it to a proper end later. The interpretation of snapshots of motion is most prominent in human bodies, in the gestures of the hand, and in the postures of the head. Mimic is represented in the painting because it contributes to a language of the body which is so restricted in its categorization that one instance, one moment may stand for the whole (episodic) gestalt.

1.3. Geometry and dynamics in Leonardo's semiotics of art Under the label "geometry" I will deal with the first part called "linear perspective" by Leonardo.9 It concerns primarily the different shapes of objects at different distances. The painting is like a plane inserted between the eye of the viewer and a scene. Any object may be conceived as the basis of a pyramid pointing to the eye and cut by the plane of the painting. Geometry is involved also in the proportions of parts and the whole, e.g., the whole groups of the apostles (with Christ) and the subgroups they form, the whole body and its parts, such as the head, trunk, and limbs, and also the proportion of the whole room where the supper takes place, from the ceiling, to the windows, the table, etc. (as in the case of the "Last Supper"). This type of geometrical proportion may be arithmetically expressed (e.g., by fractions) and can thus be related to musical harmonies. Dynamics in Leonardo's thinking are Aristotelian, i.e., the impetus trans- ferred by the mover to the moved object diminishes with time and finally goes

its meaning due to small changes in perspective, illumination etc., it is highly significant for attention and perception. Cf. also Bundgaard (2009). 8 Cf. Pedretti (1995: 134). In the modern context of elaborated theories of vision and the invention of photography, the complex activities of Leonardo as drawing anatomist and physiologist, and as observer of the dynamics of machines and bodies are redistributed into the sub-fields like scientific photography and creative art work. 9 Leonardo distinguishes: Linear Perspective, Perspective of Color and Vanishing Perspective; cf. MacCurdy (1977, vol. II: 222 f., translated from MS 2038 Bib.Nat 18r). The perspective has itself a dynamical interpretation as it is associated with an energy which has its seat in the eyes. I W. WILDGEN

to zero. It forms another type of pyramid now pointing to the state of rest."1 In the Codex Adanticus Leonardo says:

Speak first of the movement then of the weight because it is produced by the movement, then of the force which proceeds from the weight and the movement, then of the percussion which springs from the weight of the movement and often the force (C.A. 155 v.b.; cf. MacCurdy 1977, vol. I: 477).

Force I define as an incorporeal agency, an invisible power, which by means of unforeseen external pressure is caused by the movement stored up and diffused within bodies which are withheld and turned aside from their natural uses (MS A, Institut de France, 34v; cf. MacCurdy 1977, vol. I: 493).

The mathematical and physical horizon is still that of Aristotle and his medieval and Renaissance commentators11; only after Galilee and philosophically after Hobbes (e.g., in his "Leviathan", 1659) will these types of dynamics be overthrown. As I shall deal primarily with a kind of psychological dynamic shown in the composition of Leonardo's painting this difference will be of minor importance. In bodily movement, any movement of one body-part has a counterpoise in another body-part. The underlying law is that of the arms of a balance or of levers in general.

The arms of the balance make of themselves a counterpoise from the one to the other; which counterpoise will have with these arms as many varieties as the proportions of these arms will be varied (Forster Bequest MS. II, 155v.; cf. MacCurdy 1977, vol. I: 573).

The balance, the weight and counterpoise mean in painting that a single body is in balance if the weight of the movement of one part of the body, e.g., the head, has a counterpoise in another, e.g., in the movement of the shoulders or the trunk. The balance could be easily realized if all bodies were static. But this would make them "wooden", i.e., unanimated. The painter who wants to show the mind of the persons in the scene must show them in movement, and the balance of a person or group of persons has to be a dynamical balance. The primacy of movement is grounded in the primacy of the mind. The painter should represent these actions as an expression of the mind:

Every action must necessarily find expression in movement.

10 The semantics of "force" has been analyzed by Talmy (cf. W'ildgcn 2008a: chapter 4) and in catastrophe theoretic semantics (cf. Wildgen 2008b for a recent discussion). 11 The medieval authors are Grosseteste, Bacon, Yi'itelo and Pecham; cf. Frosini (1998a: chapter 7)· GEOMETRY AND DYNAMICS IN THE ART OF LEONARDO DA VINCI | 71

To know and to will are two operations of the human mind. To discern to judge to reflect are actions of the human mind. (Codex Trivulziano, 65a; cf. MacCurly 1977, vol. 1: 65)

The two basic pillars of Leonardo's semiotics of art are:

- Perspective (linear, of color, vanishing), which we shall deal with under the topic oi geometry. - Dynamics (force, weight, counterpoise, balance, movement, percussion, etc.).

In the case of Renaissance art there is a consciousness of the geometrical and even technological relevance of art (Edgerton 1980 calls the Renaissance artist therefore a "quantifier"). Moreover, geometry is considered as a practical science, nearer to craftsmanship than to natural philosophy (= physics). I will ask in the last chapter if geometry and dynamics are also relevant for the understanding of modern art.

2. Stability and dynamics in Leonardo's "Last Supper" The general disposition is given by the rectangular table on which all thirteen persons (Christ and his twelve aposdes) are placed almost in a line (two of them are sitting or standing to the right and left of the table). The total surface of the wall in the dining room is subdivided into three layers: a basic layer including a door which has been broken into the wall later, the painting of the Last Supper and three painted arcades above. The middle layer showing the Supper produces the illusion of a deep hall with rectangular tapestry on the sides and three openings going into a landscape. The linear perspective points to the head (the right ear) of Christ. The body of Christ with his extended right and left arm forms the central pyramid. Jesus' ear is significantly related to the narrative content. Jesus has just uttered that one of his pupils will betray him, and he listens to their answers.12 His head is slighdy decentered in relation to the open door in the background, which puts his face into relief against the landscape. The upper border of the landscape defines a line on which the eyes of Christ are placed.13

12 Luca Pacioli, the friend of Leonardo, gave the interpretation that Christ has just said: "Unus vostrum me traditurus est" ("One of you will betray me") and that Christ in this moment accepts the beginning of his martyrdom. 13 In the restored painting it became clear that the eyes are at the level of the nearer landscape, i.e. on a line with the tower of a church to the left of Jesus (cf. Baralon & Marani 2001: 157). 72 I W. WILDGEN

2.1. Leonardo's painting in the sequence of traditional treatments of the topic The geometrical arrangement of the thirteen actors in the scene has a basic symmetry: Christ versus twelve apostles, with six of them sitting to his left and six to his right. The linear arrangement, which includes Judas the traitor, is new. In most paintings that have treated the same topic before, Judas is sitting on the other side of the table and turns his back to the viewer. This is the case in Leo de Castagno's Last Supper (1447) shown in Figure 1. Even in Ghirlandaio's painting made in 1480 (cf. Wölfflin 1904: 26 f.) and in Perugino's painting for Sant'Onofrio in Florence around 1490 (cf. Luchinat 2001: 48) this traditional constellation is still used.

Figure 1. Leo de Castagno 's Last Supper (1447).

Different sketches by Leonardo (ca. 1493/94) show that he originally aimed at a more traditional composition where Judas is sitting with his back to the viewer and John is sleeping on the table in front of Jesus. In a partial sketch Judas is standing up to grasp at the bread according to the dictum of Christ that the traitor is the one who first takes the bread. In the final version Judas is still grasping at the bread while he holds his purse with the other hand. The final solution integrates John and Judas into the line of twelve aposdes and allows four groups of three persons around Christ. John and Judas are in the same group and form an opposition inside this group. GEOMETRY AND DYNAMICS IN THE ART OF LEONARDO DA VINCI | 73

Figure 2. The whole painting in the restored shape.

Leonardo's subdivision of a group of entities on a line may be compared to a system of tones on a scale, and Leonardo has also said that since music is composed following the proportions, the same intention could be realized in a painting (cf. Braunfels-Esche 1984: 113). The geometrical features of the painting become significant by their deviation from the standards of illusionist representation or common visual experience:

1. The viewpoint is above the heads of the real viewers in the hall; therefore, the scene on the wall becomes similar to a scene in a theater. It is explicitly a symbolic representation of the Last Supper.14 2. The table fills the whole breadth of the hall φι the painting) and leaves the room behind almost empty; i.e. the table is the scenario of the narrative content. 3. The persons and the associated gestures and mimics are made as large as possible. They could not sit down and eat at the table, and they have to turn their side for the viewer in order to be fully visible. Leonardo thus reduces the realism of his painting in order to maximize its expressive power. 4. The "semiotic" body-parts: hands and faces are dramatically emphasized by the variability of their forms.

14 As a basic rule of linear perspective the central point attracting the rays of the perspective should be at a distance from the floor corresponding to that of a viewer, i.e., 150 to 180 cm (cf. also Kemp 2001). Leonardo has given an illustration of such a stage-perspective in the Codex Huygens, Folio 119; cf. Panofsky (1940: 78). 74 I W. WILDGEN

2.2. Motion, percussion, and vortices in the central scene of the painting Thus, geometry, perspective and proportion are subordinate to the expression, which is the message of the painting; the geometrical "mistakes" become iconic signs constituting the skeleton of a narrative. The dynamics of the painting are also partially narrative. The utterance of Christ "one of you will betray me" is a force, and the effect of this force creates a "percussion" in the group of apostles. Like a shock-wave it hits the two groups most strongly that are sitting at the right and left of Christ and to a lesser degree the exterior groups. If we consider the nearer groups, Jack is pushed back, whereas John, although displaced in relation to Christ, stays calm. Judas steps back just in the moment he is grasping at the bread. These two groups are more agitated than the calmer outer groups. Thus, the dynamical effect of the words of Christ may be compared to a wave with repercussions and vortices (cf. Wildgen 2005: 161).

Figure 3. The two interior groups: Judas-Peter-]ohn and Thomas-Jack-Philipp.

We can notice that the three persons subdivide again into two plus one:

John / Peter + Judas (in the center, obscured by the light coming from the left). Thomas / Philip + Jack (in the center, receiving the full blow of Jesus' utterance).

The exterior groups have also one central figure to which the others turn. If we analyze their postures and gestures, we can further decompose all four groups of aposdes into two plus one (center). The central person neutralizes and then stops the movement initiated by Jesus and thus brings it to rest. This is the natural locus of movements in Aristotle's physics. It is as if the blow of the utterance has dynamically shaped the four groups and their subgroups as a result of the "percussion"; the grouping would in this view be a natural consequence of the underlying dynamics. If a flow of water is obstructed, GEOMETRY AND DYNAMICS IN THE ART OF LEONARDO DA VINCI | 75 specific forms are created by the flow; if three objects stand in the flow, as shown on sheet 13B of Leonardo's Codex Leicester (ibid: 90), they are bound together.

Figure 4. Details of Leonardo's treatment of hydrodynamics; c. 1509—11; cf. Codex Leicester, 2007: 47 (left) and 90 (right).

The gestures have their own language, and Leonardo asks the artist to observe people who are unable to speak in terms of how they express themselves with the help of gestures.15 Thus, the dynamics in the picture concern the effect of Jesus' utterance in the postures of his disciples and in their gestures and mimics. Dynamically, Judas is clearly separated from the other apostles, he seems to be lost for any positive effect, holds his money in the right hand, and stops grasping for a moment at the bread. He shows a closed, sinister face, which is in full contrast to the face of John who is illuminated by the light from the left. In relation to the geometry of the painting we could say that Leonardo tries to organize his composition as an instant in a process which shows the origin of the force and the immediate effects and multiple structures created by the percussions of the force, which are in and come from Jesus. The geometry is subordinated to these dynamics. As the emotional and intellectual effects of the central force are the main topic of the painting, Leonardo reorganizes the geometry of the scene in order to arrive at an optimal representation of the percussion in body-postures, gestures and mimics. The anchoring image schema is a kind of fluid wave, and the pattern shown in the painting is a snap-shot of the process of pattern-formation. The word of Jesus has the effect of a 'Turing instability", which generates morphology by self-organization in the group of apostles; the four triads of persons are the product of the interaction of the expanding instability in the subgroups. The

15 Cf. Pedretti (1995: 219). 76 I W. WILDGEN

whole is then a type of self-organized pattern resulting from the instability created in the center (Jesus).16 Leonardo's construction is a visualized solution under these premises. This explains why Leonardo was perceived as a radical innovator in his time. His solution was partially taken up by Raphael and Dürer in their representations of the "Last Supper", but the dynamics were less radical. They chose a compromise between the lesson given by Leonardo's Last Supper and tradition.

2.3. An anchor of stability: the shape of the table and its position In the medieval and orthodox tradition the front line of the table is often straight, whereas the three other borders are disposed along a semi-circular line. In the group of persons the central person and his antagonist (e.g. Jesus and Judas) are highlighted, the remaining aposdes are treated as a group (cf. Wildgen 2004: Figure 1). In Dürer's woodcut (1510) the table is less broad and the aposdes form a circle with an opening on the front-side. Circular arrange- ments of the aposdes are also characteristic for Dürer's woodcut in the series of his "small passion" (1509-1511; cf. Hamann 1932: 500). In a drawing of 1523 Dürer even returns to the classical (Italian) oblong table and thus departs radically from the younger tradition by placing Jesus (and John) to the left of the table (cf. Panofsky 1977: 295; plates 277f.; and Marani 2001: 334-337). Dirk Bouts' "Last Supper" (1467, town hall, Leuven) shows the aposdes sitting along a quadratic table, with two sitting with their back to the viewer, three on each side of the table, and four at the side of Jesus. Tintoretto has painted at least four different versions of the Last Supper in 1543, 1566, 1579 (or 1581) and in 1592 (or 1594). In this last painting, Judas is again sitting on the opposite side of the table, below the diagonal which separates the aposdes and the persons serving food. As Jesus is giving bread to John, the scene is after the betrayal question and before Judas has left the group.

16 The Aristotelian idea (De cáelo, II, 6) that the wave of water accelerates until it reaches the mid-line of the path towards the limit and decelerates later has been commented on and discussed in medieval rime and during the Renaissance (cf. Frosini 1998a: 138-148). This could explain why the two groups next to Jesus at his left and right show a stronger movement. Leonardo also considered backward waves called "retrosi" and "vortici" (cf. the comparison of Leonardo's sketch of the flow behind a cylinder and modern representations in: Gombrich 1969: 191). The motion of Peter and Philip towards Jesus in the nearer groups could be explained by the vortex-effect of the wave behind an obstacle. In the same period, the theory of waves (in water and air) became de-mechanized (cf. Frosini 1998a: 144) and could be applied in semiotic contexts. GEOMETRY AND DYNAMICS IN THE ART OF LEONARDO DA VINCI | 77

Figure 6. Jacobo Tintoretto (1518—1594) "Last Supper" in San Giorgio Maggiore (Venice), painted between 1592 and 1594.

The shape of the table influences the grouping of persons sitting at the table. Although Leonardo did not give a rule for such groupings, he commented extensively the proportions of body parts in men/women and horses applying the method of fractions. The basic units in human bodies is 1/16 and the different body parts are described by rules containing fractions on the basis of 16: 2/16, 3/16., 5/16 etc. If further levels of distinction are considered, the basis is 162 =256 (cf. Panofsky 1940 and 1955/1970). In the case of the Last supper the basis is 12 and Leonardo considers the fractions 6/12 = Vz; 4/12 = 1/3. The two symmetric groups of apostles and the three triads are the result of these proportions.17 Thus the painting has also a rhythmical order based on arithmetic proportions like musical harmonies. After the analysis of this highlight in Leonardo's work I shall turn to more basic compositions and their dynamics.

17 Peirce considers the triad as the basic semiotic configuration, because more complex configurations may be understood as compositions of triads or "degenerate" triads. Simpler configurations are understood as truncated ("degenerate") triads. He did not give a general proof of this conjecture, but Burch (1991) was able to prove the conjecture in the context of his reconstruction of a "Peircean Algebraic Logic" (PAL) and he showed that Quine's attempt to reduce triads to dyads follows from an extensionalist view of relations, which is not compatible with Peirce's anti-nominalism (cf. Burch 1991: ix and Quine 1953). 78 I W. WILDGEN

3. Figurai composition with three (or four) constituents Christ and the Virgin are central topics of Chrisdan paintings. The "Last Supper" shows Christ in the group of apostles. The Virgin appears very often with Jesus as a child (in different postures and functions). Sometimes both are surrounded symmetrically by two saints or by two angels. These configurations are basically dyadic; i.e., the central pair, the Virgin and Jesus, are the main topic (the figure), and the surrounding persons are part of the background. In the first painting of Leonardo as independent master (ca. 1481/2), the visit of the three Magi, the Magi kneel in front of the Virgin and Jesus in a triangular construction, and they keep a distance which marks their secondary role (in relation to Mary and Jesus). A half-circle of about twenty persons separates them from a background which has no direct relation to the episode.18 The dominant geometrical figure is a broad triangle pointing to the head of the Virgin and including the three Magi. The triangle is dynamically asymmetric, as the right side is the line between the gaze of Mary and Jesus that points to the face of a kneeling Magus. This may be called the central force-line inscribed into the triangle. The head of Mary is at the same time the vertex of a semicircle along which the surrounding figures are arranged. In his further development Leonardo reduces the prominence of linear perspective and concentrates on the features he calls "perspective of color and vanishing perspective". The background becomes idealized, vague, and neutral (cf. his "Mona Lisa").

18 Some authors have counted 66 persons and animals in the painting. Sketches for the painting show that geometrical considerations were basic for its composition. Thus, the linear perspective is defined by the architecture in the background. GEOMETRY AND DYNAMICS IN THE ART OF LEONARDO DA VINCI | 79

Figure 6. Leonardo's Epiphany (Adoration of the Magi) and the underlying geometry and view-tines.

In "The Virgin of the rocks" we find a new constellation with four persons.19 The geometrical scheme is that of a pyramid with the angel on the second (visible) side of it. The front side has on its vertex the face of the Virgin, on its edges the Jesus-baby at the right and St. John Baptist as a baby on the left (above the base-line of Jesus and the angel).

19 In the London version of the painting, which is a modified copy, the pointing gesture of the Angel is eliminated. This and the different gaze direction of the angel simplify the quaternary construction to a triadic one. 80 I W. WILDGEN

Figure 7. 'The Virgin of the Rocks" by Leonardo (Louvre, Paris; 1492-94) and detail of the second version (London before 1508; detail).

This painting shows the complexities Leonardo tried to manage. Jesus is separated from Mary, who holds St. John with her right hand. In the field of Jesus' head we find three different hand-gestures: - Mary protects/grasps Jesus (she has no contact with him and her gesture is interrupted by the gesture of the angel), - Jesus blesses St. John, - the angel, who looks at the spectator/interpreter outside the painting, points to St. John, and - St. John looks at Jesus and kneels in front of him.

The hand gestures alone define five different force-lines, and all four persons have different gaze-lines involving different angles of the head. All these forces imply some narrative context; i.e., Leonardo tries to compress a story in the static configuration of a painting. The basic episodes are: GEOMETRY AND DYNAMICS IN THE ART OF LEONARDO DA VINCI | 81

- The angel Gabriel announces the birth of John (Luc. 1,13)20 - The angel Gabriel announces the birth of Jesus (Luc. 1,35) - John announces the activity of Jesus (Mat. 3,11; Luc 3,16; Joh. 1,26) - Jesus calls John the prophet, who prepares his paths (Mat. 11,10; Luc. 7,27).

The biblical report is represented in its different episodes; nevertheless, the quintet Elizabeth (mother of John) - Mary / John - Jesus / and the angel Gabriel is simplified to a quartet. As the relational link between John and Elizabeth is lost, John is now linked to Mary, who in the biblical story has only a light link to him. At her arrival John moves in the womb of Elizabeth.21 Geometrically, the four persons who interact occupy the edges of the pyramid of persons; the gestures of their hands and their gaze-lines constitute a "pyramid" of forces, as shown in Figure 8. One could say that this painting lacks a proper center of gravity. Its thematic units are distributed in space, and although there is a geometrical schema for this distribution, the dynamics of weights, forces, and gestures remain somewhat vague. In a second version (in London) the intriguing gesture of the angel and its vision-line out of the frame are eliminated. This creates a better balance, but diminishes the vividness.'

Angel

John

Y*t£UT6 8. ι fJC f/jfurr/tu in trjc ν ujejrt uj nJt IUAJU .

20 Some commentators call the angel "Uriel". If we compare the two angels in the present painting and Leonardo's "Annunciation", they have similar clothes and colors. Gabriel is in general associated with Mary and with John. Piccaluga (1998) also calls the angel in Leonar- do's painting "Gabriel". 21 Gabriella Piccaluga (1994) relates Leonardo's painting to the teaching of Amadeo Mendes da Silva in Milan. In Mendes' teaching, the angel introduces the third immaculate being, Jesus, to the first two, Mary and John Baptist. Together the Virgin, John and Jesus form a trinity in the theological interpretation of Mendes. This level of analysis corresponds to Panofsky's third level called "iconologica!" or with Cassirer "symbolic" (cf. Panofsky 1939/1970: 66). 82 I W. WILDGEN

The elaborated cartoon for St. Anne (with Mary and Jesus) was finished in 1498/99, i.e., after the "Last Supper" and a decade after "The Virgin of the rocks". It shows another quaternary relation: St. Anne, Mary, Jesus and St. John. The major narrative and topical contents are: - St. Anne is the mother of Mary, - In the tradidon of this pictorial topic, Mary is sitting on the knees of St. Anne and Jesus is held by Mary. - The presence of John or the lamb is facultative.

In the elaborated cartoon (now in London), Leonardo cuts the upper edge of the pyramid such that the heads of Mary and Anne are on one horizontal line. Nevertheless, the contours of Mary at the left and the gaze-line of Mary — Jesus — St. John form a triangle. Dynamically, we have a central triad given by Mary (holding Jesus), Jesus (blessing St. John) and St. John (receiving the blessing); St. Anne is a bystander. She supports Mary (on her knees), looks at her and points to heaven; her pointing hand defines the upper edge of a smaller pyramid, with the head of Jesus and St. John as the bottom-line. Figure 9 illustrates this analysis.

Hand of Anne

Mary Anne \ John

Figure 9. Geometrical figures andforce-lines iti the London cartoon of St. Anne.

The lines of the gazes and hand gestures fit into a rectangle; its diagonal is the basic force-line, which has Jesus as its central attractor. The weights of the GEOMETRY AND DYNAMICS IN THE ART OF LEONARDO DA VINCI | 83 central groups of adults Anne/Mary are balanced by the body of Jesus, which is a counterpoise to the body of Mary sitting on Anne's knees. The whole composition is, therefore, centered on Jesus (as its "barycenter"). In 1503 and later, Leonardo tried different constructions. In one sketch, the head of St. Anne in proximity to that of Mary is scratched, and a new version where the heads of Mary and St. Anne are at a certain distance, so that St. Anne does not look at Mary, is given (cf. Clayton 1992: 245, fig. 22). The basic geometrical figure is now a triangle, with the base line on top as shown in Figure 10.

Mary

Lamb

Figure 10. Triangular construction in the sketch of 1503.

The final stage (ca. 1510) of St. Anne (now exhibited in the Louvre, Paris) comes back to the asymmetric pyramid with the main force-line at the right, placing St. Anne, Mary, Jesus, and the lamb on one gaze-line. However, the dynamics are new. Although Mary is still sitting on the knees of St. Anne, she is moving towards the child (Jesus). This complicated decentralization clears the space for a full portrait of St. Anne, who now joins the main gaze-line instead of breaking it into two gaze-lines as in the early cartoon. Jesus is also in motion, and while climbing the lamb, which is struggling against it, he is held back by Mary. The dynamical scale of events goes from: - St. Anne (passive, just observing), - the lamb (just reacting, struggling), - Mary (gently withdrawing Jesus from the lamb), to - Jesus (trying to ride on the lamb and moving against the force-line of Mary). 84 I W. WILDGEN

The force-lines of gaze and gesture are reinforced by (real) forces like: - support (Anne, Mary) - hold back (Mary, Jesus) - mount (Jesus, lamb) - separate from (Jesus, Mary)

The dynamics of support serve as a background for the dynamics of holding, separating, and mounting. Jesus is the mediating force between Mary and the lamb.

Figure 11. The final version of St. Anne completed in 1509/10 (Louvre, Paris). GEOMETRY AND DYNAMICS IN THE ART OF LEONARDO DA VINCI | 85

The strong dynamism of this painting and the stricter geometrical construction are obvious, if we compare it with Raphael's treatment of a similar topic such as the "Holy Family" (1507). Although an influence by a lost cartoon of Leonardo is visible (e.g., the pyramidal construction and the gaze-lines on the left edge of it), it is more static (see Joseph's bar in the center and the linear array of the figures without weights and counterpoise).

4. Figurai composition and some levels of interpretation22 In the last sections, I have mainly analyzed what can be seen in the painting in terms of geometrical and arithmetical structures, force fields defined by obvious movement (frozen in the picture), and of gestures and directions of glance. These visible features (I did not consider color, shades, transient zones, the landscape or the floor on which the figures stand) establish a basic signification of the piece of art, independent of the literal or metaphorical meaning of the objects, animals, and persons represented, or of the narrative context of the actions frozen in the picture. In the last section on the paintings with St. Anne I have introduced the names of Anne, Mary (the Virgin), Jesus, John (Baptist), and the lamb, and I have renamed several recognizable actions, such as support, hold, mount, and separate. These referential meanings relate what is visible in the painting to the knowledge of the viewer (ideally to the knowledge of a 16th century viewer at court or in the clergy). This level of interpretation, which links visible features to content, goes beyond the painting; i.e., we have to consider another semiotic stratum, on which persons, animals, their attributes, actions, episodes, and stories are organized. In the case of sacred art, the bible is the narrative background; i.e., we refer to some linguistically organized and retrievable knowledge. As a consequence, the interpretation of the painting at the content level has to operate on three separate mental strata in the mind of the viewer:23

(a) The visual features (including the geometrical and dynamical features discussed in the last sections) and the entities they constitute (b) The relations between these entities (recoverable visually) and of forces, events and actions frozen in the painting (c) The meaning of entities recognizable in the painting

22 Panofsky (1939/1970) distinguishes three strata of meaning in everyday life and in works of irt (cf. Panofsky 1939/1970: 53-67). We prefer to treat such strata as methologically motivated and thus variable. In cognitive and cultural reality a kind of processual hierarchy of complexity may exist, which allows for a number of meaningful subdivisions. 23 Panofsky's second "iconographies!" level does not distinguish the relational, dynamic reatures from the forms recognized. I W. WILDGEN

Visually recoverable relations (b) are:

1. Anne — background to — Mary — background to — Jesus 2. Anne's head — above — Mary's head — above - Jesus' head 3. Anne — looks older than — Mary - looks older than — Jesus 4. Anne - supports - Mary - supports - Jesus 5. Anne and Mary form one group, Jesus and the lamb another group (by proximity).

The interpretation of immediately given data in search for a richer understanding is called "attribution" in psychology, e.g. the attribution of causes and motivation for a perceived event or action. In Figure 11, the three persons Anne, Mary, and Jesus (of different age and sex) are easily identifiable as the major topics. From our knowledge of the bible we know that there is a kinship relation: Anne — mother of — Mary — mother of -Jesus (by transitivity we know that Anne is the grandmother of Jesus, by inversion that Jesus is the son of Mary and Mary the daughter of Anne). Even the colors may help to match the knowledge structure with the visual structure. Traditionally, Anne has a green mande and a red suit. In Leonardo's painting mother and daughter have a green mantle, and Mary wears a red suit. This knowledge level constitutes a space of conventionalized meanings related to the visual space that we establish by looking at the painting (a) and inferring in a kind of imagined mirror-action24 the relevant forces (b). The knowledge structure is, however, much richer: - The lamb is a symbol of the sacrifice of Jesus and an attribute of John (Baptist). - From John, who replaces the lamb in the cartoon, we may remember the specific relation of John to Jesus; John baptized him and called him the Messiah. - Further links exist between John's mother, Elisabeth, and Anne, the mother of Mary; both had been waiting long to become pregnant. Fur- ther, Elisabeth's pregnancy announced Mary's, and Mary stayed with Eli- sabeth during the first months of her own pregnancy.

I will not further elaborate the contents of the attributional space and its consequences for the final interpretation. My concern is rather the background of knowledge which governs the interpretation process (the component called archetypal by Thom, and generic by Petitot (2004) and Bundgaard (2009). The first space (the painting and its visual effect) contains a rich geometric and dynamical structure (weights, barycenters, force-lines, gaze-directions, etc.)

24 Mirror cells in higher primates allow the observer to reproduce in his mind the movement or action he is witnessing. In the case of a static image, he/she may reproduce the process which has led to the static situation. GEOMETRY AND DYNAMICS IN THE ART OF LEONARDO DA VINCI | 87 which is used in many of Leonardo's paintings (and in those of his contempo- raries). A purely static schema would be insufficient; an adequate anchoring of the analysis of verbal and visual signs asks for a dynamic schema. The art of Leonardo shows clearly that we need a concept of dynamic vaknce.2S In the case of St. Anne we have on the surface a quaternary constellation: Anne — Mary - Jesus - lamb. If one considers the force fields and actions, one notes that a basic interaction links mainly Mary - Jesus -the lamb. - Mary draws on Jesus. - Jesus draws on the lamb. - The lamb resists.

There is a conflict between Mary, who tries to prevent Jesus from mounting the lamb, and Jesus, who notices this (he looks back to her) but resists her. This triad constitutes a force field which dominates the message of the painting. A first schematic representation introduces two vector-fields with attractors:

Mary < Jesus > lamb

Jesus is in the metastable position between two attractors; the narrative (biblical) content of these attractors is:

Mary: His mother; she cares for her baby. Jesus: He feels the duty to sacrifice and to leave Mary behind.

The archetypal schema can be derived from catastrophe theory (cf. Wildgen 1982,1994). The semantic archetype of tranrfer, is shown in figure 12.

Figure 12. The dynamical archetype of transfer (giving) and a fiber on it (with attributed contents).

25 Dynamic schemata are non generic in the sense of Petitot (2004); they are integrational in the sense of Oakley (2009). 88 I W. WILDGEN

This archetype does not describe completely the basic interactions in the composition. We have to consider two complications:

a. Anne supports/anchors the whole event (physically and genealogically), and she is a fourth attractor, who does not directly intervene but sustains the event (which is happening on her knees). In catastrophe theory, a second internal variable (and dimension) is introduced, and Thorn (1983: 205) calls the basic archetype with four centers of attraction a "Messen- ger". It works like a background cause, which is a medium (cf. Wildgen 1982: 88-92, 1994: 129-134) which enables the interaction and transfer implying three attractors and their force-fields. b. There is a complication in the manner of "transfer" which corresponds to the difference between the following sentences: i. Mary sits Jesus on the lamb, ii. Mary withdraws Jesus from the lamb, iii. Mary prevents Jesus from riding the lamb, iv. Mary tries to prevent Jesus from riding the lamb.

In the visual structure we just observe the hands of Mary seizing Jesus and the hands (and feet) of Jesus seizing the lamb, and we see that Jesus has a stronger grip on the lamb than Mary has on him. The turning of his head creates an opposition to the force-direction of Mary's hands. The preference scale of the sentences: iv > iii > ii > i is linked in the context of the painting to visual cues, but we need further knowledge which comes from our experience with human interaction. In a phenomenological perspective, the body-centered and enacted schemata are fundamental.

Archetypal schema (Independent of the human sphere)

First level of attribution

Phenomenology of human experience (Typically the human sphere)

Second level of attribution

Narrative interpretation in the context of the biblical story (The religious/cultural sphere)

Figure 13. The archetypal schema and two further levels of attribution. GEOMETRY AND DYNAMICS INTHEARTOF LEONARDO DA VINCI | 89

If one observes a mother interacting with her child, the interpretation at the second level occurs. In cultural traditions that have elaborated a collective visual or linguistic memory (e.g., of major biblical contents), the more artificial third level is superimposed on the quick routines of level two. It is typical for art that cultural knowledge and special "reading" skills are presupposed; just "perceiving" and recognizing is not enough. This became obvious when in abstract art the narrative content and the reference to established canons of content were abolished.

5. Outlook: Are geometrical and dynamic features relevant in the analysis of abstract art? Abstract art gave up the realism implied by Leonardo's theory of art and abandoned narrative contents and recognizable representations. Did geometry and dynamics lose their relevance in this context? It is true that spatial illusion, the major effect of the correct use of linear perspective, either became secondary or was replaced as artists put the accent on the two-dimensional nature of paintings, the relevance of texture and the balance of colored surfaces, characteristically in the work of Cézanne (1839- 1906) and van Gogh (1853—1890). As a consequence the geometry of the plane and the dynamics of strokes and lines came to the foreground. The trend towards abstraction and minimalism enforced the geometrical and dynamic features; in a sense it eliminated many attributional and representational processes dependent on specific cultural presuppositions. In the synthetic cubism of Picasso (1881-1973) new artificial perspectives, and modes of integration of parts to wholes, appeared. The parts, such as of a body or plant, were still representational, but the composition was artificial in relation to everyday experience. When Kandinsky (1866-1944) devised his first abstract paintings, or when Pollock (1912-1956) created his action paintings, geometrical order seemed to disappear, and stochastic (noisy) patterns and irregularity seemed to dominate. Nevertheless, geometry did not disappear. Since the mid 19th century new models of geometry (non-Euclidean and hybrid geometries) and new types of dynamics (non-linear and chaotic dynamics) have been formally described. In relation to these geometries, Picasso's cubistic paintings and Pollock's dripping actions correspond to geometrical and dynamic principles in a similar way to Leonardo's paintings; i.e., the link between mathematics and art is still existent and productive. It is clear that the higher attributional levels were not just cancelled; they were replaced by new and often very complicated attributions, referring to specific cultural experiences in a global and rapidly changing world. Most of modern art criticism tries to analyze these often very unstable and unpredicta- ble processes. This has the consequence that such analyses cannot cope with the standards of modem experimental and mathematically structured sciences. The underpinning of the attributional processes by geometrical and dynamical I W. WILDGEN

principles remains to be the backbone of our understanding of visual art because they are cognitively founded in our sensorial and motor-capacities. This level is accessible for scientific methods and makes a scientific visual semiotics feasible. Via the evolution of these capacities it is even rooted in the physics of the surrounding world, and therefore realistic, but in a less immediate sense than in Renaissance art. As a consequence the cognitive semiotics of art should follow a double strategy: - Find the universal underlying mechanism of visual understanding. - Describe the infinity of attributional processes dependent on cultural traditions and context of usage.

References Baralon, P. B. & Marani, P. C. (2001). Leonardo: The Last Supper. Chicago: University of Chicago Press. Baur, O., Bott, B., Braunfels-Esche, S., Keele, K. D., Ladendorf, H., & Putscher, M. (Eds.) (1984). Leonardo da Vinci. Anatomie, Physiognomik, Proportion und Bewegung. Köln: Kölner Medizinische Beiträge, vol. I, Kohlhauer. Braunfels-Esche, S. (1984). Aspekte der Bewegung. Umrisse von Leonardos Propor- tions· und Bewegungslehre. In O. Baur, B. Bott, S. Braunfels-Esche, K. D. Keele, H. Ladendorf & M. Putscher (Eds.), Leonardo da Vinci. Anatomie, Physiognomik, Proportion und Bewegung (pp. 84—117). Köln: Kölner Medizinische Beiträge, vol. I, Kohlhauer. Bundgaard, P. F. (2002). Presentation and Representation in Art - Ontic and Gestaltic Constraints on Aesthetic Experience. VISIO 7 (1-1): 187-203. Bundgaard, P. F. (2009). The Grammar of Aesthetic Intuition - On Ernst Cassirer's Concept of Symbolic Form in the Visual Arts. Synthese. Available at: , Retrieved 13/09/2010. Clark, K. (1958). Leonardo da Vinci. An Account of this Development as an Artist. Har- mondsworth: Penguin Books. Clayton, M. (1992). Studi per St. Anna. In Leonardo & Venera, Mostra Paiamo Grassi, Catalogue (pp. 242—255). Milano: Bompiani. Clayton, M. (1996). Leonardo da Vinci. One hundred Drawings from the Collection of her Majesty the Queen. The Queen's Gallery, London. Codex Leicester (2007). Leonardo da Vinci. The Codex Leicester (ed. by M. Ryan, P. Cottrell, M. J. Gorman & D. Cross). London: Scala. Edgerton, S. Y. (1980). The Renaissance Artist as Quantifier. In M. A. Hagen (Ed.), The Perception of Pictures (pp. 179—212). New York: Academic Press. Frosini, F. (1998a). Filosofia naturale e sdenta della pittura nel pensiero di Leonardo da Vinci. PhD-thesis, University of Urbino (Italy). Frosini, F. (Ed.) (1998b). "Tutte le opere non son per instancarmi". Raccolta di scritti per i settant' anni di Carlo Pedretti. Rome: Edizioni Associate. Gombrich, E. H. (1969). The Form of Movement in Water and Air. In C. D. O'Malley (Ed.), Leonard's Legacy. An International Symposium (pp. 171-204). Berkeley: University of California Press. GEOMETRY AND DYNAMICS IN THE ART OF LEONARDO DA VINCI |

Hamann, R. (1932). Geschichte der Kunst. Vol. 2: Mittelalter und Neuheit. München: Droemersche Verlagsanstalt. Jahn, J. (1952). Leonardos Aufzeichnungen. In H. Lüdecke (Ed.), Leonardo da Vinci: Der Kunstler und seine Zeit (pp. 49-61). Berlin: Henschel-Verlag. Kemp, M. (2001). "Fate come dico, non fate come faccio!" Lo spazio e lo spettatore nell' Ultima Scena. In P. C. Marani (Ed.), Il Genio e le Passioni, Leonardo e il Cenacolo. Precedenti, innovazioni, riflessi di un capolavoro (pp. 53—59). Mostra Milano, Palazzo reale, 21 march—17 june 2001, Skira, Milano. Luchinat, C. A. (2001). Note sulla psicologica die commensali nei Cenacoli fiorentini prima e dopo Leonardo. In P. C. Marani (Ed.), Il Genio e le Passioni, Leonardo e il Cenacolo. Precedenti, innovazioni, riflessi di un capolavoro φρ. 47—52). Mostra Milano, Palazzo reale, 21 march—17 june 2001, Skira, Milano. Lüdecke, H. (Ed.) (1952). Leonardo da Vinci: Der Künstler und sane Zeit. Berlin: Henschel- Verlag. MacCurdy, E. (1977 [1938]). The Notebooks of Leonardo da Vinci. Voi. 1, 2. London: Jonathan Cape. Marani, P. C. (Ed.) (2001). Il Genio e le Passioni, Leonardo e il Cenacolo. Precedenti, innovazioni, riflessi di un capolavoro. Mostra Milano, Palazzo reale, 21 march-17 june 2001, Skira, Milano. Merleau-Ponty, M. (1989). L'oeil et l'esprit. Paris: Gallimard. Nicodemi, G. (Ed.) (1939). Leonardo da Vina. Gemälde, Zeichnungen, Studien. Zürich: Fretz and Wasmuth. Oakley, T. (2009). From Attention to Meaning. Explorations in Semiotics, Linguistics, and Rhetoric. Bern: Peter Lang. Panofsky, E. (1970 [1939]). Iconography and Iconology. An Introduction to the Study of Renaissance Art. In E. Panofsky, Meaning in the Visual Arts (pp. 51—81). Har- mondsworth: Penguin Books. Panofsky, E. (1940). The Codex Huygens and Leonardo da Vinci's Art Theory. London: The Warburg Institute (Reprint Klaus Reprint, Nendeln, Liechtenstein, 1968). Panofsky, E. (1977). Das Leben und die Kunst Albrecht Dürers, Rogner and Bernhard. Frankfurt/Main: Zweitausendeins. Pedretti, C. (Ed.) (1995). Leonardo da Vinci. Libro £ Pittura. Codice Urbinate lat. 1270 nella Biblioteca Apostolica Vaticana (trascrizione critica di Carlo Vecce). Florence: Giunti. Peirce, C. S. (1986). Semiotische Schriften. Voi 1. Frankfurt/Main: Suhrkamp. Petitot, J. (2004). Morphologje et esthétique. La forme et le sens che\ Goethe, Lessing Lévi-Strauss, Kant, Valéry, Husserl, Eco, Proust, Stendhal Paris: Maisonneuve & Larose. Piccaluga, G. F. (1994). Ζοφια. Achademia Leonardi Vinci, VII: 13-56. Piccaluga, G. F. (1998). Una nuova copia della Vergine delle rocce. In F. Frosini (Ed.), "Tutte le opere non son per instancarmi". Raccolta di scritti per i settant' anni di Carlo Pedretti (pp. 113-127). Rome: Edizioni Associate. Quine, W. V. O. (1953). Reduction to a Dyadic Predicate, In W. V. O. Quine, Selected Logic Papers. Reprint. Cambridge, MA: Harvard University Press. Reti, L. (Ed.) (1974). The Unknown Leonardo. London: Hutchinson. Schumacher, J. (1974). Leonardo da Vinci der Maler-Philosoph. Frankfurt/Main: Makol. Strong, D. S. (1979). Leonardo on the Eye. An English Translation and Critical Commen- tary of Ms. D in the Bibliotèque Nationale, Paris with Studies of Leonardo's Me- thodology and Theories of Optics. New York: Garland. Thom, R (1983). Mathematical Models of Morphogenesis. Chichester: Horwood. I W. WILDGEN

Verstegen, I. (2005). Amheim, Gestalt and Art. A Psychological Theory. Wien-New York: Springer. Wildgen, W. (1982). Catastrophe Theoretic Semantics. An Elaboration and Application of René Thorn's Theory. Amsterdam: Benjamins. Wildgen, W. (1994). Process, Image, and Meaning. A Realistic Model of the Meaning of Sentences and Narrative Texts. Amsterdam: Benjamins. Wildgen, W. (2004). Éléments narratifs et argumentatifs de T'Ultime Cène" dans la tradition picturale du Xlle au XXe siècle. In S. (Ed.). Espaces perçus, territoires imagés en art (pp. 77-97). Paris: L'Harmattan. Wildgen, W. (2005). Visuelle Semioük der elementaren Kräftefelder der Hände (Gestik) und der Augen (Blicke) in einigen Werken von Leonardo da Vinci und Barocci. In W. Nöth & A. Herding (Eds.), Körper- Verkörperung- Entkörperung (pp. 149-179). Series: Intervall, vol. 9. Kassel: Kassel University Press. Wildgen, W. (2006). The Dimensionality of Text and Picture and the Cross-cultural Or- ganization of Semiotic Complexes. In R. Köhler and A. Mehler (Eds.), Aspects of Automatic Text Analysis (pp. 421—442). Festschrift in Honor of Prof. Dr. B. Rieger. Berlin: Springer. Wildgen, W. (2008a). Kognitive Grammatik. Klassische Paradigmen und neue Perspektiven. Berlin: Mouton de Gruyter. Wildgen, W. (2008b). The 'dynamic turn' in cognitive linguistics. In H. Tissari (Ed.), Studies in Variation, Contacts and Change in English. Electronic series. Available at· , Retrieved 13/09/2010. Wölfflin, H. (1904). Die kJassische Kunst. Eine Einführung in die italienische Renaissance. Munich: Bruckmann (3rd edition). Ivan Darrault-Harris

Non-Genericity as an Invariant of the Readability of Pictures1

Starting from the semiotic analysis of a diptych by Cranach the Elder picturing Lucrece committing suicide and Judith after the murder of Ho- lophern (Gemäldegalerie Alte Meister of Dresden, Germany), we examine the plastic layouts that make those pictures readable. It is a well-known fact that those two paintings are a pictorial conversion of legendary texts (respectively the book of Judith in the Old testa- ment and Livy's mythical history of Rome). Particular interest is given to what interpretation of the texts is thus made visible - which provides a first level of readability. But, more precisely, our aim is to highlight the adequacy and fruitful- ness of the concept of "non-genericity" (put forward by Jean Petitot in his Morphologe et esthétique, 2004), as it may prove a valuable invariant of readability not only for 16th century works, but also in general, while being efficient even today to the extent that it "guides" the contemporary spectator's interpretive perception. Indeed, a phenomenon as non-genericity, such as it is exploited by an artist in his work, can be seen as the point of origin of a whole new process that generates the signification of the artwork.

CORRESPONDENCE: Ivan Darrault-Harris. School for Advanced Studies in Social Sciences (EHESS), Paris, France. EMAIL [email protected]

The concept of non-genericity We shall use the semiotic study of a well-known diptych by Cranach the Elder (Lucretia and Judith, Gemäldegalerie Alte Meister of Dresden), to show the full significance of the probability concept of non-genericity, originally introduced by Jean-Pierre Morel and his colleagues, in the processing of pictures. The key point in Morel and his colleagues' approach (Desolneux et al. 2000) is that shapes, spatial correlations can take on a non-conceptual signification on their own, i.e., independendy of any knowledge of the shape itself. Such a meaningful purely spatial event is "an event that, according to probalistic estimates should not happen in the image and there therefore is significant" (Desolneux

1 Full color reproductions of all images and paintings can be found online at ht^//www.cogniavesemiotics.com (click on "issues overview" in the sidebar and scroll down to #5 to access image files for individual contributions).

Cognitive Semiotics, Issue 5 (Fall 2009), pp. 93-102 I I. DARRAULT-HARRIS

et al. 2000: 8). In other words, any sufficiently important deviation from a generic situation generates perceptual saliency (Petitot 2004: 82). Within the frame of morphodynamic structuralism, which he champions and illustrates, Petitot introduces the notion of "intrinsic signification" defined in purely morphological terms: "The morphological and semiotic - gestaltist - phenomenon of the emergence of intrinsic significations is in fact manifest at the very first levels of perception and it is then possible to elaborate a precise abstract model" (Petitot 2004: 80). But one realizes immediately that this notion of intrinsic signification implies that the signification is not already provided: "[...] the task is not to conceptuaba the interpretation of a given form, but to bring out in a non-conceptual way what is significant in it" (Petitot 2004: 82) ,2 As mentioned above, preconceptual signification is determined in terms of a deviation from a statistical standard case. Thus, a non-predictable arrangement of elements in the picture makes it all the more intrinsically meaningful. Some unlikely alignments will be maximally perceptible. The question, in this case, is not to find an interpretative model of the image, but whether it is possible to extract from it morphological features that are intrinsically significant. Hence the efficiency of a "probability version of the opposition between generic and non-generic" (Petitot 2004: 83).3 The reader will be spared the difficult mathematical formulas in terms of which non-genericity is defined. However, a simple example may illustrate what is meant by "intrinsic signification" and enhanced "saliency" in the domain of vision. In figure 1, it is obvious that the statistically highly improbable alignment of segments induces an immediate perceptive pop-up effect. In other words, it is the very improbability of a given morphological constellation or configuration which constitutes its intrinsic signification.

• "·Λ· fflsr^v j 'W^qm ¡rj — . WMA - rwk ss ilSsSS I. Mlir4'11 9 .9.1 {Ηm & Ί ΓϊτΐΙΒ»; γ '9 ! "νν:- mλ έΑ m¡¡Sg * 1 -t. Τ f>~ y, WSBÊkwν · «

Figure 1. Petitot, 2004:56.

2 In the original text: "[...] il ne s'agit pas d'interpréter conceptueHement une forme donnée mais de dégager de façon non conceptuelle ce qui s'y trouve de significatif '. 3 In the original text: "[...] version probabiliste de l'opposition entre génétique et non générique [...]". NON-GEN ERICITY AS AN INVARIANT OF THE READABILITY OF PICTURES |

Finally, it is suggested that those non-generic morphological features can be the basic level of a new generative process leading, through articulation, to an upper level of meaning (conceptual, interpretative, mythical, etc.). In the conclusion of his semiotic re-reading of Goethe's analysis of the Laocoon, Petitot shows the existence of a new generative process of meaning which he calls "montée morpho-sémiotique" (Petitot 2004: 65) (morpho-semiotic "raising"). The concept of non-genericity and its importance for perception is well- known (cf. for example Hoffman 1998, Ramachandran & Hirstein 1999). Its import on aesthetic meaning construction has, in contrast, only rather recently been foregrounded. We owe these developments to Jean Petitot and his semiotic analyses of Piero della Francesca, Raphael Mantegna, and Poussin (Petitot 2004, 2010 and more particularly of sculptures, for instance the famous Laocoon, whose analysis by Goethe, according to Petitot's demonstration, is the first structural analysis in history; cf. Petitot 2001). Bundgaard (2009, 2010) has also demonstrated the semiotic function of non-generic spatial configurations in art. The analysis below follows the principles in Petitot's claim to the effect that non-generic structure is exploited as a means of meaning making in art; the key idea being that when an artist introduces non-generic (critical, statistically improbable) relations in an image he thereby selects certain spatial structures or relations and qualifies them as semanticalty relevant: in other words, he exploits what is inherently significant for the human eye in order to semiotize space in a specific way.

The diptych This diptych (painted around 1525 by Cranach the Elder (1472-1553)) represents, in two matching parts, Lucretia committing suicide and Judith carrying Holophern's severed head. Cranach is known to have painted many representations of these legendary characters, not counting those which, most probably, remain forever unknown to us (cf. Schade 2003). We reproduce here the diptych as seen by Michel Leiris:4

4 The French writer Michel I^eiris was fascinated by the diptych when he first saw it. Indeed, it sparked off his début in writing with the production of his autobiographical masterpiece L'Age d'homme (The Age of man). We thank the Pr. Herman Parret (University of Leuwen) for his help to find again the genuine reproduction of the diptych. 96 I I. DARRAULT-HARRIS NON-GENERICfTY AS AN INVARIANT OFTHE READABILITY OF PICTURES |

Figure 1. Cranach the elder, 1540, Lucretia and ]udith.

Let us start with the case of Lucretia. The story of her rape and suicide is told by Livy (History of Rome, The Rape of Lucretia, Book I, 57-60, translation by B. I I. DARRAULT-HARRIS

O. Foster. London: William Heineman, 1919). We give here a brief summary of the text with some important quotations:

During the war between Rome and the Rutuli, after an attempt to capture the town of Ardea, some young princes are drinking and speaking about their wives. Collatinus, the husband of Lucretia, proposed to ride a very long distance until Rome to verify the excellence of the behavior of their wives. Lucretia is discovered, though late in the night, "engaged upon her wool". The other women attend a luxurious banquet. "The prize of this contest in womanly virtues fell to Lucretia". As Collatinus and the Tar- quinius approached, they were graciously received, and the victorious husband courteously invited the young princes to his table.

Collatinus and the princes return to the camp but "Sextus Tarquinius was seized with a wicked desire to debauch Lucretia by force".

A few days after, Sextus Tarquinius returns secredy to CoUatia, to Lu- cretia's house, is kindly welcomed. After the dinner, he came to the sleeping Lucretia. Holding the woman down with his left hand on her breast, he said, 'Έε still, Lucretia! I am Sextus Tarquinius. My sword is in my hand. Utter a sound, and you die!" When he found her obdurate and not to be moved even by fear of death, he went farther and threatened her with disgrace, saying that when she was dead he would kill his slave and lay him naked by her side, that she might be said to have been put to death in adultery with a man of base condition. Lucretia is so obliged to give up.

Lucretia asks her father and the husband to come quickly, "for a frightful thing had happened". So speaks her: "The print of a strange man, Col- latinus, is in your bed. Yet my body only has been violated; my heart is guildess, as death shall be my witness. But pledge your right hands and your words that the adulterer shall not go unpunished. Sextus Tarquinius is he that last night returned hostility for hospitality, and brought ruin on me, and on himself no less - if you are men - when he worked his pleasure with me".

Although ever/body seeks to comfort her, "it is for you to determine," she answers, "what is due to him, for my own part, though I acquit myself of the sin, I do not absolve myself from punishment; nor in time to come shall ever unchaste woman live through the example of Lucretia".

Taking a knife which she had concealed beneath her dress, she plunged it into her heart, and sinking forward upon the wound, died as she fell. The wail for the dead was raised by her husband and her father.

If we look for non-generic morphological features in Cranach's representation, we find indeed a most unlikely disposition of Lucretia's body: NON-GENERICITY AS AN INVARIANT OF THE READABILITY OF PICTURES | 99

- The whole lower part of the body from the waist down, including the left arm, is shown facing the spectator: the left hand holds the veil which hides, indeed imperfecdy, the pubis; - The face is not shown firontally; it is tilted to the right of the picture and bent backwards: the eyes look upward; - The torso, together with the right arm in the act of stabbing, shows a pronounced twist towards the left of the picture, as indicated by the right breast seen in profile (blue line).

The spectator is confronted with a strange, incoherent, three-party body and this strange unlikelihood is a powerful fact of non-genericity. This fact, simply perceived, is going to appear later as a favored place of semiotic investment. Moreover, the central axis of the face, if continued downward, meets the exact point where the dagger enters the chest (green line); the blood trickling from the wound, if continued, meets the pubis (red line): so appears a link between the rape as act of bloody penetration and the act of stabbing oneself, repetition of and redress for the rape. It can also be noticed that the axis of the dagger meets the axis of the face at an almost right angle (black line): that right angle is a fact of non-genericity, place of an important meaning, the non- assumption of the mortal gesture. These morphological features constitute so many immediately perceptible intrinsic significations because of their saliency. Crucially, they can be related to higher order conceptual significations. The case of Lucretia (much debated at the time) presented the Renaissance theologians with a considerable problem: on the one hand, suicide was a sin of extreme gravity (forbidding religious burial), while, on the other hand, Lucretia represented the humanistic value of individual freedom, much valued at the time. It should be remembered that Cranach was on friendly terms with Luther, of whom he painted many portraits. What the artist proposes is more or less a formal, pictorial resolution of this powerful contradiction, based on the manifestation of non-generic features that summon up our perception of the artwork. Lucretia is not one, but threefold: the frontal view is that of the modest Lucretia, offended by Tarquinius, the aggressor; the bent face is that of Lucretia submitting to God, towards whom she humbly looks; only the strongly twisted torso, a minor part of the body, assumes the responsibility of the suicidal act: the angle of the dagger with the axis of the face and the direction of the eyes, clearly tells the non-responsibility of the act. The representation of Lucretia stabbing herself carries on mobilizing the contemporary spectator. Cranach has used that powerful means of composition, which is non-genericity, as a permanent mean supported with the natural tropism of human eye for unlikely formal facts. We have tried to develop the interpretation of the legend that the painter wanted to illustrate, updating the story of Lucretia in his time where tensions 100 I I. DARRAULT-HARRIS

between religion and humanist values are strong. The invented solution is remarkable: the replication of Lucreda's body and the division of responsibili- ties. But the non-generic organization of the painting can also generate other lectures by the contemporary spectator. It is enough to modify the values interplaying. The painting remains, nevertheless, a morphologic device for the resolution of contradictions.

Judith and Holophern5 Below is the extract of the book of Judith with the story of Holophern's murder (Chapter 13,1-10):

1 When the evening was come, his servants made haste to depart, and Bagoas shut his tent without, and dismissed the waiters from the presence of his lord; and they went to their beds: for they were all weary, because the feast had been long. 2 And Judith was left along in the tent, and Holofernes lying along upon his bed: for he was filled with wine. 3 Now Judith had commanded her maid to stand without her bedchamber, and to wait for her coming forth, as she did daily: for she said she would go forth to her prayers, and she spake to Bagoas according to the same purpose. 4 So all went forth and none was left in the bedchamber, neither little nor great. Then Judith, standing by his bed, said in her heart, O Lord God of all power, look at this present upon the works of mine hands for the exaltation of Jerusalem. 5 For now is the time to help thine inheritance, and to execute thine enterprizes to the destruction of the enemies which are risen against us. 6 Then she came to the pillar of the bed, which was at Holofernes' head, and took down his fauchion from thence, 7 And approached to his bed, and took hold of the hair of his head, and said, Strengthen me, O Lord God of Israel, this day. 8 And she smote twice upon his neck with all her might, and she took away his head from him. 9 And tumbled his body down from the bed, and pulled down the canopy from the pillars; and anon after she went forth, and gave

5 The book of Judith is a deuterocanonical book, which means it can only be found in the Greek Bible. It is an edifying narrative that tells the adventures of a courageous woman by the name of Judith. The historical frame of the adventure is totallv fictitious. In the introduction itself, Nabuchodonosor is said to be the king of Assyrians and the king of Ninive, which is a double mistake: he was king of the Babylonians and Ninive was destroyed before he came to power. This is as strange as if we opened a book that read: "This story took place when Napoleon, king of Russia, lived in London..." It is very likely that the author was well aware of the historical facts and thus made these mistakes intentionally. His purpose is to get his readers interested, not in historic truth, but in the moral value of the story. The king of "Ninive" sends his general Holophern to wage war against Israel. Holophern, who is shown as a bloodthirsty tyrant, eventually lays siege to the little town of Bethulia. The citizens are desperate but Judith draws up a plan: she decides to go the enemies' camp by night and seduce Holophern. Once the general is asleep in a state of drunkenness, she murders him and brings his head back to Bethulia. Seeing that their chief has died shamefully, the enemies give up the siege and go home. The purpose of the story is to enhance Jewish resistance, and the heroine's name is quite symbolical on that respect since Judith means "the Jewess". NON-GENERICÍTY AS AN INVARIANT OF THE READABILITY OF PICTURES |

Holofernes' head to her maid; 10 And she put it in her bag of meat: so they twain went together according to their custom unto prayer and when they passed the camp, they compassed the valley, and went up the mountain of Bethulia, and came to the gates thereof.

A similar examination for non-conceptual significations brings out the following features:

- The backward tilt - quite prominent - of Holophern's head is in obvious contradiction with the laws of gravity, well represented by the vertical axis of Judith's body. It is also noticeable that the axis of the severed head is strictly parallel to the axis of the sword, whose point is next to Judith's right foot (yellow lines). We show with the yellow dotted line the axis of the gravity which should be respected by the representation of Holophern's head. The unbroken yellow line is the axis of the head es- caping the rules of gravity and parallel to the axis of the spade. - Although the axis of the face meets Judith's right hand, the eyes are dreamy, unfocused. - The left hand holds the head by the hair, an exact illustration of the text: "she took hold of the hair of his head". - The fingers of the right hand - in complete contrast with the left - have a definitely non-generic arrangement: the sword is not seized (as the hair is), but delicately held by fingers in diverging position (whereas the fingers of the left hand converge in their contracted grip).

How can those non-generic features be linked to higher-order significations? Contrary to what is assumed by the representation of Lucretia - whose action, as mentioned, raises a difficult issue, generating a strong contradiction - Cranach here, in our opinion, seems to act both as an illustrator and an exegete of the biblical text: it is not just Judith chopping off Holophern's head, it is Judith receiving, in the instant of the act, the strength that she has prayed God to give her. It is then possible to articulate and link the elements of non-genericity with the interpretation, looking further back in the process that generates the meaning of the artwork. If the sword and the severed head are clearly associated in space — perceptibly — in what can be interpreted as a cause-effect relation, the arrangement of the fingers of the right hand, the hand that has severed Holophern's head, clearly indicates that Judith, now devoid of the divine strength, can no longer grab hold of the weapon as an executioner, but only touch it with the tip of her fingers. Thus she regains her status as a woman, albeit inspired by divine vengeance for a short while. The non generic position of the fingers (Lucretia's right hand) constitutes the morphologic place of an important meaning: she has not assumed the horrible act (for a woman) of beheading: God has guided her arm. Here Cranach follows exactly the text. I I. DARRAULT-HARRIS

By contrast, as is indicated by the position of the fingers of the left hand, she fully assumes the possession of the trophy which will cause the terrified enemies to leave the siege of Bethulia immediately. We reali2e that the proximity of Lucretia and Judith is far from being for- tuitous. Both women have committed an act that is incompatible either with religion or with their woman status. The problem clearly lies in their responsibility, in the extent of their involvement in those acts. The task of solving this problem is set by Cranach the Elder first of all at the first level of the pictorial meaning, where the non-conceptual, intrinsic significations can appear. The question of picture variability which is at the center of this issue may find here an original approach, taking into account the different levels at which such variability may take place in the process that generates the meaning of the picture. What we propose amounts to this: not only the perceptive organs are universal; so is the phenomenon of non-genericity which generates perceptive prominences and leads the spectator - any spectator - to immediately select formal dispositions that make sense in a non-conceptual way. Seen from this angle, variability is to be found further up, at the higher levels of the generative process, where spiritual and mythical meanings are gready influenced by culture.

References Bundgaard, P. (2009). The grammar of aesthetic intuition: on Ernst Cassirer's concept of symbolic form in the visual arts [electronic article]. Synthese. , Retrieved 13/09/2010. Bundgaard, P. (2010). Toward a Cognitive Semiotics of the Visual Art-work - Elements of a grammar of intuition. Cognitive Semiotics #5 (present issue). Desolneux, Α., Moisan, L., & Morel, J-M. (2000). Meaningful Alignments. International Journal of Computervision 40 (1), 7—23. Hoffman, D. D. (1998). Visual Intelligence. New York: Norton. Petitot, J. (2001). L'espace et sa relation à l'esthétique: l'exemple du Laocoon de Goethe. Visio 6 (2-3), 263-273. Petitot, J. (2004). Morphologe et Esthétique. Paris: Maisonneuve et Larose. Petitot, J. (2010). Non-generic viewpoints as a method of composition in Renaissance paintings. Cognitive Semiotics #5 (present issue). Ramachandran, V. S. & Hirstein, W. (1999). The Science of Art. Journal of Consciousness Studies 6 (6-7), 15-51. Schade, W. (2003). Lucas Cranach: Glaube, Mythologie und Moderne. Berlin: Hatje Cantz Verlag. Zoï Kapoula, Qing Yang, Marine Vernet & Maria-Pia Bucci

Eye Movements and Pictorial Space Perception: Studies of paintings from Francis Bacon and Piero della Francesca1

We propose a conceptual model and empirical studies showing that eye movement patterns are an essential part of visual coding and retrieval from memory of artwork images; also they reflect richness of perception of pictorial space composition including depth and movement. The first study examines eye movement patterns while viewing a painting by F. Bacon; areas important for spatial composition were more firequendy fixated by observers with art training than by observers without art training. The second study concerns the painting, l'Annunciamone, by Piero della Francesca, explored by subjects without art training. Piero's pictorial 3D space and perspective captured immediately the observer's eyes, imposing an inquisitive exploration, with large size eye movements and unstable fixations both in direction and in depth. Fixation instability in depth (i.e. changes of vergence angle of optical axes) could be induced in a top-down manner by the vivid sense of depth experienced by the subjects.

CORRESPONDENCE: Zoï Kapoula. Laboratoire IRIS, National Center for Scientific Re- search (CNRS) París, France. EMAIL· [email protected]

Introduction Eye movement exploration - conceptual models Study of eye movements is a handy tool to understand perception of artwork, the aesthetic experience and cognitive interaction. Buswell (1935) and Yarbus (1967) conducted pioneering studies in this field in the USA and the Soviet Union, respectively. Yarbus reported with perspicacity that the eyes fixate mostly the meaningful points regardless of the density of details, that the centre of the image attracts the eyes more, that with additional exploration time the eyes repeat fixations between the same meaningful points rather than exploring new ones. Yarbus was the first to demonstrate that the eye movement pattern depends on the observer's objectives.

1 Full color reproductions of all images and paintings can be found online at http://www.cognitivesemiotics.com (click on "issues overview" in the sidebar and scroll down to #5 to access image files for individual contributions).

Cognitive Semiotics, Issue 5 (Fall 2009), pp. 103-121 I KAPOULA, YANG, VERNET& BUCCI

The issue of eye movement patterns over time is also central to recent models of visual exploration of paintings. Locher (1996, Locher et al. 2007) conducted eye movement studies in relation to perception of pictorial balance; they proposed a conceptual model according to which there is a pre-attentdve stage of visual processing of the painting that results in a representation of its global structural organization, including balance and symmetry. Such global visual treatment is extremely fast, can occur within 100 ms, i.e. before the eyes move. Subsequendy, on the basis of this global vision, eye movements are made to fixate with the fovea selected areas of painting for fine visual analysis. Cognitive factors, such as prior knowledge, may also influence where the eyes will move for foveal fixations. Eye movements and fixations of selected areas allow to extract local features and detailed information about the content and structural composition of the painting. Although we agree with such a model, we believe that eye movements cannot be reduced to a motorizing device of vision. Many ocular motor physiology studies provide evidence for the existence of internal signals, efferent (corollary discharge), and afferent proprioceptive signals, controlling planning and execution of every single eye movement (see Leigh & Zee 2006). We propose that such signals available in the central nervous system form the basis of active vision of paintings (see Fig. 1). This idea is in line with the scanpath theory proposed earlier, according to which a particular image causes a particular eye movement sequence and this sequence is important to code visual information, store it in memory or retrieve it from memory (Brandt & Stark 1997; Noton & Stark 1971). Although controversial in its absolute form (as images can be perceived, analyzed, and memorized without eye movements), this theory can still be valid, as we do naturally move our eyes to view paintings, visual scenes and objects. EYE MOVEMENTS AND PICTORIAL SPACE PERCEPTION |

Cognition Art knowledge, task

Internal motor EYE signals MOVEMENT SYSTEM

Selective Fixation for details of Image coding content and Storing & retrieving syntax Memory

Figure 1. Conceptual model: Holistic pre-attentive visual analysis of the painting takes place immediately; a few milliseconds later (usually 150 to 300 ms), the eye movement system is activated to bring eyes to selected regions of particular interest. Holistic visual processing and eye movement control are influenced in a top-down manner by cognition, art knowledge, specific objectives and tasks aimed by the observer. Execution of eye movements produces internal signals, i.e. efference copy andproprioceptive signals from extraocular muscles. They provide a motor basis for coding and storing the image of the painting in visual memory. Eye movements are thus an action by which the observer 'enters', 'inhabits' the painting. I KAPOULA, YANG, VERNET& BUCCI

ψ' cil fy** W % I 4. < -c" 'Τ

tó *

»: ilwuiH « -um· W.m

-qg- # ' f\

Figure 2. The artwork is a sculpturedfragment of asphalt (A). The artist explored the image of the asphalt fragment on a PC screen while his eye movements were recorded with videooculargraphy. (B) the artist viewed on the PC screen the asphalt fragment for 30 s. (C) An empty square frame was EVE MOVEMENTS AND PICTORIAL SPACE PERCEPTION | shown on the screen for 30s; the artist explored by memory the imagined asphalt fragment. (D) the artist saw again the asphaltfragment for 30s. (E) the empty frame was shown on the screen; the artist this time was instructed to draw with his eyes the imagined fragment of asphalt. Conditions Β - C, and D - E wen separated by 20s, a short-term memory interval. For all conditions, column (1) shows the eye scanpath during the totalperiod of 30s; (2) the eye scanpath during the first 5s; (3) the eye scanpath during the next 5s. Note that the scanpath is highly reprodudble for the two visually-driven conditions (B and D); in the memory-driven conditions (C and E), the scanpath is also rimilar at least for the first 5 to 10 s. Reproduäbiäty of eye movement trajectory is consistent with the conceptual modelproposed in Fig. 1.

In order to test the idea of importance of eye movement pattern for coding and retrieving a visual image from memory, in the following experiment we examined how reproducible the eye movement pattern is when the observer views or explores by memory an artwork. We have recorded the eye movement pattern of a contemporary French artist (Michel Paysant) while viewing an image of an object he created (sculptured piece of asphalt). His eyes explored the image of the asphalt piece in a particular organized sequence, drawing its contour, fixating repetitively contrasted areas popping out in 3D; the eye movement was reproducible, and when the artist was asked to imagine the asphalt piece placed in front of an empty PC screen, his eyes were still reproducing the same pattern (see Fig. 2). Similar results were obtained from the members of his family (see Paysant 2005), and also from five laboratory students completely unfamiliar with this asphalt piece. For all such naïve subjects, the eye movement pattern was reproducible over time and when asked to explore the image by memory, the pattern was still similar, at least its first part, 5 to 15 seconds (unpublished observations, master dissertation of A. Castex 2006). Examples from a naïve subject are shown in Fig. 3. Relative to the artist data shown above, the evolution of the pattern is slower for the naive subjects, particularly in the memory conditions (C, E); still one can identify common sequences. Such reproducibility even for naïve subjects support the idea that the eye movement pattern is essential for coding and retrieving the image of the artwork from memory. The next two studies examine to what extent the eye movements can also give information on perception of pictorial space composition and of pictorial movement. I KAPOULA, YANG, VERNET& BUCCI

h (" K' i Ρ i 4t; k * m/ 1 Τt ψ -ν- Λ i.

jT Λ ÏÏ Λñ ; « * (C f J\ V* « ..»s V 4v U « J : ν U 1

i V 4 4

Figure Ì. Evolution of eye movement patterns of a student (20yrs) viewing the asphalt fragment on the PC for the first time; note that the image is now a view of the back of the asphalt piece. The student performed the same sequence of conditions described for the artist; all notations as for Fig. 2. The evolution of the eye movement pattern, particularly in the memory conditions (C and E) is slower than that for the artist; the pattern is shown for 20s, four successive time windows of 5s each (2,3,4,5); the red dotted lines connect the last and first fixations over successive time windows. EYE MOVEMENTS AND PICTORIAL SPACE PERCEPTION |

Pictorial space and movement perception from original paintings by Francis Bacon It is common belief that art experts can perceive more in a painting than non- experts; however, such belief lacks experimental evidence. The goal of this experiment was to investigate the pictorial space and movement perception for observers with various degrees of knowledge of contemporary art. Our hypothesis was that art trained observers would be more sensitive to pictorial space construction and perhaps to movement; second, that such sensitivity would be reflected in their eye movement pattern. Indeed, Nordine et al. (1993) and Zangemeister et al. (1995) reported larger eye movement excursions for art trained subjects when exploring reproductions of paintings, and this is compatible with the idea of difference in pictorial space perception. The study of such questions with original paintings is of particular interest as they provide optimal visual input and optimal aesthetic experience to the observer. We share here some current ideas (e.g. Zeki 1999; Livingstone 2002) seeing artwork as a result of research, the visual artists being, in a sense, neurobiologists of vision challenging the potential and capacity of the visual brain with techniques that are unique to them. In 2004, we had the opportunity to conduct an eye movement study on some paintings by Francis Bacon exposed at the Maillol Museum in Paris, during the exhibition Le sacré et le profane. The methodology combines a questionnaire on perception with observed eye movement patterns and aims to establish the relationship between the two. The motivation for conducting this study on Bacon's artwork is related to the power with which the painter depicts both space and motion. Deleuze (2002), a philosopher who analyzed the work of F. Bacon in depth, qualifies Bacon as an architect, a play director who stages a play on the canvas, by isolating the figure from the background in many different ways; movement is another important dimension in which Bacon's artworks excel: it is generally achieved via representation of body distortion (see also Betancourt 2002). According to Deleuze, Bacon's painting is driven by the logic of sensation, believed to be irrational, primitive and distinct from perception. Thus from a scientific point of view, Bacon's painting provides an excellent ready-made experimental configuration to investigate physiologic correlates of art reception. In our study participated a group of observers with art training (painters, photographers, film directors, video artists, art professors) and a group with no art training (students & laboratory assistants). The paintings selected for the study were not known by the observers. Observers were standing at 2.7 m from the painting that was initially covered with a white curtain. Upon removal of the curtain, they spontaneously explored the painting for a period of 1 min, with no particular instruction. Subjects were not informed that they would be asked questions after the eye movement recording; such precaution was taken in order to capture the properties of their spontaneous eye movement exploration. Their eye movements were recorded with video-oculography (Chronos, Skalar). Here we will report results from the painting Study of a dog, Francis Bacon (1952, I KAPOULA, YANG, VERNET& BUCCI

Oil on canvas, 198 χ 137 cm, Tate Gallery, London, Britain). This painting depicts a plane in depth with a dog in a circle, cars, trees and a horizon line; another conflicting horizontal line induces a vertical plane like an interior wall, shutting, closing the plane in depth (see Fig. 4). Such spatial, configurational structure is used by many Still Life paintings in the 17th century (representing a table with objects against a wall, see Fig. 4). Following eye movement recording, observers were asked whether they perceived 2 planes. Responses were mitigated (70% for non-experts and 50% for art experts reported perceiving two planes). Yet, to a subsequent question does the panting evoke a sense of closure or disclosure the large majority of art experts (85%) responded closure (versus 50% among the non art experts). Was this closure effect related to pre-conscious, pre-attentional perception of the vertical conflicting plane? Whatever the reason, the subjective evaluation of space was different for these two groups of observers. Another difference concerned subjective responses on perception of movement. When questioned on the sensation evoked by the dog at the circle (movement, serenity, contemplation), 70 % of the art experts assigned movement to the dog in the circle which was described as continuously turning while only 20% of the non-experts in art reported a feeling of movement. To what extent such reports correspond to a real sensation of movement, or to higher order prior knowledge, is not known either. Indeed, the painting contains many movement cues (the dog is depicted with five feet, its back is doubled, blurred and shadowed; its face both frontal and in profile). All these genius cues were known better by observers with art knowledge. A complementary possibility would be a specific subtle difference in eye movements and fixations reinforcing such high level motion perception. EYE MOVEMENTS AND PICTORIAL SPACE PERCEPTION |

Figure 4. Francis Bacon, Study of a dog (1952, Oil on canvas, 198 χ 137 cm, Tate Gallery, London, Britain). © Francis Bacon / billedkunst.dk. A plane in depth contains an octagonal carpet, a árele, the dog, can, trees, and a horizon line. A second tine parallel to the horizon tine induces a perpendicular plane closing the plane in depth.

Indeed, the analysis of eye movements revealed two very different patterns of exploration: some observers (mostly those without art training) explored mainly the narrative elements, fixating repetitively the dog and the cars; while others (most of the art experts) produced a global scanning of the painting, also fixating points important to the spatial composition (the upper part, the lines, the edges and corners of the carpet, see Fig. 5A). The examples shown in Fig. 5 are typical. The surface of the painting explored by the eyes was larger for the art expert. In the questionnaire, the non-expert subject shown in Fig. 5A reported perceiving two planes; yet, her eye movements did not show interest in the construction points, namely the vertical upper plane. The art expert shown in Fig. 5B reported perceiving a single plane but reported closure effect I KAPOULA, YANG, VERNET& BUCCI

suggesting two planes, perhaps due to pre-attentional perception of the second vertical plane. His eye movement pattern seems to be more in agreement with such pre-attentative analysis, as he fixates the upper part and the critical second horizontal line. The art expert also reported a vivid sense of dog movement, contrasting the response from the non-expert. Perhaps the larger size of his eye movements and the larger variety of movement direction he performed (Fig. 5B versus 5A) contribute to such movement perception. Indeed, the movements of the eyes and perhaps the more unstable fixations may contribute to movement perception. Zanker et al. (2003) working with Op Art and movement illusion consider eye movements as a possible source of movement illusion. Yet, further fine analysis of degree of fixation instability is needed to substantiate this isue in our study. Finally one should note that points that were key to the spatial composition were not always salient (e.g. the upper part of the painting is featureless); yet, such points were fixated almost immediately by the art-trained observers (see Figures 5C and 5D). Patterns of eye movement exploration were different from that of non-experts from the first few seconds: the pattern in Fig. 5D is more expanded than that in Fig. 5C. Both these observations go against engineering models: for example, Itti and Koch (2000) predict that featureless areas will not be fixated, and that differences between observers appear only towards the end of exploration; as our results show, such models are simply insufficient and cannot be applied to artwork.

initial fixation EYE MOVEMENTS AND PICTORIAL SPACE PERCEPTION |

C D

Figure 5. Eye movement patterns by a non-expert (A) and an art expert subject (B) over the total period of exploration (i.e. 1 min of recording); in C and D are shown only the first five seconds of exploration, i.e. the first 8—9 saccades andfixations. Lines indicate saccade efe movements connecting succesnve fixations to different portions. Saccades and fixations were determined using standard offline algorithms: the saccade started when eye velocity exceeded 15°/s, its offset when eye velocity dropped below 10Ί s. Fixation was determined as the time period between two saccades (minimal duration 50 ms), during which eye displacement was less than 1°. Blue tines connect successive fixations that have durations within the known normal range (50 to 400 ms); black tines are used when the eyes move temporarily outside the painting space; green tines indicate the first fixations of exploration of the painting. Initially (A, B), both subjectsfixated on a cross target displayed on the lower Ιφ corner of the screen. Upon uncovering of the painting the first 1, 2 fixations (green tines) by both subjects were directed toward the central dog elements. The deference between the subjects relies on the subsequent fixations. The total pattern of eye movements is very different. A striking difference is presentfrom the first five seconds: the first 8—9 fixations by the non-expert are all focused on the dog (Fig. 5C), while those by the art expert are spread to the dog corners of the carpet and the upper part (Fig. 5D).

Our observations of Bacon's original painting extend and confirm laboratory studies carried out on reproductions of other modern paintings (see Nodine et al. 1993, Zangemeister et al. 1995). They show two types of exploration: one driven mainly by figurative elements, the other more global, including points important for the space composition. Spatial global eye scanning is subject- dependent but art training promotes such behavior. Importantly, our study also suggests that eye movement patterns can show spatial sensitivity better than verbal reports. In fact, the link between eye movement patterns and verbal reports is rather loose; the movements of the eyes seem to be more in line with pre-attention analysis. A more complete report of this study can be found elsewhere (Kapoula & Lestocart, 2006). A more recent study re-examined perception and eye movement patterns by students in art versus science schools, exploring this time a digital reproduction I KAPOULA, YANG, VERNET& BUCCI

of the same painting, Study of a dog, presented on a computer screen. Subjective tests confirmed differences in space and movement perception between the two groups. Interestingly, eye movements over this reproduction showed that the areas most fixated were the same as for the real painting. Again, science students with no deep art knowledge focused on representational elements while art students produced a more global scanning including points key to spatial structure. Thus, studies with reproduction of paintings are still valid, even though aesthetic pleasure relative to the originals is presumably less (Kapoula 2007, Oudiette 2007).

Space perception and eye movement exploration of the painting L'Annunciazione del polittico di Saint'Antonio by Piero della Francesca The motivation of this study was to examine how viewers without art training perceive and explore space and perspective in Piero's paintings. Do they focus on representational elements (as for modern art) or does Piero's unique, geometric and perspective formal purity call for a real spatial navigation regardless of prior art knowledge? The choice of analyzing the Annundation was due to its clear-cut double bound character: on the one hand, it has a strong representational import (by virtue of the Biblical motif), on the other hand, it displays an explicit emphasis on purely formal, geometrical, and spatial constructive elements (by virtue of its use of perspective). It thereby enables two distinct perceptual, exploratory trajectories: one centered on elements pertaining to the motif, another on the spatial, presentational elements by means of which the motif is pictorially interpreted. Petitot (2004) presents an excellent analysis of the painting emphasizing the dual dimension - content- spatial structure; see also Bundgaard (2002). The participants in the study were four students from the physiology laboratory and a research collaborator; no subject had seen the painting before and no one was art expert. The study was carried out in the laboratory. A high quality digital image of the painting was presented on a PC screen for 30 seconds. Observers seated in front of the computer at a distance of 50 cm. Their eye movements were recorded with video-oculography (Chronos, Skalar, system, at 200 Hz). Initially, the observers fixated a line marker at the lower left part of the screen, assuring the same starting position. Then the painting was shown for 30 seconds with no instruction (to avoid influence); observers spontaneously explored the painting and movements from both eyes were continuously recorded and stored on the computer for further analysis. After exploration, observers were asked to describe the painting, to give as many details as possible, then to draw it by memory. Note that during eye movement recording, observers were not aware of the additional tasks (description, drawing). EYE MOVEMENTS AND PICTORIAL SPACE PERCEPTION |

Figure 6. Lines connect successive fixations following saccades; note that the analyñs integrated only saccades larger than 1 degree bringing the eyes to different areas. (B) First five fixations in the order done by subject MV. At the beginning the subject fixates at the left down corner of the screen; upon appearance of the painting the first fixation is directed at the center perspective area; fixations 3 and 4 are also devoted to central architectural items. (C) Eye scanning of the painting for the total period of 30 s during which the painting was present on the screen; note the large movements along the vertical midline, the horizontal movements linking symmetrically placed figures. After the 30 s of free exploration, the subject was instructed to fixate successively the four comers of the painting; this procedure allowed us to calibrate the placement offixations on the painting during the free exploration.

Time analysis of eye movement patterns This analysis takes into account only movements larger than half a degree ignoring the multiple small refixation movements within the same area. The first fixation was directed toward the central perspective area, the next one I KAPOULA, YANG, VERNET& BUCCI

toward the Virgin and the third toward the Angel. Thus, perspective placed at the centre is the first salient element capturing viewer's eyes (see Fig. 6B). Subsequently, sequences of fixations common to the observers could be also identified. Namely, the symmetric saccade steps between the faces or the limbs of the angel and the virgin, large vertical steps along a vertical line in the middle passing through the perspective area; the external staircase frame of the chapel was also explored by these observers (see Fig. 6C). Saccade steps were large (mean size 5.66°) connecting symmetric points, and forming geometric forms. Successive movements and fixations to the series of columns depicted at different depths were indicating fine grained exploration of space. The eye explored almost all important points of the space construction and was not limited to characters. The observations contrast those previously reported for modern art paintings by Bacon. Albeit naïve, observers do not focus on representation.

Eye movement patterns - loci of frequent fixation We selected five key areas of the painting, the three central representational figures (Angel, Virgin, Holy Spirit), the key perspective zone which is a commonplace symbolic representation of Christ, and the lower central, decorated architectural front. For each of these regions we calculated the percentage of fixations. It is the perspective area at the centre that received most of the fixations. Many art-historical studies state that the center of the composition is a perceptual space of great relevance (see Arnheim 1988). Central areas are known to attract fixations (see Locher et al. 1996) but in this case both the power of a centre and of the depth make this area irresistible for observers' eyes. In total, 85 % of the fixations were devoted to the surface of the five areas together. Thus, the selection of these areas by the investigator, seems to correspond to the most fixated surface by the observers (see Fig. 7). ETE MOVEMENTS AND PICTORIAL SPACE PERCEPTION |

1 13% 2 8%

4 28% 5 18% 3 12%

6 8%

Figure 7. Boxes show the regions chosen by the investigator; for each reff on is indicated the percentage of fixations made by the group of 5 subjects. The centralperspective area (4) receives most of the fixations.

Verbal reports The first aspect described by all viewers was the perspectival central part of the painting; various terms were used (columns in perspective, corridor, background blank plane, arch, arcade). The external frame was the next architectural element described with terms such as staircase, chapel, building exterior, cut off of the painting.

Drawings Despite their rather poor drawing skills, almost all observers reproduced perspective and stair-case elements, respecting more or less the composition of the painting. Thus, space composition, perspective, and architecture were perceived and retained as the major elements together with the representational elements (see Fig. 8). Eye scanpaths, verbal reports, and drawings were thus tightly linked. A more complete report of this study can be found elsewhere (Kapoula et al. in press). I KAPOULA, YANG, VERNET & BUCCI

Wuvt w.t. ju yJ;^»«

8. After the verbal description, a page and a penal was given to the subjects to draw the painting by memory. This is the drawing by memory of the painting by subject MV. The figures are accurately placed, the depth is depicted by converging ânes, and the columns an inàcated. Verbally, this subject described the painting as follows: Mary to the right, and to the left the angel Gabriel; there is a series of columns in perspective; the painting is cut on the upper part in the shape of stairs. - Note the congruence between verbal report, drawing and the firstfive eye fixations.

Navigation in 3D space — vetgence changes When we explore objects with different orientations and situated at different degrees of depth in the natural 3D environment, we perform saccades (movements of both eyes in the same direction, as presented hitherto), as well as vergence of eye movements. The latter allow to adjust the angle of visual axes according to spatial depth (the eyes converge for an object close by and diverge for a distant object). Binocular visual disparity and accommodation of the lens of the eyes are the main stimuli eliciting vergence of eye movements. Monocular depth cues such as perspective, texture, blur, overlapping of objects, could also assist vergence, as well as high level cues such as a sense of proximity (for a review see Howard 2002). As the painting is 2D, the vergence angle during eye movement exploration should not change and should be appropriate for the distance of the PC screen (50 cm). An inappropriate vergence angle would interfere with clear single vision. From the physiologic point of view, saccades are not perfectly yoked for the two eyes. When the abducting eye (the outer eye) makes a larger movement, the eyes diverge slightly during the saccade; this divergence error is corrected by a convergence movement during the fixation that follows the saccade (see Yang & Kapoula 2003). Such physiologic changes of vergence are very small in size (<0.5° for large saccades of 20°, i.e. <5%). Most likely they are harmless and do not compromise single binocular vision. In contrast, while exploring the painting by Piero della Francesca we observed increased changes in vergence EYE MOVEMENTS AND PICTORIAL SPACE PERCEPTION | angle both during saccades and during fixations. The average vergence change between the onset and offset of the saccade was 0.8° (about 14% of the saccade size); the vergence change during the fixation that followed the saccade was also high 0.7° (11%). Such vergence changes brought instantaneously the point of intersection of the two visual axes 6 to 8 cm in front or beyond the screen where the painting was displayed (at 50 cm). Notice that for saccades and fixations, in a control study involving free exploration of letters spread over a surface similar to the one of the painting on the screen, we observed smaller variation of vergence angle (7% of average saccade sizes, 3.17°). On the basis of these observations we are tempted to conclude that the large variation of vergence angjie while exploring the painting is due to a high-level component of pictorial depth perception driving the eyes out of the physical plane and into the virtual 3D pictorial space. An alternative interpretation would be that marked changes in vergence angle are responses to various depth ambiguities and compression of perspective such as described by Petitot (2004). In other words, the vergence angle oscillates perhaps in mirror with the uncertainty of the depth perceived. For instance, as noticed by Arasse (1999), the interior of the walls of the colonade in the back is shown with high details as if it were to be seen close up; this is another conflict, which the observer's visual system either tries to solve or is subjected to. Whatever the mechanism is, allowing intersection of visual axes in front or beyond the physical plane, is reminiscent of the natural strategy of viewers in front of original paintings in museums, approaching or moving away from the canvas to appreciate the work better. The modulation of vergence angle and the resulting blurred vision may aim to produce a holistic, interior view with increased sense of depth — a source of aesthetic pleasure.

Conclusions In summary, we propose a conceptual model that is moderate version of die scanpath theory, according to which eye movements provide important physiologic substrate for coding and retrieving visual images and paintings from memory. We provide experimental evidence for this with a study on Paysant's artwork. Even naïve subjects when asked to explore the artwork by memory reproduce an eye movement pattern similar to their pattern when they visualized the artwork. Experiments 2 and 3 deal with the importance of eye movement patterns in the study of analysis and perception of pictorial space construction including perspective and depth. The methodology we introduce in these experiments includes spontaneous eye movement exploration without any instruction, followed by questionnaires, and free verbal descriptions and drawings allow to explore analysis and perception of paintings at multiple levels. The study on Bacon's painting clearly indicates that analysis and perception of hidden spatial complexity is promoted by art knowledge and training. By contrast, the Renaissance painting by Piero della Francesca causes spatially inquisitive exploration for naive, non art-trained subjects, guided by I KAPOULA, YANG, VERNET& BUCCI

the architecture, perspective and depth power of the painting. It still remains to be tested how art-trained subjects are behaving in front of such paintings. The analysis of fluctuations or changes of vergence angle is a novel eye movement parameter to consider in the field of eye movement research and perception of paintings. It deals with a fundamental question in art history and creative process studies regarding the importance of viewing distance and viewpoint for given artwork. We hope these experiments, although preliminary, will stimulate research with various types of paintings.

Acknowledgments This research was supported by the program Esthétique, Complexité, Expéri- mentation & Modélisation of the Institut of Complex Systems, Paris Ile de France, see .

References Arasse, D. & Damisch, H. (1999). De la perspective en peinture. Paris: In Medias Res. Arnheim, R. (1988). The Power of the Center. A Study of Composition in Visual Arts. Berkeley: California Press. Betancourt, M. (2002). Motion Perception in Movies and Painting: Towards a New Kinetic Art [electronic article]. Ctheoiy.net. , Retrieved 13/09/2010. Brandt, S. A. & Stark, L. W. (1997). Spontaneous Eye Movements During Visual Imagery Reflect the Content of the Visual Scene. The journal of Cognitive Neuroscience 9, 27-38. Bundgaard, R. F. (2002). Presentation and Representation in Art - Ontic and Gestaltic Constraints on Aesthetic Experience, Visio 7 (1-2), 187-203. Buswell, G. T. (1935). How people look at pictures: a study of the psychology of perception in art. Chicago: Chicago University of Chicago press. Deleuze, G. (2004). Francis Bacon: The logic of sensation. Paris: Continuum International Publishing Group-Mansell. Howard, I. (2002). Seeing in depth. Toronto: I. Porteous. Itti, L. & Koch, C. (2000). A saliency-based search mechanism for overt and covert shifts of visual attention. Vision Research 40,1489—1506. Kapoula, Z. & Lestocart, L. (2006) Perception de l'espace et du mouvement dans Study of a dog de F. Bacon. Intellectica 2, 44. Kapoula, Z. (2007). Un autre regard sur Francis Bacon. Pour la Science 346, August. Kapoula, Z., Bucci, M. P., Yang, Q., & Bacci, F. (in press). Perception of Space in Piero della Francesca's Annunciation: An Eye-Movement and Art-Historical study. Leonar- do. Leigh, J. & Zee, D. (2006). The neurology of eye movements. New York. Oxford University Press. Livingstone, M. S. (2002). Vision and Art: The Biology of Seeing. New York: Harry N. Abrams. EYE MOVEMENTS AND PICTORIAL SPACE PERCEPTION |

Locher, P. (1996). The contribution of eye-movement research to an understanding of the nature of pictorial balance perception: a review of the literature. Empirical Studies of the Arts 14, 143-163. Locher, P., Krupinski, E., Mello-Thoms, C., & Nodine, C. (2007). Visual interest in pictorial art during an aesthetic experience. Spatial Vision 21, 55-57. Nodine, C., Locher, P., & Krupinski, E. (1993). The role of formal training on perception and aesthetic judgment of art compositions. Leonardo 26, 219-227. Noton, D. & Stark, L. (1971). Scanpaths in eye movements during pattern perception. Science 171, 308-311. Oudiette, D. (2007). Espace et mouvement pictural: de la perception subjective à la phydologe oculomotrice. Master Sciences et Technologie - Neurosciences, Université Paris VI, 2007 (director Ζ. Kapoula). Paysant, M. (2005). Inventarium. Paris: Archibooks/ La Fonderie. Perito t,J. (2004). Morphologe et esthétique. Paris: Maisonneuve et Larose. Yang, Q. & Kapoula, Z. (2003). Binocular coordination of saccades at far and at near in children and in adults. Journal of Vision 3, 554-561. Yarbus, A. L. (1967). Eye movements and vision. New York: Plenum Press. Zangemeister, W. H., Sherman, K., & Stark, L. (1995). Evidence for a global scanpath strategy in viewing abstract compared with realistic images. Neuropsychologia 33, 1009-1025. Zeki, S. (1999). Inner vision. Oxford: Oxford University Press. Alessandro Pignocchi

What Is Art? A methodological framework for a pluridisciplinary investigation

Over the last decades, disciplines such as cognitive psychology, evolutionary psychology and the neurosciences have shown an increasing interest for art. It remains unclear what kind of relation these "young disciplines" should have with more traditional endeavors and, more generally, in which way they can enrich our understanding of art. In this paper, I lay down the foundations of a methodological framework which distinguishes between three basic topics: (1) the investigation of the cognitive phenomena elicited by the experience of things that we call 'artworks', (2) the investigation of the psychological structures determining the intuitive categorization of something as art and (3) the study of the intuitions and arguments used to build justifications about whether something is a work of art. Each of these topics can benefit from the cognitive approaches, provided that some specific methodological recommendations are respected.

CORRESPONDENCE: Alessandro Pignocchi. Institut Jean Nicod (CNRS ENS-DEC EHESS), Paris, France. EMAIL [email protected]

1. Introduction Over the last decades, "young disciplines", such as cognitive psychology, evolutionary psychology and the neurosciences (which I will group for simplicity under the label "cognitive approaches"), have shown an increasing interest for art (Dutton 2008, Freedberg, & Gallese 2007, Cavanagh 2005, Ramachandran & Hirstein 1999, Zeki 1999, Solso 2005, 1994, Dissanayake 1992). Art is certainly one of the most complex domains of human activity: it has a long and rich history, miscellaneous manifestations, intricate interconnec- tions with many other human activities and, crucially, it can mobilize nearly all kinds of human abilities and competences. Thus, we can surely welcome the fact that the theoretical investigation of art is now enriched by disciplines which give a new perspective — a naturalist perspective - on human abilities. More- over, traditionally, the philosophy of art — a leading discipline in the theorization on art - has had normative ambitions (Schaeffer 1992). The goal was, for instance, to find the right function of art, or the right way to interpret artworks. After the influence of Wittgenstein's Philosophical Investigations, some authors took a descriptive watershed (Weitz 1956, Kennick 1958, Dickie 1969, Carroll

Cognitive Semiotics, Issue 5 (Fall 2009), pp. 122-135 WHAT IS ART? I

1993, Schaeffer 1996, Michaud 1999, Gaut 2000). They realized the importance of appropriate description, and the importance of clearly distinguishing description from normative discourse. The new aims were, for instance, to describe what people actually call 'art', or the way people actually interpret artworks. Cognitive approaches, as scientific approaches, can now help to accentuate this descriptive watershed. This is another reason to encourage cognitive approaches to integrate the ongoing reflection on art. However, the kind of relation cognitive approaches should have with more traditional disciplines - such as philosophy, history, sociology or anthropology - and, more generally, the way in which they can enrich our understanding of art remains unclear (Walton 2007). In this paper I build the scaffoldings of a methodological framework with the aim of both clarifying what kind of contribution should be expected from cognitive approaches in a pluridisciplinary investigation of art and helping their positive interaction with traditional disciplines. I will first identify two faux pas firequendy committed by cognitive approaches to art - ηοη-speäficity and esserttiali^ation - which are responsible for some misunderstanding between the tenants of cognitive approaches and the tenants of traditional disciplines. I will then show that these faux pas can hardly be avoided without addressing, in one way or another, the question "what is art?" I suggest that behind this question there is a variety of different questions, and that distinguishing among these different questions is the first step towards a methodological framework for a pluridisciplinary investigation of art. Notably, it is essential to distinguish the questions which concern the properties of the things that we call 'artworks' (or the cognitive phenomena elicited by those properties) from the ones linked with the concept of art. The main difficulty related to the inquiry into the properties of artworks consists in the fact that the set of things that we call 'artworks' does not constitute a natural kind with an associated set of necessary and sufficient properties. Thus, stricdy speaking, there is no 'nature of art' to study. There- fore, it is necessary to question explicidy the nature of the fundamental aims of a cognitive approach to art, in order to understand how such approaches have to choose their objects of investigation. The investigation of the concept of art also requires some methodological precautions. Notably, it is essential to distinguish between the intuitive categorization of something as art and the elaboration of reflective arguments and justifications to defend or reject the arthood of a thing. The properties of artworks, the intuitive categorization of something as a piece of art and the reflective justifications concerning the arthood of a thing are three clearly distinct topics of investigation which are the armature of the methodological framework that I propose to endorse. As we shall see, the relevance of this three-stage distincdon comes from the fact that for each of these topics, cognitive approaches to art have to respect particular methodological recommendations. 124 I A. PIGNOCCHI

2. Two Problems for Cognitive Approaches to Art: Non-specificity and Essentialization The main faux pas of cognitive approaches to art is non-specifiäty. An approach to art is non-specific when it investigates a property of some artworks (or a psychological or neural phenomenon elicited by some artworks) which is also both shared by many things other than artworks and not shared by many other artworks. Since this property (or phenomenon) is neither necessary nor sufficient for something to be art, it is unclear how its investigation can enrich our understanding of art as art. Ramachandran and Hirstein (1999) have spelled out a set of ten principles inspired, at least the most worked out, from the neurosciences, which suppo- sedly help understanding the nature of art, or the nature of great artworks. The problem of non-specificity in their work arises from the fact that, as noted by Hyman (2006), many artworks fulfill only one or very few of these principles, whereas many things that are not art fulfill many of them if not all. As an illustration, consider the first principle, the peak-shift effect. Roughly, the claim is that artworks exaggerate some features used by the brain in object recognition. The perception of these exaggerated features hyper-activates the dedicated neural circuit which activates the limbic system, a system responsible for feelings of pleasure. For instance, the perception of a Hindu bronze sculpture of the divinity Parvati hyper-activates the neural circuit dedicated to the recognition of the female body, which explains the pleasure to look at it. This claim is non-specific because (1) plenty of visual displays, the majority of which would not be considered artworks, hyper-activate some neural circuit dedicated to object recognition (a smiley, for instance, certainly hyper-activates the circuit dedicated to the recognition of happy facial expressions), and (2) plenty of artworks do not exaggerate any features used for object recognition (for instance many artworks aspiring to naturalism, such as Constable's Wivenhoe Park, conceptual art such as Duchamp's Fountain, or minimalist artworks such as Malevitch's White Square on a White Square). The same kind of argument can be developed for all of Ramachandran's and Hirstein's principles. As a consequence, it is not clear that these principles teach us anything specific about art. What is the difference, for instance, between claiming that some artworks fulfill some of Ramachandran and Hirstein's principles and claiming that some pieces of visual art activate the visual areas of the brain, or the visual area and the limbic system (claims for which non-specificity is flagrant)? In a similar vein, Freedberg and Gallese (2008) suggested that some pieces of visual art activate the mirror neurons system, either through their representational content or through the traces left by the artist's actions. Again, as noted by Casati and Pignocchi (2008), many things other than artworks activate the mirror neurons system, whereas many artworks presumably do not activate the mirror neurons system. WHAT IS ART? I 125

Another example is provided by Denis Dutton (2008), who proposes to characterize artworks with a cluster of twelve features inspired by evolutionary psychology, which shares some similarities with the set of principles spelled out by Ramachandran and Hirstein. Here also non-specificity is blatant because, as underlined by Casati (2009), many artworks possess only one or few of Dutton's features, whereas many things that are not art possess many of them. The problem of non-specificity is not easy to deal with. In fact, the only way to solve it direcdy would be to look for a property or a psychological phenomenon that is characteristic of artworks. However, at least so far, no one was able to find such a property or phenomenon and, as I will explain in the next section, it is likely that none is to be found. Thus, on the one hand, we would want cognitive approaches to enrich our understanding of art as such and, on the other hand, it is possible that there are no properties or phenomena characteristic of art as such which could be investigated by cognitive approaches. Then, are cognitive approaches condemned to be non-specific? In trying to address the problem of non-specificity, several authors made an even worse error, which brought to essentialization. Essentialization is the arbitrary transformation of a descriptive claim (e.g. 'some artworks have the property Ρ") into a normative claim (e.g., 'a thing must have the property Ρ to count as real art7). Essentialization appears, for instance, when an author investigates a property, or its cognition, and claims more or less explicitly that things that do not present this property are not art. Denis Dutton, in his book The Art Instinct, is brought to essentialization when he claims that some pieces of minimalist conceptual art are not art (or very bad art) because they do not fit his own definition of art. Ramachandran's and Hirstein's rhetoric also contains traces of essentialization, even if in a more implicit way than in Dutton's book. They claim that their principles characterize great artworks. Since all their principles describe ways to activate the limbic system, and since this system is responsible for feelings of pleasure, their theory implicidy contains the normative claim that a piece of art must be pleasurable to count as great art. Notice that normativity as such is not a bad thing. As said in the introduction, it has been the rule in many fields of the theorization of art. What is fallacious is to hide normativity behind professedly descriptive and scientific theories (the so-called 'naturalist fallacy"). Thus, as scientific accounts, cognitive approaches have to avoid any form of normativity. Of course it is always possible to use the results obtained by a cognitive approach to build a normative discourse about art. But in this case, normativity has to be very explicit, and clearly distinguished from the objective description and explanation of facts provided by cognitive approaches. Thus, the first question to be addressed to improve our understanding of the way cognitive approaches to art should proceed is: how can the problem of non-specificity be avoided without falling in essentialization? I A. PIGNOCCHI

This question is not an easy one, since, in order to be specific, it seems that it is necessary to find, first, what is specific to art; in other words, the above question forces us to address the question of what is art?

3. What is Art? Artworks and the Concept of Art As said, no one has found, so far, a property or a phenomenon necessary and sufficient for something to be art. It would be poindess, thus, to try to answer this question here. However, as we shall see, discussing this question and, more precisely, specifying what is problematic about it, helps finding a way to avoid the problem of non-specificity. One thing, among many others, which explains the difficulty of the question 'what is art?' is that it hides a variety of different questions. First, the question 'what is art?' can be a question about the properties of artworks and the way we cognize them, or a question about the concept of art. Focusing on some properties of artworks is quite a different thing than studying the concept of art. To clarify this distinction between an object and a concept, consider the concept 'water'. It is clear that studying the physico-chemical properties of water is completely different from studying the way people identify, conceptualize and talk about 'water'. The physico-chemical properties of water are a research topic for physics and chemistry, whereas the concept of water as it is commonly used is a topic for disciplines such as sociology, psychology, anthropology or history. For art as well, questions about the properties of artworks have to be clearly distinguished from questions about the way people use the concept of art in their thought and conversation about art or in their interactions with artworks. We are strongly inclined to understand the question "what is art?", or "what is water?" as a question about the objective properties of the object, not about the way people use the concept. For water, this is not a problem, since there is a natural kind — the things composed by molecules of H2O - the members of which share a set of characteristic properties independent of what people call water, and which coincides more or less with it. Moreover, we willingly accept that what science says about the things composed by molecules of H2O has a priority over what people call "water". For instance, if someone calls "water" something that is not composed of H2O, we are naturally inclined to accept that this person makes a mistake. For art things are more complex. The concept of art has a quite complex history during which its use in language has changed very significandy. For the Greeks, the word 'art' was only used to refer to practices that required a high level of skill (e.g. the 'art of medicine' or the 'art of soldiering' (Carroll 1999)). The Greeks already had a set of practices that we call art today (sculpture, painting, poetry...) but they had no unifying concept for them. The unification of the different practices that we consider today as paradigmatic of art occurred only in the eighteenth century with the publication of The Fine Art Reduced to a Single Principle by Charles WHAT IS ART? I 127

Batteux. According to this author, painting, poetry, music, sculpture and dance must be unified under the common concept of Tine-Art' because they all share a common function: the imitation of the beautiful in nature. It is not clear whether this principle could unify these different practices as suggested by Batteux (for a discussion, see Carroll 1999). In any case, the evolution of art after the eighteenth century has quickly and largely exceeded the function assigned by Charles Batteux. Artists have produced many artworks that were clearly not designed to imitate beautiful natural things. Nevertheless, the theory of Charles Batteux significantly influenced the work of artists, the way people thought about art, as well as the theoretical work about what art is. Other theoreticians argued that the theory of imitation was wrong, and proposed alternative unifying principles (the possession of significant form (Bell 1915), the expression of artist emotions (Tolstoy 1899), the clarification of its feelings (Collingwood 1938) or the function of eliciting an aesthetic experience (Beardsley 1983)). Again, it was soon acknowledged that, despite the attractive- ness of those theories, none of them was able to capture the ensemble of things that people considered as art. However, again, those theories influenced the works of artists and the way people thought about art. These evolutions inspired other theoreticians, who elaborated even more powerful theories about the nature of art (such as the theory of embodied meaning (Danto 1981) or the institutional theory (Dickie 1969)). As their predecessors, those theories failed to capture the ensemble of things that we commonly call art, but they inspired artists and audience. As shown by this extremely simplified history of the concept of art, what is considered as 'art' at a given historical period is a function of what was considered art at a preceding period, and the driving force of this evolution is a complex interaction among theoreticians, artists and audience (for a development of this argument see Weitz 1956, Kennick 1958, Levinson 1979, Carroll 1993). As a result, for the case of art, contrary to water, there is no natural kind, with a set of unique and characteristic properties independent of what people consider as art, and which coincides more or less with the set of things that people call art. Neither is there a natural kind for which science would have the same sort of natural legitimacy to claim that this and only this is real art, as it has in the case of water. Of course, it is possible to isolate a subcategory of things with a set of common and characteristic properties (having certain formal, expressive or semantic properties, for instance, or activating the limbic system) and claim that only those things are 'real art' and that people who call art things that do not share those properties make a mistake. But this move is what I have called essentialization. In fact, this kind of claim is necessarily purely normative, since it necessarily excludes from the category of art things we commonly consider as art, and includes things we are not used to consider as art. Again, this kind of normative move can be interesting, but since it is normative, it is not appropriate for cognitive approaches. I A. PIGNOCCHI

Maybe the problem can be solved by claiming that art is simply what we call art, and by investigating the properties of these things. This move has two problems, however. First, the set of things that we call 'art' has extremely fuzzy boundaries because of important individual, social and cultural variations. Secondly, it does not solve the problem of non-specificity, since the things that we call art do not share any characteristic properties. Two reactions to this conclusion are possible. The first is to explicidy aban- don the idea of studying a property which is specific to art and to reconsider the goals of cognitive approaches, in order to understand how the objects of investigation of such approaches have to be isolated, given that the characterization of the nature of art, in a strict sense, is impossible. The second possible reaction is to study the concept of art itself, asking questions such as: how do we use it? How is it possible that, despite the fact that it does not refer to a natural kind, we understand what we are saying, at least to some extent, when we speak of 'art'? What determines the intuitive categorization of an object as art? What determines the kind of arguments that we use to argue for the arthood of an object? These two possible solutions to the problem of non- specificity are the topic of the next two sections.

4. What is 'Relevant' for a Cognitive Approach to Art? The major conclusion of the last section is that cognitive approaches have to give up the ambition of investigating a property or a phenomenon that would be, strictly speaking, specific to art, because art is not a natural kind with a set of characteristic and unitary properties which could be an object of scientific investigation. Any property or phenomenon which could be studied by the cognitive approaches will be at best a feature of some, but not all, artworks, and will always be a feature of things that are not art. However, among the set of properties or phenomena which are shared by some artworks, not all seem to have the same degree of relevance for our understanding of art. The capacity to activate the visual areas of the brain, for instance, or the capacity to activate the color perception areas, is a property which is shared by many artworks. However, the study of these properties does not seem to be relevant for our understanding of art. Then, what properties are 'relevant' for a cognitive approach to art? To answer this question, it is necessary to notice that, contrary to the study of quantum mechanics, for instance, or of other areas of fundamental research, art is a domain that is strongly intertwined with our ordinary life. We interact daily with artworks, we think about them and we have conversations about them. Thus, what we expect from a cognitive approach to art - or from any theoretical approach to art - is that it could, directly or indirecdy, enrich our real relations with art, i.e. enrich our interactions with artworks and the way we are able to think and speak about artworks and art. Therefore, two constraints determine the kind of properties or phenomena that are 'relevant' for cognitive approaches to art. On the one hand, of course, WHAT IS ART? J they have to be potential objects of scientific investigation. This excludes, as I have shown, art as such, since the concept of art does not refer to a set of things with common and characteristic properties which could be investigated scientifically. For the same reason, it is likely that this first constraint also excludes very general and undefined topics such as aesthetic experience, the value of art, or the nature of great artworks. On the other hand — and this second constraint is quite unusual for a scientific approach - the topic of investigation must be connected with questions and theories that have the power of enriching our real and ordinary relations with art (our experience of some artworks, the way we reflect on them, the way we can speak of them...). This excludes very circumscribed properties or phenomena, such as the activation of visual areas or the perception of color (see Casati 2003 for a similar point). In other words, it is necessary to find what Casati and Pignocchi (2008) called 'mid-level hypotheses', i.e. the hypotheses which are sufficiendy circumscribed to be object of scientific investigation and, at the same time, sufficiendy connected with our real relations to art. One method to find this kind of hypotheses would be to take inspiration from theories developed by 'traditional' disciplines, such as philosophy of art or history of art, which are, at least some of them, explicidy concerned with the potential enrichment of our real relations with artworks (Baxandall 1985, Wollheim 1987). Generally, these theories are not concerned with the constraints of finding objects of scientific investigation. Thus, the method consists in isolating in these theories the points or questions which satisfy the constraints of scientific investigation in order to enrich, precise or improve the overall theory, which itself has the potential to enrich our relations with art. Examples of mid-level hypotheses could possibly be found, for instance, in theories concerned with artist's intentions and the role played by their retrieval in our understanding and appreciation of an artwork. On the one hand, the causal role of intentions on behavior and the way we retrieve the intentions of an agent on the basis of observable cues is now a standard topic for disciplines such as cognitive psychology or neuroscience (for a review see Frith & Frith 2006). On the other hand, many very influential theories in philosophy of art (Levinson 1979, Danto 1981, Wollheim 1987) and in history of art (Baxandall 1985) are based on the concept of intention. Moreover, at least some of these theories are explicidy concerned with topics direcdy connected with our everyday relations to art ('what role could or should play my knowledge of the artist's intention on my interpretation and appreciation of an artwork?', for instance, is a question that any person who is even a litde interested in art has asked). Thus, it seems likely that there are some hypotheses concerning artist's intention which would satisfy the two constraints discussed above (other examples of mid-level hypotheses can be found in Cavanagh 1999, Casati 2003 and Pignocchi, submitted (a)). A first step in the cognitive investigation of the artist's intentions and their retrieval would be using the concepts of cognitive psychology to provide a 130 I A. PIGNOCCHI

taxonomy of the different kinds of intentions which play a role in the production of an artwork, such as 'meta-cognitive intentions' (the intention that a property of an artwork will be perceived as being produced with the intention to elicit a given effect), 'simple prepositional intentions' (the intention to elicit a given effect) and 'intentions in action' (the intentions which represent a given goal in the format of concrete actions), as well as a taxonomy of the different kinds of mechanisms used to retrieve these different kinds of intentions (inferential and simulationist mechanisms, for instance) (for an example of a cognitive approach of artist's intentions see Rollins 2003 and Pignocchi, submitted (a)). As an illustration, consider the painting Woman taking Tea from Chardin. If we are asked about the intentions of the painter, our first spontaneous answer could be, for instance, that Chardin wanted to evoke a certain feeling of loneliness mixed with a subtle kind of melancholy. But this is only one (hypothetic) state of mind which could have determined the aspect of the painting. For the painting to be completed, nearly any Chardin's gestures had to be caused by a complex and organized set of intentions, ranging from the most abstract, such as the ones mentioned above, to the most concrete visuomotor planning of actions, such as the simple white brush stroke thanks to which he evokes the reflection of light on the teapot. In order to bridge the gap between very abstract intentions and very concrete gestures, Baxandall (1985) has described the influences exerted on Chardin by the scientific theories of visual perception of his century. The main topic of these theories, inspired by Locke and Newton, was to investigate the active role played by the subject during perception. According to Baxandall, Chardin took inspiration from some principles of these theories in order to create a set of complex perceptive effects. Notably, he blurred the contour of the face, because he knew that faces strongly attract attention, and he did not want that the observer could stabilize his attention on a single point. To counterbalance the attraction of the face, he also created an area running from the nape to the hand holding the cup, on which attention has multiple attraction points created either by contrast, luminosity or clearness. He further destabilized perception by creating conflicts between linear perspective and various principles describing the role played by colors on illusions of size and distance. He also played in multiple ways with the principle according to which acuity diminishes from the fovea toward the borders of the visual fields. The intentions connected with the use of all these principles are of course organized inside more abstract plans, concerning properties of the general experience elicited by the painting. To recover the different kinds of intentions involved in the production of Woman taking Tea, we use very different kinds of psychological mechanisms. To retrieve very abstract intentions we possibly use inferential processes which are quite automatic for many of us, possibly because we are culturally trained to understand painting in this way (Schaeffer 1992). Instead, when we pay WHAT IS ART? J attention to the brush strokes we possibly activate a motor simulation of some aspect of Chardin's gestures. When we think about the influence of the scientific theories of visual perception we need to make an effort to put ourselves in the shoes of a painter of the eighteenth century who is impressed by the progress of science and who wants to play with scientific principles. We also recruit inferential mechanisms to understand, for instance, whether Chardin wanted the spectator to recognize that he used the contemporary scientific theories of vision or if he just wanted to use these theories to produce certain effects. All this is exaggeratedly simplistic of course, and much more work is needed to come up with a plausible scenario of the intentional process of creating Woman taking Tea (a process that we experience, whatever our knowledge, as highly coherent) and to describe possible psychological mechanisms used to recover iL More work is needed also to show that it is possible to draw the contour of a general taxonomy of the different kinds of intentions involved in the production of an artwork and of the psychological mechanisms used to recover them. The point here is only to show that this topic could possibly fulfill the two constraints: being a potential object for science and being able to enrich our real relations with art. Let us now move to the other possible answer to the problem of non- specificity: the investigation of the concept of art.

5. Intuitive Categorization and Justification Here also, some key distinctions have to be drawn, since it seems unlikely that there is such a thing as the concept of art, which could be a single and clearly defined object for science. In order to draw the relevant distinctions, it is useful to take inspiration from the field of ethics, where the cognitive theories are more advanced than in the field of art. In the psychology of morals, one of the most basic distinctions which frame the field is the one between moral intuitions and moral justifications. Moral intuitions are what determines our first and spontaneous answers to questions of the kind "is this behavior moral or immoral?" Moral justifications are what we say when we are asked to provide justifications for our answer. It has become clear by now that moral intuitions and moral justifications are determined by distinct psychological mechanisms, and it is often the case that when asked for justification we end up contradicting our first intuitive response (Green & Haidt 2002, Hauser et al. 2007, Green et al., 2008). In the case of art, the very same distinction can be drawn. The object of the first level of investigation would be the mechanisms which determine our intuitive and spontaneous categorization of something as art. The object of second level of investigation would be the set of arguments and justifications that we build when asked to justify our spontaneous categorization of X as art or as non art (or as art "to some extent", or as art "in a sense", etc.). I A. PIGNOCCHI

At the first level, a very simple protocol to begin the investigation of the mechanisms responsible for the intuitive categorization of something as art would simply parallel the one used to investigate moral intuitions. It would consist in asking subjects to answer questions such as "is X an artwork?" where X is either a real object or a fictive one constructed for the purpose of the design. The aim would be to connect the investigation of the intuitive use of the concept of art with the existing psychological literature on the concepts of artifacts (e.g. Bloom 1997, Casier & Kelemen 2005, Malt & Sloman 2007) and to clarify the possible specificity of the intuitive concept of art. Here also, a possible close collaboration with traditional disciplines, such as analytic philosophy of art, has to be envisaged. In fact, in the analytic philosophy of art there is a long tradition of designing thought experiments to test our intuitions relatively to borderline cases of arthood (Danto 1981). For the moment, of course, philosophers of art have used these thought experiments only to test their own intuitions. In order to investigate the intuitive concept of art from a psychological perspective, a first step would be to use these thought experiments to test the intuitions of various groups of subjects of different social and cultural origins. At the second level of investigation, much more work is needed to clearly frame the work expected from cognitive approaches. Cognitive approaches may, for instance, help to understand why some ideas about what art is have been more easily integrated than others by a given culture (the general paradigm for this branch of research could be, for instance, the epidemiological approach presented in Sperber 1996).

6. Conclusion: a Pluridisciplinary Investigation of Art I have suggested that cognitive approaches to art frequently suffer from a kind of non-specificity: they focus on a phenomenon which, despite the fact that it can be elicited by some artworks, does not inform our understanding of art in general, or does not clearly do so. It is necessary to solve this problem without falling into essentialization, i.e. without claiming, implicitly or explicitly, that the phenomenon in question is the real "essence of art". This has forced us to address a difficult question: What is art? The discussion of this question allowed us to distinguish among three different possible topics of investigations for the cognitive approaches: (1) the properties of the things that we call 'artworks', (2) the intuitive categorization of something as art and (3) the reflective justifications concerning the arthood of a thing. One difficulty at the first level of investigation is to choose the right object, since the set of things that we call "art" does not coincide with a natural kind which could be investigated scientifically. I suggested the 'relevant' objects of investigation for cognitive approaches to art at this first level are determined by two criteria. The first is trivial: the object of study must be sufficiently circumscribed to be a potential object for scientific investigation (this exclude, for instance, art as such, and possibly also aesthetic experience or the evaluation of WHAT IS ART? I

artworks). The second criterion is that the object must have the potential to enrich our actual relation with art, through the enrichment of our experience of some artworks and/or the way we converse and think about art and artworks. One way to satisfy this second criterion is to take inspiration from the traditional literature and, more precisely, from theories explicitly concerned with our real relations with art. The simultaneous application of the two criteria would thus amount to finding in traditional theories concerned with our real relation with art those points that can be the object of and benefit from a cognitive perspective. At the second level of investigation — the intuitive categorization of something as art — cognitive approaches could adapt the method they use to investigate moral intuitions. The material for experiments could be directly inspired by the classical thought experiments in analytic philosophy of art. On the one hand, this method provides a radically new perspective on very traditional questions for the analytic philosophy of art. On the other hand, it could provide to cognitive approaches new material, already richly analyzed by other disciplines, to explore the psychological structure on which concepts are built. The last topic - the way people build arguments when discussing the arthood of things — is extremely vast, and much more work is needed to subdivide it into more circumscribed topics, and to understand what kind of contribution could be expected from cognitive approaches. I have suggested that an epidemiological approach could help to describe the feedback effects of theoretical work on the folk notion of art.

Acknowledgment I would like to thank Roberto Casati, Valeria Giardino and two anonymous referees for helpful comments on earlier versions of this paper.

References Baxandall, M. (1985). Patterns of intention: on the historical explanation ofpictures. New Haven: Yale University Press. Bell, C. (1915). Art (Revised ed.). London: Chatto and Windus. Bloom, P. (1996). Intention, history, and artifact concepts. Cognition 60 (1), 1-29. Carroll, N. (1999). Philosophy of art a contemporary introduction. London: Roudedge. Carroll, N. (1993). Historical narratives and the philosophy of art. The Journal of Aesthetics and Art Criticism 51 (3), 313-326. Casati, R. (2003). Methodological issues in the study of the depiction of cast shadows. A case study in the relationships between art and cognition. Journal of Aesthetics and Art Criticism 62 (2), 163-174. Casati, R. & Pignocchi, A. (2007). Mirror and canonical neurons are not constitutive of aesthetic response. Trends in Cognitive Sciences 11 (10), 410. Casati, R. (2009). Book review: The Art Instinct, by Denis Dutton. [electronic version]. , Retrieved 13/09/2010. 134 I A. PIGNOCCHI

Casier, Κ. & Kelemen, D. (2005). Young children's rapid learning about artifacts. Developmental Science 8, 472—480. Cavanagh, P. (2005). The artist as neurosciendst. Nature 434, 301-307. Cavanagh, P. (1999). Pictorial art and vision. In R. A. Wilson & F. C. Keil (Eds.), ΜΓΓ Encyclopedia of Cognitive Science, (pp. 648—651). Cambridge, MA: MIT Press. Collingwood, R. G. (1938). The principles of art. Oxford: Clarendon Press. Danto, A. C. (1981). The transfiguration of the commonplace: a philosophy of art. Cambridge, MA: Harvard University Press. Dickie, G. (1969). Defining Art. American Philosophical Quarterly 6 (252). Dutton, D. (2008). The Art Instinct: Beauty, Pleasure, & Human Evolution. New York: Bloomsbury Publishing PLC. Dissanayake, E. (1992). Homo Aestheticus: Where Art Comes From and Why. New York: Free Press. Freedberg, D. & Gallese, V. (2007). Motion, emotion and empathy in esthetic experience. Trends in Cognitive Sciences 11 (5), 197—203. Frith, C. D. & Frith, U. (2006). The Neural Basis of Mentalizing. Neuron 50, 531-534. Gaut, B. (2000). 'Art' as a Cluster Concept. In N. Carroll (Ed.), Theories of art today (pp. 25—44). Madison: University of Wisconsin Press. Greene, J. & Haidt, J. (2002). How (and where) does moral judgment work? Trends in Cognitive Science 6, 517—523 Greene, J., Morelli, S. Α., Lowenberg, Κ., Nystrom, L. E., & Cohen, J. D. (2008). Cognitive load selectively interferes with utilitarian moral judgment. Cognition 107 (3), 1144-1154. Hauser, M. D., Cushman, F., Young, L., Jin, R. Κ.-Χ., & Mikhail, J. (2007). A dissociation between moral judgments and justifications. Mind & language 22 (1), 1- 21. Hyman, J. (2006). Art and Neuroscience. Art and Cognition workshop [PDF] (pp. 164— 179). , Retrieved 13/09/2010. Kaufman, D. A. (2002). Normative Criticism and the Objective Value of Artworks. Journal of Aesthetics and Art Criticism 60 (2), 151-166. Kennick, W. E. (1958). Does Traditional Aesthetics Rest on a Mistake? Mind 67, 317— 334. Levinson, J. (1979). Defining art Historically. The British Journal of Aesthetics 19 (3), 232- 250. Malt, B. C. & Sloman, S. A. (2007). Category essence or essentially pragmatic? Creator's intention in naming and what's really what. Cognition 105, 615—648. Michaud, Y. (1999). Critères Esthétiques et Jugement de Goût. Rayon Art: Jacqueline Chambón. Pignocchi, A. (submitted a). How the intentions of the draftsman shape perception of a drawing. Pignocchi, A. (submitted b). The Intuitive Concept of Art. Ramachandran, V. S. & Hirstein, W. (1999). The science of art: a neurological theory of aesthetic experience. Journal of Consciousness Studies 6, 15—51. Ramachandran, V. S. (2002). The neurological basis of artistic universals. Art and Cognition workshop [PDF] (pp. 59-62). , Retrieved 13/09/2010. Rollins, M. (1999). Pictorial representation: when cognitive science meets aesthetics. Philosophical Psychology 12, 387-^13. WHAT IS ARI? J 135

Rollins, M. (2004). What Monet Meant: Intention and Attention in Understanding Art. The Journal of Aesthetics and Art Criticism 62, 175—188. Schaeffer, J. M. (1992). LArt de I Age Moderne. Nrf essais: Gallimard. Schaeffer, J. M. (1996). Les Célibataires de IArt. Pour une 'Esthétique sans Mythes. Nrf essais: Gallimard. Sperber, D. (1996). Explaining culture: A naturalistic approach. Oxford: Blackwell. Solso, R. L (2005). The Psychology of Art and the Evolution of the Conscious Brain. Cambridge, MA: MIT Press. Solso, R. L. (1994). Cognition and the Visual Arts. Cambridge, ΜΑ: ΜΓΓ Press. Tolstoy, L. N. (1899). What is art? (A. Maude, Trans. 3rd ed.). London: W. Scott. Walton, K. (2007). Aesthetics - What?, Why?, and Wherefore? Journal of Aesthetics and Art Criticism 65 (2), 147-161. Weitz, M. (1956). The Role of Theory in Aesthetics (Vol. 15, pp. 27-35): Blackwell Publishing on behalf of The American Society for Aesthetics. Wollheim, R- (1987). Painting as an Art. London: Thames and Hudson. Zeki, S. (1999). Inner Virion: an Exploration of Art and the Brain. Oxford: Oxford University Press. Ellen Dissanayake

The Artification Hypothesis and Its Relevance to Cognitive Science, Evolutionary Aesthetics, and Neuroaesthetics

An etiological description of mother-infant interaction suggests that it was adaptive during human evolution. From their early weeks, infants universally respond to certain affect-laden elements of maternal communication that can be called "proto-aesthetic". I hypothesize that these capacities and receptivities, which evolved from about 1.7 million years ago to enable mother-infant bonding, became a reservoir of affective mechanisms that could be used subsequendy by ancestral humans when they first began to "artify" - that is, when they invented the "arts" as vehicles of ceremonial religious experience. The etiological, evolutionary, and cross-cultural foundations of the Artification Hypothesis challenge current influential ideas in cognitive science, evolutionary aesthetics, and neuroaesthetics about human aesthetic capacities and behavior. The hypothesis offers a broader and deeper view to these fields by emphasizing preverbal, presymbolic, pancultural, cross-modal, supra-modal, participative, temporally organized, affective, and affinitive aspects of aesthetic cognition and behavior. It additionally proposes that artifying is an unre- cognized but vital evolved component of human nature.

CORRESPONDENCE: Ellen Dissanayke. School of Music, University of Washington, Seattle, WA. EMAIL: [email protected]

Introduction Since the late 1950s, a number of once distinct areas in psychology and philosophy have gradually gathered under what might be called a cognitive umbrella or canopy. After incorporating an influential newcomer, cognitive neuroscience, it is not too much to say that two thousand years of theory about what once was called epistemology have been irrevocably altered—subsumed by and re-conceptualized as "the brain" and "mind". The successes of cognitive science, along with those of its fellow-traveler evolutionary psychology, have led some researchers recently to train their field glasses on philosophical subjects that have traditionally been of little interest to neuroscience or evolution — aesthetics and the arts. Thus we have "neuroaesthetics" (Ramachandran & Hirstein 1999, Skov & Vartanian 2009. Zeki 1999), "evolutionary aesthetics" (Voland & Grammer 2003), and now, in this volume, "aesthetic cognition".

Cognitive Semiotics, Issue 5 (Fall 2009), pp. 136-158 THE ARTIFICATION HYPOTHESIS |

My own academic work in arts and aesthetics was underway before the prominence of cognitive studies, although I was one of the first to approach these subjects from the perspectives of ethology and what soon became evolutionary psychology (Dissanayake 1974, 1979, 1984). Working in parallel with but outside the cognidvist project, I gradually developed a hypothesis that locates the emergence of aesthetic (or "proto-"aesthetic) capacities and sensitivities in the evolution of a universal biologically-essential behavior — coordinated interactions between human mothers and their immature infants that contributed to the survival of the latter and consequently to a mother's reproductive success. I argue that this "bonding ritual", as it began to evolve in early hominin mother-infant pairs - perhaps as early as 1.7 million years ago (Falk 2009, Leakey 1994) - relied upon distinctive emotionally affecting "proto- aesthetic operations"1 on common vocal, facial, and kinetic expressions of friendliness or affiliation, making them attention-getting and pleasurable. Importandy, the interaction became temporally-organized, allowing for further aesthetic effects (see below and Dissanayake 2000,2008, 2009). Consideration of proto-aesthetic components that inhere in this generally overlooked human behavior - notable for being dyadic and collaborative rather than individual and self-interested2 - provides important additions to or modifications of the orthodox cognitivist project as well as to orthodox evolutionary psychology. Indeed, it is not too much to suggest that "cognition" is not only embodied (Lakoff & Johnson 1999, Slingerland 2008) but that the readiness to respond to affect-laden (proto)-aesthetic operations, in place at the beginning of human life, precedes and influences the development of embodied cognition (or any other kind). Additionally, the evolutionary appearance and development of "artification" - a behavioral predisposition for deliberate use of aesthetic operations by adult individuals and groups in contexts and circumstances of uncertainty (occurring perhaps as early as 200 kya [thousand years ago])3 - has implications for prevalent cognitivist and evolutionary ideas about the relationship between ritual, religion, and the arts. Moreover, the ethological, evolutionary, and cross-cultural perspectives of the artification hypothesis

1 The modifier "proto" indicates that although these alterations of communicative signals can be considered "aesthetic", mothers use them spontaneously and unconsciously with infants, not intentionally as they are later employed in deliberate artification. 2 Certainly parent and infant have competing interests (Trivers 1974) and altruism provides benefits to the self (Hamilton 1964, Maynard Smith 1964). Yet let us not forget that the lives of our Pleistocene ancestors required not only resourceful, competitive individuals but also strongly bonded social groups that could work together with confidence and loyalty. The usual view of humans as selfish, cooperating only so they can advance their own interests, cannot account for the tenacity and resilience of their emotional bonds to one another (Cacioppo & Patrick 2008, Gintis 2000, Rossano 2007, Sober & Wilson 1998, Sterelny 1996, Taylor 1992). 3 The beginning of art(ification) depends on what one considers to be "art". Cupules - patterned, repeated hemispherical indentations in rock surfaces, which I consider to be an early instance of artification - have been minimally dated from as early as 200-150kya in Auditorium Cave in Bhimbetka, India (Bednarik 2003). I E. DISSANAYAKE

broaden and deepen the current cognitivist purview of human aesthetic capacities and behavior by emphasizing their preverbal, presymbolic, pancultural, cross-modal, supra-modal, participative, temporally-organized, affective, and affinitive aspects. The artification hypothesis as presented here is necessarily speculative, but grows out of the synthesis of an array of established findings and well-researched conjectures about early human and hominin behavior.

PART I. THE ARTIFICATION HYPOTHESIS Proto-Aesthetic Dispositions in Infants During the years that cognitive science was becoming consolidated, developmental psychologists began to discover remarkable cognitive abilities in young infants. Before these investigations, common pediatric wisdom held that apart from a few innate reflex behaviors - crying, suckling, clinging, startling - babies were pretty much wax tablets for their elders to inscribe as they (and their cultures) decreed. Today, however, it is well established that newborns come into the world with decided preferences and motivations, so that one can speak intelligendy of "infant semiosis" and even "fetal psychology" (Trevarthen 1994, 1997). We now know that newborns have innate capacities that predispose them not only to solicit physical care but specifically to elicit social and emotional interaction with others. For example, newborns prefer human faces to any other sight and human voices to any other sound. They can imitate face, mouth, and hand movements (Meltzoff & Moore 1977) and respond appropriately to another person's emotional expressions of sadness, fear, and surprise (Field et al. 1982). At birth, infants can estimate and anticipate intervals of time and temporal sequences and remember these temporal patterns, categorizing them both in time and space and in terms of affect and arousal (Jaffe et al. 2001). By six weeks of age, these perceptual and cognitive abilities permit normal infants to engage with adult partners in complex communicative interchanges — the playful behavior that is colloquially called "baby talk". In all cultures,4 human mothers (and other adults) behave differently with infants than with adults or even older children. They use a special vocal register: a high, soft, breathy, singsong tone of voice, which babies prefer to typical adult conversational speech. Vocal contours are much more labile than in ordinary speech directed to other adults (Fernald 1992), and the tempo of speaking is slower (Longhi 2003).

4 Mother-infant interaction has been well studied in a number of different cultures over the past quarter century (Field et al. 1981, Leiderman et al. 19"7"7). Although differences have been found among the practices of different socioeconomic classes and ethnic groups in the United States and Europe and variations observed among these and some Native American and African societies, for the most part the differences arc small, having to do with the proportion of time spent looking, smiling, and vocalizing bv the mother or infant and with the pace and intensity of the encounters. THE ARTI Fl CATION HYPOTHESIS |

Even though they realize that infants cannot understand words, many mothers speak to small babies as if they expect a reply: "You going to tell me a story? Tell me a story then. Tell me a story. Really! Ah, that's a good story! That's a good story, tell me more. Tell me more. Tell me more, yes?" As evident from this example, the utterances are simplified, rhythmic, and highly repetitive. Transcripts such as these reveal poetic features: talk to babies is composed of phrases, each (whether having two or nine syllables) about three and a half to five seconds in length - the temporal length of a poetic line, a musical phrase, and a phrase of speech in adults (Lynch et al. 1995, Turner 1985) - and these phrase segments are combined with others in "stanzas" that comprise a theme with variations. Subtler poetic features of language such as metrics, phonetics, and foregrounding are discernible with further close analysis (Miall & Dissanayake 2002). Despite the fact that in some societies there is no tradition of talking to babies, other rhythmically regular noises such as tongue- clicking, hissing, grunting, or lip-smacking may be used and supplemented by physical movements and exaggerated facial expressions. Even deaf mothers use a "baby talk" version of sign language (simplified, exaggerated, and repetitive), which a deaf baby prefers to adult-oriented signing (Masataka 1996). Along with this special vocal behavior, adults engage and keep infants' attention by the use of rhythmic body movements (touching, patting, stroking), hugging and kissing, unusual facial expressions (uninterrupted gazing into the baby's eyes, sustained smiles, an open mouth, widened eyes, raised eyebrows), and characteristic head movements (backward bobs, nods, and wags). The baby responds to these peculiarly repetitive, exaggerated, and elaborated antics with wriggles, kicks, and smiles. Indeed, it is infants' own positive and negative reactions that persuade caretakers to entertain them with only certain kinds of movements, expressions and sounds, for without an expected infant response (and one's own wish to provoke it) there would be no reason to behave in this unusual and odd way. It is not too much to say that the behavior is dyadic, since infants actively elicit, shape, and otherwise influence the pace, intensity, and variety of signals that adults present to them. Another important and relevant finding is that infants perceive their mother's vocal, visual, and kinetic signals simultaneously as a package (Schore 1994) and in general are capable of cross-modal processing. For example, three- week-olds can perceive the similarity between bright colors and loud sounds (Lewkowicz & Turkewitz 1980). At six months babies recognize that a pulsing tone (heard) and a dotted line (seen) are alike, as are a continuous tone and an unbroken line (Wagner et al. 1981). Thus in early interactions, behaviors are not only direcdy mirrored or imitated but also may be matched by either partner in supramodal qualities such as intensity, contour, duration, or rhythm — qualities that apply to any sense modality. That is to say, the loudness of a sound may be matched by a strong arm or leg movement (or vice versa) or the downward contour of a head movement by a downward fall of the voice (or vice versa) 140 I E. DiSSANAYAKE

(Beebe & Gerstman 1984, Eimas 1984, Marks, Hammeal, & Bomstein 1987, Stern 1985). The temporal dimension is also significant. Microanalyses of videotaped engagements reveal that the ongoing packages of simultaneous sounds, movements, and facial expressions exchanged by the pair are exquisitely coordinated in time.5 Both mother and baby adjust their responses to each other within seconds or fractions of seconds, according to discernible "rules" of mutual regulation that are made up as they go along (Beebe 1986, Beebe, Stern & Jaffe 1979, Beebe et al. 1988, Nadel et al. 1999, Papousek & Papousek 1981, Stern 1971). Over much of the first year of the infant's life, the pair engage and disengage, synchronize and alternate, practicing their physical, physiological, and emotional "attunement" by means of these multimodal expressive signals.

Mother-Infant Interaction is Adaptive Studies show that interaction with caretakers, as described in the preceding section, has practical cognitive and social benefits for infants, such as helping them to practice and perfect the give-and-take of social participation. It helps them to recognize and regulate their own shifting levels of emotion and assists their recognition of prototypical sounds of the language they will eventually speak. By anticipating what comes next in a familiar sequence, the baby "hypothesizes" or predicts when a climax will occur and then experiences its fulfillment. Being able to recognize pattern in the behavior of others - what psychologists call "sequencing" - is essential to eventual social and intellectual competence, making it possible to comprehend and predict others' behavior (Greenspan 1997: 6, 67). I claim that mother-infant interaction is, in fact, a bona fide adaptation that evolutionary psychologists have largely overlooked, despite its ubiquity. In evolutionary theory, an adaptation is an anatomical structure, a physiological process, or a behavior pattern that contributed to ancestral individuals' ability to survive and reproduce in competition with other members of their species (Crawford 1998). I argue that the evolution of mother-infant interaction as described above helped to address an important problem for our Pleistocene ancestors. Evolving hominins came to differ from earlier primates in various ways, including upright walking. Bipedalism required numerous gradual alterations in human skeletal and muscle anatomy, including a shortened and narrowed birth canal. At the same time, human brains were enlarging quite dramatically. These conflicting trends presented a serious problem at parturition and resulted, over millennia, in a gradually reduced gestation period that produced increasingly

5 There may be, of course, "miscoordinations," as when one or the other member of a pair is unresponsive or when a mother is "intrusive," unable to read the baby's signals of overstimu- lation and its wish to disengage (Beebe & Lachmann 1988). THE ARTIFICATION HYPOTHESIS |

immature infants that could pass through the reshaped birth canal.6 Compared to other primates, the resulting hominin babies were unusually helpless, requiring care and vigilance for an unprecedentedly long period of time. It has been calculated that for an infant human to be as mature at birth as an infant chimpanzee, gestation would last for 21 months and the baby would weigh 25 pounds (Gould 1977, Leakey 1994: 44, Portmann 1941). Based on fossil evidence, I suggest that the mother-infant interaction (with its proto-aesthetic operations) was underway in Homo ergaster, which lived approximately 1.7 million years ago,7 or its close relative or descendant H. erectas (Leakey 1994). The universal mother-infant interaction is, I propose, a behavioral adaptation that assured that mothers would be emotionally bonded to their immature offspring and thus willing to provide the necessary extended care over months and years. Such a hypothesis gains credibility when it is pointed out that each vocal, facial, and gestural component of mothers' signals to their babies is drawn from common ordinary vocal, facial, and gestural signals of friendliness and affinity used among adults. Soft, higher-pitched, relaxed voices; smiles, an open mouth, eyebrow-flashes, head bobs and nods; and touches, pats, hugs, and kisses are expressions and behaviors spontaneously and habitually used by individual adults in everyday friendly interaction with each other. Because higher primates also use similar expressions in ordinary sociable circumstances (King 2004), it is not farfetched to assume that our hominin ancestors also used them.

Ritualization Without even thinking about it, mothers make ordinary, everyday affinitive sounds, facial expressions, and head and body movements «tfraordinary. Just how extraordinary becomes clear if we imagine interacting with our friends and associates as we do with babies. Ethologists describe a similar or analogous phenomenon in other animals, especially birds. When ordinary movements and sounds are made conspicuous and distinctive, ethologists call this evolutionary process "ritualization" and the behaviors themselves "ritualized". In this process, head and body movements and/or vocalizations that are ordinarily used in an instrumental context (say, pecking for food, preening, plucking grass for a nest, making a sound that attracts parental attention) become altered in special ways. They are formalized (simplified or stereotyped), repeated, exaggerated, elaborated, and used in a new context to communicate a different, non-ordinary message. In the garganey duck, for example, the casual everyday

6 Other anatomical adaptations developed simultaneously to address this critical problem. Compared to other primates, considerable brain growth of human infants occurs after birth; the skull of human neonates can be compressed during parturition; and parturient human mothers have a separable pubic symphysis. 7 H. ergaster had a narrower pelvis than its predecessor, H. habilis, in which rapid brain expansion was occurring between 3-2mya (Wade 2006: 18-22). "Turkana Lad", a specimen of H. erectos of l.ómya, also has a narrow pelvis (Falk 2009: 51). I E. DISSANAYAKE

behavior of turning the head and cleaning the wing feathers is transformed by these special alterations into a noticeably different formal, regulari2ed and repeated activity in which preening does not occur at all. The head turns back and forth, swinging the bill so that its tip simply touches one spot on the wing, over and over again. In this context, the message is no longer "I'm preening" (as if anyone would care) but, rather, "I want to mate with you". Females have evolved to understand this new, acquired meaning. In mother-infant interaction, the instrumental context for, say, touching is grooming and tending; for smiling it is showing relaxed pleasure. The other signals similarly also communicate friendliness or affiliation. When simplified and regularized, repeated, exaggerated, and elaborated, however, the new message expresses and communicates intense affection, attentive reassurance, and delighted regard far beyond the usual casual contexts with other adults of friendly acknowledgement or companionship. Importandy, these particular ritualized signals of ancestral mothers were adaptive because - having origi- nated in innate affinitive expressions - they reinforced (when formalized, repeated, exaggerated, and elaborated) the neural circuits for affinitive feelings in a mother's own brain through proprioceptive feedback (Scherer & Zentner 2001). An infant who called forth from its mother frequent positive reinforce- ments of loving concern presumably enjoyed greater survival as its mother enjoyed greater reproductive success. Neither of the pair "knew" why they acted and responded as they did. Pleasure was the reward, as it is in other crucial adaptive behaviors relating to sex, food, rest, being liked and admired, and staying warm.

Proto-Aesthetic Operations and Artification The process of ritualization makes specific signals salient. Its operations are all ways of attracting and sustaining attention. Salience — prominence or emphasis of any sort - is potentially emotional. Normally our daily lives are spent in a generalized, unremarkable state of ordinary consciousness in which we do not experience "emotion" so much as what might be described as mood fluctuations, whose eddies are more or less good (positive), bad (negative), or indifferent. Emotion enters (or potentially enters) the scene when there is some discrepancy or change, provoking an interest. We appraise a salient or novel cue, anticipating what it means for our vital interests (Watson & Clark 1994). I suggest that artists in all media deliberately perform the operations described by ethologists as they occur instinctively during a ritualized behavior: they simplify or formalize, repeat (sometimes with variation), exaggerate, and elaborate ordinary materials, bodies, surroundings, tones, beats, body movements, semantics and syntax, motifs, ideas — thereby making these things more than ordinary (Hogan 2003, Jakobson 1971, Mukarovsky 1964/1932, Shklovsky THE ARTIFICATION HYPOTHESIS |

1965/1917).8 By doing so, artists attract attention, sustain interest, and create and mold emotion in their audience (which is what mothers also achieve with babies).9 I suggest that the earliest arts appeared when humans invented ceremonial rituals, where they drew upon already evolved capacities to use (and sensitivities to respond to) proto-aesthetic operations. Ritual ceremonies are meant to affect biologically-important states of affairs that humans necessarily care about — assuring food, safety, health, fertility, prosperity, and so forth. These occasions are times of uncertainty when circumstances can become better or worse (Turner 1969, van Gennep 1960/1909).10 I suggest that uncertainty - leading to emotional investment or "caring about" - was the original motivating impetus for the human invention of religion and its accompanying behavioral expression, which in earlier writings I have called "making special," "making the ordinary extraordinary," and — in recent years — "artificatiori'. The artification hypothesis conceptualizes art differendy from most other schemes - as a behavior ("artifying"), not as the results (paintings, carvings, dances, songs, or poems) or their putative defining qualities (beauty, harmony, complexity, skill). By considering human art as something that people do, it is possible to ask what its adaptiveness might be. At its most basic, artifying is the deliberate use of the proto-aesthetic operations that evolved — as described above — as mechanisms used unconsciously by ancestral human mothers in the highly adaptive context of reinforcing emotional bonds with ever more helpless infants. In these early interactions, we cannot yet speak of arts or even artifying, although the proto-aesthetic operations came to exist as a sort of reservoir of capacities and sensitivities that when used deliberately in a new context (culturally-created ritual ceremonies) again became adaptive. In other words, the proto-aesthetic operations are Tier One of my hypothesis, and the results of artification (music, dance, shaping and embellishing surroundings, poetic language, dramatic performance and the like), based on these operations and developed in humanly-unique and culturally- invented religious ceremonies, are Tier Two.1'

8 The universal aesthetic principles cited by neuroscientists Ramachandran and Hirstein (1999) can be distributed between two of the aesthetic operations described in my scheme as forma- lÌ2ation (grouping, binding, contrast extraction, perceptual problem solving, symmetry) and exaggeration (peak shift, super-normal stimulus, caricature). Three of the four ways in which poets develop poetic metaphor, according to Lakoff and Turner (1989), are also covered by aesthetic operations - i.e., extension (my exaggeration and repetition), elaboration, and questioning (manipulating expectation). 9 As infants mature and require less soothing and more excitement, mothers often spontaneously manipulate their expectations as in games of "Peek-a-Boo" or "This Litde Piggy". Artists also notably manipulate expectation and "surprise" their audiences. 10 Ritualized signals in animals are also often associated with ambivalence (Hinde 1982: 126). 11 Although deliberate artification may be considered an exaptation (side-effect or by-product) of the adaptive mechanisms that promote mother-infant bonding, its ability to relieve anxiety and coordinate individuals in ceremonial contexts was and still is also adaptive. See discussion in the following sections. 144 J E. DISSANAYAKE

Religion, Ritual, and Artification Over the millennia of hominin evolution, the human mind increasingly became a "making sense" organ. Interrelated powers of memory, foresight, and imagination gradually developed and allowed humans to stabilize and confine the stream of life by making mental connections between past, present, and future, or among different experiences or observations. They could remember good and bad things and imagine them happening again. One cost of this growing awareness of the possibilities and unpredictability of life was uncertainty and anxiety. I have suggested that artification first arose in ceremonies that accompanied the invention of religion, another prominent component of human nature that in recent years has been addressed by a number of evolutionary psychologists and cognitive scientists (e.g., Atran 2002, Boyer 2001, Wilson 2002). Although Atran (2002) and Boyer (2001), consider religion to be an evolutionary by- product that parasitizes other adaptive mental processes, Wilson (2002) and others (e.g., Alcorta & Sosis 2006, Richerson & Boyd 1998, Rossano 2007) argue that religion offers fitness advantages through increased group cooperation and solidarity. However, apart from Alcorta and Sosis (2006), none of these scholars seems to recognize the important part played by the arts in instilling cooperation and solidarity. Although anthropologists usually conceptualize a society's "rituals" as part of its symbolic cognitive system, they typically overlook the fact that regardless of the meanings that are conveyed, ceremonial rituals are constituted of arts and would not exist without them. Bodies, surroundings, sounds, movements, and so forth are artified (as described earlier, with aesthetic operations) in order to attract attention, sustain interest, coordinate group effort, and provide emotional excitement and satisfaction. Because of the inseparability of religious practice and artification, it is plausible to suggest that the arts arose in human evolution as components of ceremonial behavior rather than as independently- evolved activities. Through a ceremony's artifications, its messages or meanings were reinforced and the practitioners convinced that they were addressing the matter at hand. Belief in religious dogma may or may not be adaptive, but the vehicles that installed and reinforced the beliefs, the artifications developed and practiced in religious ceremony, could inadvertently become adaptive.12 Whether or not a ceremony achieved its intended result of, say, attracting game or promoting fertility, I submit that it may have provided two general

12 Ritual practice as a concomitant of religion is commonly thought to have arisen as part of the suite of traits made possible by language and symbolizing ability - i.e., ca. 50kya, according to recent conjecture (Wade 2006: 164-165). Yet artifications such as perforated pendants are found from 300kya, ostrich eggshell disc beads from 200kya, and shaped pieces of ochre (to color bodies and objects) from these dates or even earlier (Bednarik 2003); musical/dance behavior leaves no trace and may well also have occurred at these times. I agree with Donald (1991, 2006), who suggests that pre-verbal "mimetic" culture (using gesture, pantomime, dance, visual analogy, and ritual) could have developed in H. erectas, i.e., after 1.8mya. THE ARTIF1CAT10N HYPOTHESIS | adaptive benefits that affected individuals and groups. By artifying in a culturally-sanctioned manner, individuals had "something to do" in uncertain circumstances, giving them a sense of control and thereby relieving anxiety. Additionally, through participation in a temporally-organized activity or performance, a group reinforces its social bonds. Both these outcomes can reasonably be identified as adaptive. In the first instance, it is well known that prolonged emotional stress and anxiety compromise a wide range of bodily functions including energy release, immune system activity, mental activity, digestive function, and growth and tissue repair, as well as reproductive physiology and behavior (Sapolsky 1992). Second, the physical and neurological coordination of participating with others in temporally organized activity has at least three stabilizing effects: satisfying predictability, tension-producing-and- resolving manipulation of expectation, and neurophysiological coordination that is felt subjectively as emotional coordination or integration, which ultimately provides individuals and groups with a sense of coping through social support (Caporael 1997). Several studies report that ritual and artful behavior increased in prehistory at times of resource stress, as in populations of Mimbres (Brody 1977: 210), Late Dorset (Taçon 1983), prehistoric Arnhem Land (Taçon & Brockwell 1995, Taçon, Wilson & Chippindale 1996), and Numic-speaking peoples (Garfinkel, Marcom & Schiffiman 2003).

PART II: SOME IMPLICATIONS OF THE ARTIFICATION HYPOTHESIS FOR COGNITIVE, EVOLUTIONARY, AND NEURO-AESTHETICS (CENA)13 The artification hypothesis claims that an accurate understanding of human artmaking and participation is both broader and deeper than the prevalent view of aesthetic cognition and behavior as described by CENA which, to begin with, is suffused with modern Western notions about art. For example, visual arts or pictured (or real) landscapes are typically treated as if they were static entities perceived by static individuals. Researchers in evolutionary aesthetics and neuroaesthetics assume that art can be characterized by (or even be considered synonymous with) beauty (Dutton 2009, Martindale, Locher, & Petrov 2006, Thornhill 2003), talent (Dutton 2009), originality (Miller 2000), creativity (Martindale, Locher, & Petrov 2006, Miller 2000, Dutton 2009) and rarity (Dutton 2009). But ethnologists' reports from premodern societies indicate that stationary arts and contemplative observers are rare. On the contrary, nearly everyone actively makes or participates in the arts, tradition is frequently valued over novelty and creativity, and fearsomeness or shock may trump beauty. For its part, cognitive studies of music investigate the perception and performance of fragments of musical behavior such as awareness of pitch,

13 Because evolutionär)' aesthetics and neuroaesthetics along with cognitive studies in general share some common ideas about the arts, I use an acronym, CENA, when referring to all three fields as a group. 146 I E. DISSANAYAKE

rhythm, intervals, and other measurable elements of Western tonal music. Illustrative examples of music (like neuroaesthetics' studies of visual arts) are frequendy Western masterpieces, which in important ways are unlike the sources for aesthetic engagement in premodern or Pleistocene peoples or in small children. Neuroaesthetics treats visual arts as packages of perceptual stimuli, even though these same stimuli also occur in nonaesthetic contexts and it is not clear whether or how the two can be distinguished (Brown & Dissanayake 2009). Evolutionary aesthetics and neuroaesthetics studies are typically concerned with perceptual and cognitive preferences for features that were or are adaptive - e.g., associated with salubrious landscapes, healthy mates, nutritious food, cognitively-satisfying forms, or works - e.g., pictorial or literary depictions of adaptively relevant subjects such as romance or resolved conflict (Barash & Barash 2005, Martindale, Locher & Petrov 2007, Voland & Grammer 2003). Such studies are worthwhile, but rarely considered are aesthetic capacities or mechanisms — the behavioral and emotional means and manipulations by which features or works have their effects. The arts do not have their effects (adaptive or otherwise) simply because they activate cognitive/perceptual modules that direct us to good mates, or because they contain the color red that connotes biologically salient stimuli such as blood or ripe berries. If the mere stimulus were sufficient (say, a pornographic image or a gushing wound), there would be no need at all to embed these categories or features in art works or events in which they are arranged with relation to other stimuli and otherwise manipulated. It is the manipulations — what is done, the operations or means to the end of art — that produce their emotional effects. And it is in the temporal arts — i.e., those that take place in time — that one can best appreciate how emotion can be manipulated to expressive and eventually adaptive ends. Whatever else aesthetic cognition may consist in, it should include consideration of aesthetic manipulations. Although the artification hypothesis has been gradually developing along with the maturation of cognitive science, its emphasis on ethology — considering art and aesthetics as adaptive behaviors with their roots in early hominins - suggests important and relevant observations that CENA have dismissed or overlooked.

1. Humans are animals: ethological and evolutionary thinking are important to understanding human cognition and behavior It might seem surprising that human infants are born with the rudiments of aesthetic cognition — that they are ready to respond to proto-aesthetic manipulations of visual, vocal, and gestural signals. Yet, from the viewpoint of ethology, it is no surprise at all: many animals also respond to formalization, repetition, exaggeration, elaboration, and manipulation of expectation. One might say that aesthetic cognition, based in proto-aesthetic operations that THE ARTIFICATION HYPOTHESIS | 147 evolved in H. ergaster mothers and infants, is a pivotal behavior between primate ancestors and modern humans, who are unique among other animals in deliberately performing aesthetic operations with the intention of making the ordinary extra-ordinary and consequendy having an emotional effect on other people, including ancestors and spirits. Ethology is perhaps more associated today with field naturalists who carry binoculars and a notebook than with the CENA crowd who may work with expensive brain imaging machinery. For understanding aesthetic cognition and behavior, however, the "whole animal approach" of ethology reminds us of another antecedent of artification: the behavior of play (Deacon 1997, 2006, Dissanayake 1974). In many juvenile animals, including hominins (presumably) and humans, play creates a special world in which such actions as play-stalking, play-fighting, and even play-courtship are acknowledged as being "not for real". As in ritualized behaviors, another ethological subject, participants in play realize that altering ordinary behavior in a new context bestows a new or special meaning. Ethology has been described as the naturalistic study of behavior from an evolutionary perspective (Burghardt 2005). Full appreciation of the fact that all cognition and behavior necessarily have their roots in the evolutionary past seems perplexingly and inexcusably absent in much that goes on under the cognitive canopy. Although they are aware that minds are composed of neural matter, all too many cognitive psychologists - like their academic counterparts in the humanities - still operate as though minds are detached from evolutionary reality (Slingerland 2008: 95).14 Surprisingly, this seems to be the case with embodiment philosophers who accept that minds arise from our physical being and are "embodied" or "metaphorical" yet do not appear to wonder why that sort of mentation should have evolved to characterize human cognition. For example, why should such basic-level cognitive structures as "paths" and "containers" appear as paths and containers to cognitively normal human beings (Slingerland 2008: 23)? We know that minds, like the output of other organs, work as they do because that way of working (and not some other way) contributed to survival and reproductive success. What is or was that contribution? One can approach this question with knowledge of our Pleistocene past in which humans were motivated to desire the same primal requirements that face any wild, group-living animal: sustenance (food, water, warmth); social acceptance and participation; sex and mutuality; care of offspring; safety

14 Brandt (2006: 184) is an exception. He acknowledges the "macroscopic" scale of evolutionary time (the "why?"), along with the microscopic scale of neurology (the "how?"), and the mesoscopic (historical) scale of our own space and time (the "what?") and recommends that the levels and questions be and remain in contact. Among cognidvists, the work of Merlin Donald (1991, 2006) and Terrence Deacon (1997, 2006) is noteworthy in finding the origins of aesthetic cognition and behavior in the evolutionary and even animal past - in preliterate and even preverbal hominins. I E. DISSANAYAKE

(avoidance of pain and death); and competence (knowing what to do to survive - to achieve one's goals and satisfy one's needs) (Panksepp 1998). Paths, containers, and other metaphors of embodied cognition have evolutionary as well as "bodily" meanings that have no doubt influenced their pervasiveness in language and thought.15

2. Preverbal affective and aesthetic mechanisms continue to influence human language and cognition; they are a critical, if neglected, component of CENA There are more than a few cognitivists who consider minds only as they exist in modern human adult individuals — that is, people who are verbal, rational, and tacitly literate. But such individuals are a miniscule proportion of humans, past and present, who think and have minds; such an emphasis may well exclude preverbal infants and preliterate children, nonliterate peoples in small-scale societies, and Pleistocene ancestors. According to the artification hypothesis, the precursor to aesthetic cognition began in hominin mother-infant interactions as early as 1.7 mya. These ancestors made tools and lived in small hunter- gatherer groups but did not speak — though like apes they could have made expressive vocalizations. By 200 to 150kya, Archaic hominins were in what is now India, carving cupules and quite likely singing and dancing, although they may also not have yet had much if any language. Although some have claimed that the selective pressure for the language faculty was the adaptive value of communicating complex propositions (e.g., Pinker & Jackendoff 2005: 204),16 this view of things seems too narrow if one takes into account the paralinguistic or expressive "half' of speech. It seems worth considering that, as Panksepp (1998: 302) has said, "the brain mechanisms for language were designed for social interactions, not for the conduct of science". Certainly baby talk, as described, has nothing to do with the exchange of verbal information about the world and everything to do with coordinating individuals and sharing social experience, which has obvious adaptive value. After living with nonliterate Trobriand Islanders nearly a century ago, Mali- nowski (1925) was moved to suggest that language serves not to imitate thought but to move another to act; like artification, it is intended "to produce

15 See Pinker (2007: 235-278) for an extensive discussion of reasons that the "embodied cognition" or "metaphor" theory is not biologically fundamental in the ways that its proponents presume. 16 On p. 231 of the same paper, Pinker and Jackendoff also contend that language (not "the language faculty") "is an adaptation for the communication of knowledge and intentions" and that "as a species humans rely on acquired technology, know-how and extensive cooperation among non-kin". This way of putting it is more acceptable than the "complex propositions" cited in the first paragraph of their paper (p. 204). It leaves open the possibility of accepting the evolutionary importance of artification (arising from the evolved human ability, originating in ancestral mother-infant interactions, to interact contingendy and to entrain emotional state) insofar as it contributes to socialization and cooperation. THE ARTIFICATION HYPOTHESIS |

an effect"17 (see also Donald 2006). Malinowski's contemporary, British social anthropologist A. R. Radcliffe-Brown (1948/1922: 234), proposed that ceremonies (composed, as described earlier, of arts, including verbal arts) serve to maintain and transmit from one generation to another the emotional dispositions - not only or even primarily the complex propositions - on which a society depends for its existence. With some exceptions, practitioners of CENA tend not only to emphasize language and ignore preverbal or nonverbal experience but because emotions are viewed as proximate mechanisms designed to address adaptive problems by motivating behavior (Tooby & Cosmides 1992: 99), their varieties, subdeties, complexities, and importance are typically neglected. Affective neuroscience, as described by Jaak Panksepp (1998), is therefore requisite to an understanding of human thought and behavior and, particularly, to understanding aesthetic cognition and behavior which, in their earliest manifestations, had to do with the artifications of ceremonial practices that were intended to help people obtain the life needs that they had evolved to care about. Important in the evolution of proto-aesthetic operations is their temporal organisation and coordination, which makes possible the emotionally satisfying, tension-resolving, and unifying effects described at the end of Part I. Infants are supremely sensitive to repetition and variation in the pulses, phrases, and narrative episodes of their mothers' babytalk (Gratier & Apter-Danon 2009). This suggests that both the creation and the experience of the temporal arts of poetry, music, and dance inhere in our fundamental psychobiology. Malloch and Trevarthen (2009: 8) claim that human action, thought, and feeling take place under the coordinated and integrated control of a time keeping, energy regulating Intrinsic Motive Pulse (IMP) and that the brain is a network of dynamic systems all obedient to a scale of rhythms that flow in unison - an "innate communicative musicality". Such a formulation makes clear that much is overlooked when aesthetic cognition is conceptualized (as by some CENA theorists) simply as "sensory" or "perceptual" or primarily based in language or verbal thought.

3. The bio-ecological requirements of several million years of hunter-gatherer life have affected human cognition In his study of thought in individuals who live in small-scale societies, anthropologist C. R. Hallpike (1979) has demonstrated that with regard to the development of cognitive processes such as symbolization, classification, number, measurement, conservation, space, time, and causality, premodern people do not require what Piaget (1970) called "concrete operations" and that for most purposes "pre-operatory" thought is just fine. Although they may

17 Such a view corresponds with the categories of infants' first words, which at thirteen months have to do with parents, requests and refusals, and greetings, in that order (Hauser 1996: 338). I E. DISSANAYAKE

have little need to "communicate complex propositions," they have a myriad of adaptive abilities that use other kinds of thought. Howard Gardner's theory of "multiple intelligences" is also relevant here (Gardner 1983, 1999).18 He suggests that individuals possess varying degrees of eight broad cognitive capacities: linguistic, logical-mathematical, spatial, bodily-kinesthetic, musical, interpersonal, intrapersonal, and naturalist. An adequate description of the components of these capacities and their interrelationship requires more nuance than is possible here. However, it suffices to say that the intelligences described by Gardner have definable neural substrates and would have emerged during human evolution. It is not difficult to see that they would all contribute to hunter-gatherer lives. All of them, according to Gardner, can be used in the making and experiencing of art works. All require "thinking" (cognition), even though in most of the intelligences, verbal or formal operational (prepositional) thought would not be required. In small-scale societies of the recent past, propositional thought may not have been a priority but artification was central, as in the Yirkalla, an Aboriginal population of Arnhem Land in Australia with a total population in the 1950s of about four hundred. According to Richard Waterman

[T]he arts [...] are unified to an unusual degree, in that the subject matter of all of them - graphic, plastic, folkloric, choreographic, dramatic, and musical — is largely confined to activities of totemic species and totemic ancestors. Within the largest musical category almost every song has a painted iconographie design or hard-wood carving, a story, a dance, and a segment of ritual associated with it (Waterman 1955: 40).19

A cynic might say that the Yirkalla use of language - artified and performed in a multi-media setting — was obviously not adaptive since the group could not compete with European settlers who, in addition to their greater technological prowess, were able to express and understand complex propositions. Yet in Australia such groups persisted for at least forty thousand years, whereas complex propositional thought would not have been used until the invention of writing some six thousand years ago (and may yet turn out to have been an adaptive dead-end).

18 The influential theories of Piaget and Gardner have been challenged and are not accepted in all their details by all psychologists; my use of them is general and sen es my purposes here. This is not the place to defend particular aspects of their complexities. 19 Waterman's description illustrates or implies several claims made in this article about aesthetic operations (cognition?) in the majority of societies over evolutionary time: the arts are typically presented and experienced in a complex multi-media manner; people make art about things they care about; participation (rather than passive contemplation) is prominent; the arts are intended to produce an effect; and the arts are collective and social, helping their participants to share mental representations and create culture (see also Donald 2006: 14). THE ARTIFICATION HYPOTHESIS |

4. Artification and aesthetic cognition are adaptive in their own right, not by-products of symbolizing or other cognitive ability; evolutionarily, they have been as adaptively important to successful human lives as the verbal transfer of information Although some animals and a number of birds perform art-like behaviors — and, like human infants, respond to formalizing, repeating, exaggerating, and elaborating — humans as a species are unique in that, without being constrained by season or gender or genetic program, they deliberately use these operations to make their bodies, vocalizations, movements, and surroundings extraordinary or special. I consider this phenomenon of artifying to be in the same range of importance to human evolution as tool making, speech, or the making and using of symbols. It is a grave oversight that such a crucial human activity has been inadequately recognized by those who are concerned with human cognitive abilities.20 To my knowledge no advocate of CENA has considered aesthetic cognition (as artification) to be foundational or even important in human mentation.21 Prehistorians study material artifacts and visual images, so it is perhaps not surprising that they generally assume that human art, like speech, is a consequence or offshoot of symbolizing ability.22 Evolutionary aesthetics and neuroaesthetics are concerned with preferences for perceptual signals (signs) of biological reward or cognitive coherence. In considering art to be a behavior — something that people do (not its product as an artifact or image) - the artification hypothesis challenges these assumptions. If one considers the temporal arts, it is clear that not all arts are symbolic - for instance vocalizing or playing a musical instrument, marking beats, or dancing. Their effects are emotional more than cognitive-symbolic: they attract attention, sustain interest, and create and mold emotion. Visual marks need not automatically be assumed to be representational, as in the earliest drawings of children and, arguably, the earliest rock markings of our ancestors. On the contrary, these are traces of marking as an activity in its own right, having an

20 Brandt (2006: 176), insofar as he posits a similar "aesthetic register" of the mind, is an exception, although he seems to be concerned more with cultural than biological evoludon. He characterizes art as "any expressive or instrumental doing that deliberately creates excess structure" (and uses the words "special", "extraordinary" and "elaborated"). 21 Although the tide, The Artful Mind (Turner 2006) might suggest otherwise, the subtide (iCognitive Science and the Riddle of Human Creativity) refers to creadvity, which can be applied to a multitude of activities, not specifically to a behavior of ratification. 22 As Nicholas Wade remarks, "Archaeologists [...] tend to equate full-fledged modern language with art, which only becomes common in the archaeological record some 45,000 years ago. Their argument is that creadon of art implies symbolic thinking in the mind of the artist, and therefore possession of language to share these abstract ideas" (Wade 2006: 46; see also Baiter 2009). This is so of course only if one takes for granted that "art" is representational and visually perceived, as with the famed Upper Palaeolithic rupestrian paintings and mobi- Liary artifacts. For stimulating counter-ideas about the earliest appearance of human palaeoart and language, see Bednarik (2003). I E. DISSANAYAKE

effect on the world and making the ordinary surface extraordinary. Aesthetic operations of regularizing or formalizing, repeating, exaggerating, and elaborating these marks are additionally interesting and satisfying, even when they are not symbolic. But even symbolic marks can produce a variety of effects that have little or nothing to do with their intended symbols or referents. The complex red, black, and white designs on prows of ocean-going ceremonial canoes (,masawa) of Trobriand Islanders are meant to dazzle those who see them so that they will be helplessly persuaded to trade their most valuable shells in the kttla exchange (Campbell 2002). The Trobrianders know the symbolism of these designs — stylized animals and birds that in various ways represent power. Their trading partners do not know but the optical dazzle in itself is intended to disorient them. Proponents of CENA would note that masawa designs stimulate perceivers' neuro-optic apparatus. They could suggest adaptive antecedents for the choice of curvilinear shapes and bold colors. But their theoretical emphasis on perception and signals obscures important emotional aspects that are the motivation for artifying in the first place, as well as the emotional responses that artifications produce apart from their representational or symbolic meaning. Additionally, adult perceivers, like the infants they once were, respond emotionally to "vitality effects" - supra-modal qualities of intensity, shape or contour, motion, duration, and rhythm that occur in both space and time and help to manipulate expectation (Stern 1985, 1999). Such effects as surging, fading away, fleeting, being drawn out, or being shaken to the core (dazzled) each apply to a variety of circumstances in visual, auditory, and kinetic modalities.23 Additionally, they have associations with various colors and shapes that spread out into modalities other than visual. If the expressive forms and media of art are intended to influence the minds of an audience (Donald 2006: 4, Malinowski 1925), artifications give perceptions emotional oomph, thereby bestowing conviction on the beliefs that are the occasions for ritual ceremonies. Rather than being symbolic, the vitality effects described by Stern (1999), along with aesthetic operations as described here, are examples of anabgy (Donald 1991, 2006).24 They have biological meanings of fast and slow, loud and soft, large and small, extravagant and dainty, careless and precise, which in some cases even other animals can recognize. Prosody and gesture are also

23 In this regard, one can remind advocates of embodied or narrative thinking that people find paths to be thorny, obscure, winding, or welcoming; containers can be capacious, constrict- ing, smooth, or crowded. 24 Merlin Donald's (1991, 2006) views of mimesis and analog)· arc rich and complex; they would be augmented by incorporation of the artification hypothesis and they complement my discussion here. THE ARTIFICATION HYPOTHESIS |

analogous — both may have preceded speech by 2 million years when brains expanded rapidly (Corballis 2002).25 In summary, the etiological, evolutionary, and cross-cultural perspectives of the Artification Hypothesis challenge some current CENA pronouncements by emphasizing the importance of preverbal, presymbolic, pancultural, cross- modal, supra-modal, participative, temporally organized, affective, and affiliative aspects of human art-making and response. Additionally, the hypothesis proposes that artification is a universal behavioral predisposition whose components are ancient and influential — if unregarded — features of human cognition.

References Alcorta, C. S. & Sosis, R. (2005). Ritual, Emotion and Sacred Symbols: The Evolution of Religion as an Adaptive Complex. Human Nature 16, 323-359. Atran, S. (2002). In Gods We Trust. Oxford: Oxford University Press. Balter, M. (2009). On the Origin of Art and Symbolism. Science 323 (Feb. 6), 709-711. Barash, D. P. & Barash, N. R. (2005). Madame Bovary's Ovaries: A Darwinian Look at Literature. New York: Delacorte Press. Barsalou, L. (1999a). Perceptual Symbol Systems. Behavioral and Brain Sciences 22, 577— 609. Barsalou, L. (1999b). Perceptions of Perceptual Symbols (author's response to commentary). Behavioral and Brain Sciences 22, 633—660. Bednarik, R. G. (2003). The Earliest Evidence of Palaeoart. Rock Art Research 20 (2), 89-135. Beebe, B. (1986). Mother-Infant Mutual Influence and Precursors of Self- and Object Representations. Inj. Masling (Ed.), Empirical Studies of Psychoanalytic Theories, Vol, II (pp. 27-48). Hillsdale, NJ: Erlbaum. Beebe, B. & Gerstman, L. (1984). A Method of Defining "Packages" of Maternal Stimulation and Their Functional Significance for the Infant with Mother and Stranger. International journal of Behavioral Development 7,423—440. Beebe, B. & Lachmann, F. (1988). The Contribution of Mother-Infant Mutual Influence to the Origins of Self- and Object Representations. Psychoanalytic Psychology 5 (4), 305-337. Beebe, B., Stern, D., & Jaffe, J. (1979). The Kinesic Rhythm of Mother-Infant Interactions. In A. W. Siegman & S. Feldstein (Eds.), Of Speech and Time: Temporal Patterns in Interpersonal Contexts (pp. 23—34). Hillsdale, NJ: Erlbaum. Beebe, B., Alson, D., Jaffe, J., Feldstein, S. & Crown, C. (1988). Vocal Congruence in Mother-Infant Play. Journal of Psjchotinguistic Research 17 (3), 245-259. Boyer, P. (2001). Religion Explained. New York: Basic Books. Brandt, P. A. (2006). Form and Meaning in Art. In M. Turner (Ed.), The Artful Miné Cognitive Science and the Riddle of Human Creativity (pp. 171-188). Oxford: Oxford University Press.

25 Lawrence Barsalou (1999a, b) has proposed that analog images are the sole medium of human thought. I E. DISSANAYAKE

Brody, J. J. (1977). Mimbres Painted Pottery. Albuquerque: University of New Mexico Press. Brown, S. & Dissanayake E. (2009).The Arts are More than Aesthetics: Neuroaesthetics as Narrow Aesthetics. In M. Skov & O. Varshanian (Eds.), Neuroaesthetics (pp. 43— 57). Amityville, NY: Baywood. Burghardt, G. M. (2008). Evolution and Paradigms in the Study of Literature. Style 42 (3,4), 144-149. Cacioppo, T. & Patrick, W. (2008). Loneliness: Human Nature and the Need for Sodai Connection. Campbell, S. F. (2002). TheArtofKula. Oxford and New York: Berg. Caporael, L. R. (1997). The Evolution of Truly Social Cognition: The Core Configura- tion Model. Personality and Social Psychology Review 1, 276-298. Corballis, M. C. (2002). From Hand to Mouth: The Origins of Language. Princeton: Princeton University Press. Crawford, C. (1998). The Theory of Evolution in the Study of Human Behavior An Introduction and Overview. In C. Crawford & D. L. Krebs (Eds.), Handbook of Evolutionary Psycholog/: Ideas, Issues, and Applications (pp. 3-41). Mahwah, NJ: Erlbaum. Deacon, T. (1997). The Symbolic Species: The Co-Evolution of language and the Human Brain. London: Allen Lane. Deacon, T. (2006). The Aesthetic Faculty. In M. Turner (Ed.), The Artful Mind: Cognitive Science and the Riddle of Human Creativity (pp. 21-53). Oxford: Oxford University Press. Dissanayake, E. (1974). An Hypothesis of the Evolution of Art from Play. Leonardo 7 (3), 27-31. Dissanayake, E. (1979). An Ethological View of Ritual and Art in Human Evolutionary History. Leonardo 12 (1), 27-31. Dissanayake, E. (1984). Does Art Have Selective Value? Empirical Studies of the Arts 2 (1), 35-49. Dissanayake, E. (1990). Of Transcribing and Superliteracy. The World and I, October, 575-587. Dissanayake, E. (2000). Art and Intimacy: How the Arts Began. Seattle: University of Washington Press. Dissanayake, E. (2008). If Music is the Food of Love, What About Survival and Reproducrive Success? Musica Scientia Special Issue, 169-195. Dissanayake, E. (2009). Root, Leaf, Blossom, or Bole: In S. Malloch & C. Trevarthen, (Eds.), Communicative Mumality: Exploring the Basis of Human Companionship (pp. 17— 30). Oxford: Oxford University Press. Donald, M. (1991). Origins of the Modem Mind: Three Stages in the Evolution of Culture and Cognition. Harvard: Harvard University Press. Donald, M. (2006). Art and Cognitive Evolution. In M. Turner (Ed.), The Artful Mind: Cognitive Science and the Kiddle of Human Creativity (pp. 3-20). Oxford: Oxford Univer- sity Press. Dutton, D. (2009). The Art Instinct: Beauty, Pleasure, and Human Evolution. New York: Bloomsbury. Eimas, P. D. (1984). Infant Competence and the Acquisition of Language. In D. Caplan, A. Roch Lecours & A. Smith (Eds.). Biological Perspectives on Language (pp. 109-129). Cambridge, MA: MIT Press. Falk, D. (2009). Finding Our Tongues: Mothers, Infants & The Origins of Language. New York: Basic Books. THE ARTI Fl CATION HYPOTHESIS |

Fernald, A. (1992). Maternal Vocalizations to Infants as Biologically Relevant Signals: An Evolutionary Perspective. Inj. H. Barkow, L. Cosmides, & J. Tooby (Eds.), The Adapted Mind: Evolutionary Psycholog and the Generation of Culture (pp. 391—428). New York: Oxford University Press. Field, T. M., Sostek, A. M., Vietze, P., & Leiderman, P. H. (Eds.) (1981). Culture and Early Interactions. Hillsdale, NJ: Erlbaum. Field, T. M., Woodson, R., Greenberg, K, & Cohen, D. (1982). Discrimination and Imitation of Facial Expressions by Neonates. Science 218,179—181. Gardner, H. (1983). Frames of Mind: The Theory of Multiple Intelligences. New York: Basic Books. Gardner, H. (1999). Intelligence Reframed: Multiple Intelligences for the 21st Century. New York: Basic Books. Garfinkel, A.P., Marcom, G., & Schiffman, R A. (2003). Culture Crisis and Rock Art Intensification: Numic Ghost Dance Paintings and Coso Representational Petro- glyphs. In D. Christensen & P. Whitehead (Eds.), American Indian Rock Art 33, 83- 103. Tucson, AZ: American Rock Art Research Association. Gintis, H. (2000). Strong Reciprocity and Human Sociality, journal of Theoretical Biology 206,169-179. Grader, M. & Apter-Danon, G. (2009). The Improvised Musicality of Belonging: Repetition and Variation in Mother-Infant Vocal Interaction. In S. Malloch & C. Trevarthen (Eds.), Communicative Musicality: Exploring the Basis of Human Companionship (301-327). Oxford: Oxford University Press. Greenspan. S. (1997). The Growth of the Mind and the Endangered Origins of Intelligence. Reading, MA: Addison-Wesley. Gould, S.J. (1977). Ontogeny and Phytogeny. Cambridge, MA: Harvard University Press. Hallpike, C. R. (1979). The Foundations of Primitive Thought. Oxford: Clarendon. Hamilton, W. (1964). The Genetical Evolution of Social Behaviour. Parts I, II. Journal of Theoretical Biology 7, 1-52. Hauser, M. D. (1996). The Evolution of Communication. Cambridge, MA: MIT Press. Hinde, R. A. (1982). Ethology: Its Nature and Relations With Other Sciences. Oxford: Oxford University Press. Jakobson, R. (1971). The Dominant. In L. Matajka & K. Pomorska (Eds. & Trans.), Readings in Russian Poetics: Formalist and Structuralist Views (pp. 82—93). Cambridge, MA: MIT Press. Jaffe, J., Beebe, B., Feldstein, S., Crown, C. L., & Jasnow, M. D. (2001). Rhythms of Dialogue in Infancy. Monographs of the Society for Research in Child Development 265, 66 (2), 1-151. King, B. J. (2004). The Dynamic Dance: Nonvocal Communication in African Great Apes. Cambridge, MA: Harvard University Press. Lakoff, G. & Johnson, M. (1999). Philosophy in the Flesh: The Embodied Mind and Its Challenge to Western Thought. New York: Basic Books. Lakoff, G. & Turner, M. (1989). More Than Cool Reason: A Field Guide to Poetic Metaphor. Chicago: University of Chicago Press. Leakey, R. (1994). The Origin of Humankind. New York: Basic Books. Leiderman, P. H., Tulkin, S. R. & Rosenfeld, Α., eds. (1977). Culture and Infancy. New York: Academic Press. Lewkowicz, D. J. & Turkewitz, G. (1980). Cross-Modal Equivalence in Early Infancy: Auditory-Visual Intensity Matching. Developmental Psychology 16 (6), 597-607. I E. DISSANAYAKE

Longhi, E. (2003). The Temporal Structure of Mother-Infant Interactions in Musical Contexts. Ph.D. thesis, University of Edinburgh. Lynch, M. P., Kimbrough-Oller, D., Steffens, M. L. & Buder, E. H. (1995). Phrasing in Pre-Linguistic Vocalizations. DevelopmentalPsychobiology 28 (1), 3—25. Malinowski, B. (1925). The Problem of Meaning in Primitive Languages. In C. K. Ogden & I. A. Richards (Eds.), The Meaning of Meaning: A Study of the Influence of Language Upon Thought and of the Science of Symbolism (pp. 451-510). New York: Har- court Brace. Malloch, S. & Trevarthen, C. (Eds.) (2009). Communicative Musicality: Exploring the Basis of Human Companionship. Oxford: Oxford University Press. Marks, L. E., HammeaL, R. J., & Bornstein, M. H. (1987). Perceiving Similarity and Comprehending Metaphor. Monographs of the Society for Research in Child Development 52 (1), 1-92. Martindale, C., Locher, P., & Petrov, V.M. (Eds.) (2006). Evolutionary and Neurocognitive Approaches to Aesthetics, Creativity and the Arts. Amityville, NY: Baywood. Masataka, N. (1996). Perception of Motherese in a Signed Language by 6-Month-Old Deaf Infants. Developmental Psycholog 32 (5), 874-879. Maynard Smith, J. (1964). Group Selection and Kin Selection. Nature 201,1145-1147. Meltzoff A. N. & Moore, M. H. (1977). Imitation of Facial and Manual Gestures by Human Neonates. Science 219,1347-1349. Miall, D. & Dissanayake, E. (2002). The Poetics of Babytalk. Human Nature 14 (4), 337- 364. Miller, G.F. (2000). The Mating Mind: How Sexual Choice Shaped the Evolution of Human Nature. New York: Doubleday. Mukarovsky, J. (1964 [1932]). Standard Language and Poetic Language. In P. L. Garvin (Ed.), A Prague School Reader on Esthetics, Literaiy Structure, and Style (pp. 17—30). Washington, DC: Georgetown University Press. Nadel, J., Carchon, I., Kervella, C., Marcelli, D., & Réserbet-Planty, D. (1999). Expectancies for Social Contingency in 2-Month-Olds. Developmental Science 2, 164— 173. Panksepp, J. (1998). Affective Neurosdence: The Foundations of Human and Animal Emotions. Oxford: Oxford University Press. Papousek, M. & Papousek, H. (1981). Musical Elements in the Infant's Vocalization: Their Significance for Communication, Cognition, and Creativity. In L. P. Lipsitt & C. K. Rovee-Collier (Eds.), Advances In Infancy Research 1 (pp. 163-224). Norwood NJ: Ablex. Piaget, J. (1970). Genetic Epistemologa. New York: Columbia University Press. Pinker, S. (1997). How the Mind Works. New York: Norton. Pinker, S. (2007). The Stiff of Thought: Language as a Window into Human Nature. New York: Viking. Pinker S. & Jackendoff, R. (2005). The Faculty of Language: What's Special About It? Cognition 95 (2), 201. Portmann, A. (1941). Die Tragzeit der Primaten und die Dauer der Schwangerschaft beim Menschen: Ein Problem der vergleichenden Biologie. Revue Suisse de Zoologie 48,511-518. Ramachandran, V. S. & Hirstein, W. (1999). The Science of Art. journal of Consciousness Studies 6, 15-51. THE ARTIFICATION HYPOTHESIS |

Richerson, P. & Boyd, R. (1998). The Evolution of Human Ultra-Sociality. In I. Eibl- Eibesfeldt & F. K. Salter (Eds.), Indoctrinahility, Ideology and Warfare: Evolutionaiy Perspectives (pp. 71-95). New York: Berghahn. Rossano, M. (2007). Supernaturalizing Social life: Religion and the Evolution of Human Cooperation. Human Nature 18, 272-294. Sapolsky, R. M. (1992). Neuroendocrinology of the Stress Response. Inj. R. Becker, S. M. Breedlove & D. Crews (Eds.), Behavioral Endocrinology (pp. 287-324). Cambridge ΜΑ: ΜΓΓ Press. Scherer, K. R. & Zentner, M. R. (2001). Emotional Effects of Music: Production Rules. In P. N. Juslin & J. A. Sloboda (Eds.), Music and Emotion: Theory and Research (pp. 361-392). Oxford: Oxford University Press. Schore, A. N. (1994). Affect Regulation and the Origin of the Self: The Neurobiology of Emotional Development. Hillsdale NJ: Erlbaum. Shklovsky, V. (1965 [1917]). Art as Technique. In L. T. Lemon & M. J. Reis (Eds. & Trans.), Russian Formalist Criticism: Four Essays (pp. 3-24). Lincoln: University of Nebraska Press. Skov, M. & Vartanian, O. (Eds.) (2009). Neuroaesthetics. Amityville NY: Baywood. Slingerland, E. (2008). What Science Offers the Humanities: Integrating Body and Culture. Cambridge: Cambridge University Press. Sober, E. & Wilson, D. S. (1998). Onto Others: The Evolution and Psychology of Unselfish Behaviors. Cambridge, MA: Harvard University Press. Sterelny, K. (1996). The Return of the Group. Philosophy of Science 63, 562-584. Stern, D. (1971). A Microanalysis of Mother-Infant Interaction. Journal of the American Academy of Child Prychiatiy 10, 501-517. Stem, D. (1985) The Interpersonal World of the Infant. New York: Basic Books. Stern, D. (1999). Vitality Contours: The Temporal Contour of Feelings as a Basic Unit for Construction the Infant's Social Experience. In P. Rochat (Ed.), Early Social Cognition: Understanding Others in the First Months of Life (pp. 67—90). Mahwah, NJ: Erlbaum. Taçon, P. S. C. (1983). Dorset Art in Relation to Prehistoric Culture Stress. Etudes Inuit/InuitStudies 1 (1), 41-65. Taçon, P. S. C. & Brockwell, S. (1995). Arnhem Land Prehistory in Landscape, Stone, and Paint Antiquity 69, 676-695. Taçon, P. S. C., Wilson, M., & Chippindale, C. (1996). Birth of the Rainbow Serpent in Arnhem Land Rock Art and Oral History. Archaeology Oceania 31,103-124. Taylor, S. E. (1992). The Tending Instinct: How Nurturing is Essential for Who We Are and How We Live. New York: Times Books. Thornhill, R. (2003). Darwinian Aesthetics Informs Traditional Aesthetics. In E. Voland & K. Grammer (Eds.), Evolutionary Aesthetics (pp. 9-35). Berlin: Springer. Tooby, J. & Cosmides, L. (1992). The Psychological Foundations of Culture. In J. Barkow, L. Cosmides & J. Tooby (Eds.), The Adapted Mind: Evolutionary Psycholog and the Generation of Culture (pp. 19-136). Oxford and New York: Oxford University Press. Trevarthen, C. (1994). Infant Semiosis. In W. Noth (Ed.), Origins of Semiosis (pp. 219- 252). Berlin: Mouton de Gruyter. Trevarthen, C. (1997). Fetal and Neonatal Psychology: Intrinsic Motives and Learning Behavior. In F. Cockburn (Ed.), Advances in Perinatal Medicine (pp. 282—291). New York: Parthenon Publishing Group. Trivers, R. (1974). Parent-Offspring Conflict American Zooìogst 14, 249-264. I E. DISSANAYAKE

Turner, F. (1985). The Neural Lyre: Poetic Meter, the Brain, and Time. In F. Turner (Ed.), Naturai Classicism: Essays on Literature and Science (pp. 61-108). Charlottesville: University Press of Virginia. Turner, M. (2006). The Artful Mind: Cognitive Science and the Riddle of Human Creativity. Oxford: Oxford University Press. Turner, V. (1969). The Ritual Process: Structure and Anti-Structure. London: Roudedge and Kegan Paul. van Gennep, A. (1960 [1909]). The Rites of Passage. London: Roudedge and Kegan Paul. Voland, E. & Grammer, Κ. (Eds.) (2003). Evolutional Aesthetics. Berlin: Springer. Wade, N. (2006). Before the Dawn: Recovering the Lost History of our Ancestors. New York: Penguin Press. Wagner, S., Winner, E., Cicchetti, D., & Gardner, H. (1981). 'Metaphorical' Mapping in Human Infants, Child Development 52, 728-731. Waterman, R. (1955). Music in Australian Aboriginal Culture: Some Sociological and Psychological Implications, Journal of Music Therapy 5, 40-49. Watson, D. & Clark, L. A. (1994). The Vicissitudes of Mood: A Schematic Model. In P. Ekman & R. J. Davidson (Eds.), The Nature of Emotion: Fundamental Questions (pp. 400-405). New York: Oxford University Press. Wilson, D. S. (2002). Darwin's Cathedral. Chicago: University of Chicago Press. Zeki, S. (1999). Inner Vision. An Exploration of Art and the Brain. Oxford: Oxford University Press.