Introduction

1The “méthode d’évolution des motifs” endeavours to throw some light upon the still unknown or scarcely known evolution of an art, a style or a theme, by the synchronistic collation of the development of a rather large range of decorative designs especially chosen to that end.

2It is a flexible and altogether strict method, which, for each case, tries to model itself on what it meets, like a made to measure suit, not like a ready-made one.

3In a rough form, this method is constantly and often unconsciously used. But Philippe Stern specified, developed and applied it in a much wider and more systematic way. By this method, the spurious khmer and cham chronologies have been corrected and many other studies can be found in the bibliography.

4Inscriptions alone are powerless to recall the entire evolution of an art or a style. It is dangerous to squeeze these inscriptions, in order to draw out of them more than they contain. Other methods must then be used, among which that based on the evolution of decorative designs. If well used, it is a safe one. Even in religious and traditional arts, the whim of an artist is indeed always possible, as well as the accurate copy of an older model, un to the smallest details. But because of the multiplicity of constantly confronted designs, some escape this arbitrary or this copy. So, the general synchronism of rather numerous evolutions insures the fiability of the results more than the absolute preciseness in the detail of a single case. Each evolution backs the others up. Whim or copy are precisely pointed out by the lack of synchronism in the developments.

5It is more difficult than one would think to choose the decorative designs, whose developments will be examined. The most important, the most obvious designs are not necessarily the most useful, and the most spectacular changes do not always give the safest indications.

6At the very beginning of such a research, the best method is – paradoxical though it may seem – not to have one. Let us explain. Faced with a problem, the point is to approach it without any preconceived idea, to be able to put back into the melting pot anything that is not very sure and, out of the utter multiplicity of the pictures or of the monuments, to look for the leading thread, an end of which appears suddenly and fugitively (such as the stone imitation of the fragile, light wood architecture in cave 1 at Ajantā, or the hierarchy of the columns in the same cave); that is to say to multiply even bold and at the beginning unsupported hypotheses, providing that you have courage enough to give them up, if they do not stand fast later on. The result is a constant backward and forward motion, a zigzag path.

7If severe and absolutely objective checking is obviously compulsory once the results are established, long gropings in various directions are necessary to obtain these results first, operated with a special intuition for choosing from the themes and their variations, for apprehending the general direction of an evolution, and then adhering to the sometimes sinuous lines of its development.

8The first step is generally made by collecting what seems to occur simultaneously: clearly characterized appearances of certain signs, showing the periods of an art, a style or a theme.

9The contiguity of these groupings, that are close to one another, is shown by overlapping.

10It is then necessary to find out the direction of the development, what comes before and what comes next. Some evolutions are non-reversing, and when the chain is fairly long, this direction is obvious.

11At a given time, some conceptions and datings, which you were relying on before, are sometimes proved impossible; the hypotheses multiplying, intermixing and cancelling one another out, you feel dizzy, puzzled, at sea and discouraged. You are ready to give up. Perhaps this is to be compared to the periods of aridity of nearly every mystic: a necessary passage. And often, at the very moment when, completely at a loss, you believe that you are at the furthest point from the goal, some leading ideas, some “Ariadne’s clews” appear.

12Then begins a kind of rocking motion: the good gets the better of the bad, and a phenomenon of acceleration occurs. If on the right track, the different evolutions, directions and hypotheses, as if magnetic, come nearer and catch on to one another, backing and supporting each other. This faster motion in which all is uncontradicted is the very sign that you are on the right track.

13Checking has then to be very strict, objective, and every uncertain result must be given up or, at all events, pointed out.

14So a sequence has appeared. It still remains to try to nail this continuous and floating garland on the few dates that are firmly settled by inscriptions.

15If some sudden changes and combined reactions with possible contributions from outside-copies from former times, or imitations of other techniques-are taken into consideration, the evolution of a religious and traditional art shows a never completely interrupted continuity.

Pala and Sena steles. History and chronology

16It is possible to trace a history of Pala-Sena sculpture based on the analysis of decorative motifs. Four groups of steles have been determined, each of them characterized by the concomittance of motifs and a same state of evolution of these very motifs.

First period

17The first group (fig. 31 to 34) is distinguished by:

• a slabstone made of one single piece, whose upper part is rounded; some of them (fig. 31-32) are not rounded but rectangular: these pieces are hollowed out in the lower part, around the legs of the god,

• the surface of the slabstone is not decorated, it is simply edged with a single row of beading or a series of plain leaves (flames?), or also with a latticed band with flowers (fig. 31-32),

• the pedestal will undergo a particular tangible evolution. It is now made of one step (ekaratha), adorned either with the divine vehicle (vāhana), Nandin and the lion of Śiva and Pārvatī (fig. 33), the seven horses of Sūrya, or with styled lotus flowers which are engraved (fig. 31), then the pedestal is surmounted by a flower which supports the god,

• all deities (or secondary people, like vidyādhara) wear jewellery:

• one girdle, with a heavy pendant terminating in a lance-head element,

• a little corded loop on one hip (usually the right),

• each one wears two bracelets: one on the wrist (bracelet which will exist during all the Pāla-Sena period, without any change) and one on the arm, a little triangle fixed on a single ring situated high on the arm (almost near the shoulder),

• one worked necklace with, sometimes, a row of beads,

• women and vidyādhara have a round coil of hair on the top of the head with, sometimes, little beaded loops in front of it; Viṣṇu and Sūrya wear a mukuṭa, in the form of a frustum of a pyramid,

• a beaded yajñopavīta falls vertically from the left shoulder to the left thigh.

18The concomitance of all these common elements, which determine a first very compact group, cannot be accidental.

Second period (fig. 35-48)

19If we go through the steles, we find others which combine elements from the first analysed group, in an identical or a different aspect, and new elements. The first class of motifs proves the contiguity between the two groups:

• rounded slabstone, the rectangular form having disappeared,

• persistence of the row of beads,

• generalization of the small leaves (flames?),

• pedestal with one step (ekaratha), still adorned with flowers, generally not engraved, but sculptured in low relief (fig. 38), or with vāhana (fig. 35, 39) or with worshipping figures,

• girdle with heavy pendant,

• small corded loop,

• chiselled necklace,

• generalization of the beaded necklace, and

• the round coil of hair (women, vidyādhara).

20If some motifs have disappeared (latticed band with flowers), if some others have evolved, at the same time, new ones appear. Let us see these evolving:

• the pedestal may have three steps (triratha), its surface is smooth unless a filet runs along its upper and lower edges,

• some modifications at the waist: there are in general two girdles, hung with loops of beads: one on each leg,

• the triangular armlet has grown but is placed lower down,

• the women’s coiled of hair and vidyādhara are surrounded by crowns.

21And now, let us make the list of the innovations:

• leaves of a complex type (fig. 35-36, 38),

• beaded twist (fig. 36-37, 39),

• flames of a particular type, mostly to be found on buddhistic steles, and especially on Budha images (A. Foucher, Iconographie boudhique, I, fig. 23),

• an undecorated band which, perhaps, finds its origin in the beaded twist (the band may be beaded or not).

22Thus, this second period is very well defined and characterized by an important number of innovations and a great richness of alterations. On some steles, it is possible to discover other infrequent, uncommon motifs which will be generalized during the subsequent periods (some representations of vyāla, of haṃsa: fig. 37 and M.-T. de Mallmann, Introduction à l’étude d'Avalokiteçvara, pl. VIIIc, of kῑrtimukha: fig. 37, 46, of a flower-bud on the crown: fig. 37-39...).

Third period (fig. 49-60)

23Here too some motifs are inherited from the previous period:

• the undecorated band,

• the series of leaves of complex type,

• the triangular bracelet in the hollow of the elbow (fig. 49, 53-54),

• the two girdles with the little corded loop and the pendant without its lance-head adornment.

24On the other hand, there are motifs borrowed from the second period with alterations and innovations:

• not only two but four loops of beads hang from the girdle, between which little pendants appear,

• the top of the slabstone is no longer rounded but sharp-pointed, and, innovation, a monster’s face, kῑrtimukha, decorates it (fig. 49, 53-54),

• the halo (śiraścakra, mukhamaṇḍala), around the head is now generalized: not circular, the upper part being flat,

• the pedestal is higher and has more than three steps: five at least (pañcaratha) or seven (for the seven horses of Sūrya: fig. 69); it is characterized by some innovations:

• a little «tenon» is situated in the middle of the fillets of the middle step,

• the fillet is no longer flat, but has successive recesses,

• two sculptured flowers, in high relief, adorn the middle ratha,

• there are two pairs of vidyādhara at the top of the slabstone, their rounded coils of hair have disappeared to give way to pointed mukuta,

• the women’s coild of hair hang on the shoulder,

• an āmalaka tops the mukuṭa of Viṣṇu and Sūrya; a crown encircles it, it is adorned, among other things, with the flower-bud and the beaded loops seen in the second period,

• the gods may have a stole on their breasts, the uttarῑya, worn before by females only,

• the yajñopavīta is treated in various ways,

• the theme of the royal throne (gaja: an elephant, vyāla: a kind of griffin, makara: a marine monster, haṃsa: a goose or gandharva: a celestial musician) is completely represented,

• the kīrtimukha, already noticed, will undergo a change: it is now treated very harmoniously, with two large foliated scrolls in high relief.

25As before, between the second and third periods, here also some steles provide motifs which will be generalized only during the following period (conical mukuṭa: fig. 54, large ornate girdle with two corded loops: fig. 54, 56 or with a bow of material: fig. 58, flowers on the other ratha: fig. 58, loops hanging from the armlet: fig. 56, EISMS, pl. XLIIIc).

Fourth period (fig. 58-73)

26More criteria exist here to define the period. There are no great innovations, but on the other hand, almost every motif changes (thanks to these alterations, it seems even possible to determine three phases within the period).

27Without any discredit to it, we can consider this period as an extension of the previous one:

• the sharp-point of the upper part of the slabstone is more developed,

• the monster’s face is in low relief, it is treated with much more exhuberance,

• the slabstone is hollowed out all around the god,

• the halo is flat (consisting very often of an undecorated band),

• the whole theme of the royal throne is represented, but without the simplicity of the previous period: wild fancy prevails (little men are mixed with the hybrid monsters vyāla, the makara are of a very great beauty, the two haṃsa are replaced by a couple of gandharva...),

• the vidyādhara still wear a pointed mukuṭa,

• the loops of beads and the very heavy pendant hang from the girdle, there is still the little corded loop but we must notice a new element: a little bow of material,

• a new type of necklace adorns the breast of the gods (sometimes) and of the goddesses,

• a last generalization is to be noticed: the little beaded loops hung from the armlet,

• an ornate band encircles the women’s coils of hair hanging on one of their shoulders,

• the mukuṭa is now conical, and always surmounted by the āmalaka,

• the jaṭāmukuṭa (fig. 64) is made of scrolls: it is easy to follow this evolution from C. Sivaramamurti, L’art en Inde, Mazenod, Paris, 1974, fig. 70 and our fig. 55, passing through fig. 56 and coming to fig. 64.

28We can see by the examples that some motifs undergo no real change, but slight alterations. The same is seen on the club (gadā) and the discus (cakra) of Visnu: little bows top them. The evolution easily perceived in the jaṭāmukuta is also seen in the pedestal, with seven ratha (saptaratha), with flowers and rhizomes of flowers on its entire surface.

29When one considers the whole Pāla-Sena epoch, one observes that the artistic history is straight forward, and is traced here by means of the decorative motifs alone. But we can try to go beyond this relative chronology, and ask for an absolute one: some data are supplied by history, artistic or not, Indian or not, which make some dated scale possible. In this framework of (rare) dates, one will be able to inscribe first some steles, which are «dated» (as ascribed to a particular year of the reign of a king) and then the undated ones will fill up that framework.

30But, at the same time, another approach will be necessary: to compare «dated» steles, as yet unstudied, with our results. Then, we shall first ascribe these «dated» steles to one of our four periods, after which we shall put these same works in the dated scale, and, finally, we shall try to define the chronological limits of the four periods.

31But that will constitute the second part of this research that will not be presented here.

The Khmer colonnette

32There are few designs which present, as well as the khmer colonnette, a continuity of evolution, the persistence of a direction which, though it may weaken at times, is taken up again each time. Khmer architecture equally shows an astonishing development which unmistakably reflects the rigour of the khmer mind. The khmer lintel reveals the frequent renovations of this art by drawing upon the past. The khmer colonnette combines the two tendencies.

33This trend which continues without interruption is more evident during the second epoch of khmer art, in the shift from the round colonnette to that of octagonal form, the period streching from the beginning of the ninth century to the late twelfth century.

34The salient feature of this evolution seems to be the increasing decoration in strict gradation.

35Thus emerge the two aspects which apparently are contradictory but still constantly go together in khmer art: a taste for richness (such as the proliferation of life) and a liking for order (geometrical, neat and clean graded divisions on each side of the centre).

36From the ninth century onwards, the rings and groups of mouldings of the colonnettes are more numerous, increasing thus the number of the divisions of the shaft. The number of the most important elements of these groups goes on increasing from 3 to 5, from 5 to 7, from 7 to 9. Up to the eleventh century the rings generally increase in size and relief.

37On the other hand, the blanks that separate the groups of rings and mouldings diminish in height as also the size of the leaves that decorate these blanks becomes smaller. Consequently, the number of these leaves increases, changing from one large leaf on the side to a series of leaves and tiny dog-tooth pendeloques.

38But be it in life or in art, there are periods of renovation that, like the throbbing of the heart, bring new blood to an art which time has sclerosed and impoverished. Then sharp changes appear, or reactions to the established order, an influx of the new elements or the influence of other techniques. But, as Gilberte de Coral Remusat has discovered, khmer art looks to its own past for new elements; it turns for inspiration to the richest period of its recent past. But this is not archaism; the aim is not to imitate the past, but to renovate it. The décor is taken up once again to be transformed and adapted. It is absolutely impossible not to distinguish it from the past as the new and distinct features cannot be missed. Moreover certain décors are copied without being understood: it happens more than once that the pendeloques appearing on the back of the colonnette ascend. Sometimes it also happened, especially at Banteay Srei, that the artists copied down the imitation exactly as it was, including the mistakes committed in the first imitation.

39A strange consequence of this is the increasing dismissal of elements during the transitory periods: in fact, the imitation of the past orientates us towards the despoiled models. During the period of these renovations, we find the colonnettes relatively overornate. Now, the renovational periods sought to react to the habitual, by finding something new to rejuvenate by diversity rather than to embellish themselves by the accumulation of elements. Thus, it is curious to note that the khmer colonnette at these periods is not overornate, but influenced by and also reactionary towards the past, it is simple and free from restraint.

40In the general evolution which is a sort of vital wavering between the new décors often taken up from the past, the momentary interruptions are found in four long periods of the khmer renovation: in the Kulên, first half of the ninth century (complete freedom and renovation, it is the lintel which was borrowed from the past); at the beginning of the school of Prah Kô (the last quarter of the ninth century where the past was freely imitated); in the school of Banteay Srei (967) where freedom is still more evident in the imitation of the past; at the beginning of the Angkor Vat school, in the first half of the twelfth century where the freedom is but momentary.

41Thus the development of khmer architecture and its colonnette appears to be an evolution following a continuous direction, which in a way is logical: the continuity and the regular orientation are particularly evident in the change from the round colonnette to the octagonally shaped colonnette in the second epoch of khmer art.

42The evolution of the colonnette has been examined from various points of view: mainly the décor but the form and to some extent the capital and the base are equally valid points.

43From the point of view of form, the history of the colonnette has been divided into two parts: these two divisions coincide with the two major divisions of khmer art in its entirely.

44The round shape was always used in the 6th, 7th and 8th centuries (the epoch which is often called – very unjustly in our opinion – preangkorian art). The scanty and timorous specimen of the octagonal form from which the Kûlen style was to derive, appears only in the late 7th century.

45The octagonal form is almost constant from the beginning of the ninth century up to the thirteenth century. Lending its own characteristic to the more important half of khmer art, it contrasted radically with the round form of the preceding period. The square shape of the secondary colonnette of the first half of the ninth century is a trial form which was soon dropped. The sixteen-panelled form is a short-lived experiment of the Angkor Vat which is the period of renovation, not to be taken up again. The round shape reminds us of the past renovatory periods (the Prah Kô style during the last quarter of the ninth century and the Banteai Srei in the second half of the tenth century, and the continuation of these periods) but it does not continue either.

46The décor in the first epoch of khmer art at first decorates the top, the lower part and the centre of the shaft of the colonnette; there we find a garland, a ring and leaves along with the intermediary fillets which soon increase in number and then the décor appears on the sides of the ring. We also find that, as on the lintels, the garland over the leaf-work changes for some time, accompanied by two motifs, and leaves with points folded to the side are added to the leaves seen from the front.

47We find that the evolution of décor, beginning from the early renovation of the ninth century up to its adapting the octagonal form, is distinctly divided into two parts:

• décor consisting of rings and elements of relief – these combinations, distinctly graded at the centre, the quarter (or the sixth), the eighth (or tenth or twelfth) increase in number, thus dividing the shaft of the colonnette in 10 or 12 parts, at the end of the evolution (during the first three quarters of the twelfth century when Angkor Vat school flourished). Each combination consists of more and more numerous elements: to the three elements, i.e. the ring and the two filets which enframe it, is annexed a row of lotus buds on each side (five components); later on, on each side an ascending row of petals was added (seven components) and finally, at the very end of khmer art, we find the reduplication of the petals which has already occurred (nine elements). The mouldings increase and their relief becomes strongly marked.

• décor of the blanks between the rings – the leaf-work seen full-face: the leaves, at first one on each side, increased in number – getting smaller at the same time. During the periods when imitation of the past was in vogue and the period that followed them, the pendeloques which are often copied unintelligently (not only are they descending but also ascending) were often separated. In the course of the period of renovation, the décor of leaves undergoes modification, becoming richer and more and more diversified.

48Gradually, the leaves diminish in size to remain as the dog-tooth décor in the late stage.

49The capitals are now more visible than the bases of the colonnettes which were so often destroyed. The lack of documents precludes the use of the available data which is to be regretted because the small figures which ornate it and are rarely preserved, would have provided very precious information.

50What is to be particularly noted in the capital is its shape which was so plain at the beginning: a bulb, a fillet, the widening shape and the square abacus are superposed. This schema becomes gradually more and more intricate but evolves very slowly over a period of many centuries. This change is undoubtedly due to the lack of understanding at the beginning of the tenth century (the Bakheng style): with more and more bulbs superposed which gradually flattened, we finally get the pile of mouldings superposed, as if it were a pile of dishes, as in the Angkor Vat school, (first three quarters of the twelfth century).

51It appears that the decadence of the colonette corresponds to the swan song of khmer art and to the wonderful style of the Bayon (from 1181 to 1219 approximatively), but the superabundance and the huge dimensions of the edifices lead to an economy of labour and to a loss of interest in décors.

52The overdecoration continues but the gradation disappears. We no longer find the earlier gradation, nor the composition dividing the shaft of the colonnette at regular intervals or ending it with a capital. The capital and the small dog-tooth leaves that decorated the blanks disappear. All that remains is the rings with some standard, identical combinations, each of them followed by a reduced blank and this alternation occupies all the space that is available.

53The colonnette itself is just one of the many elements of khmer décor. Perhaps the lintel is still more important. The spandrels and the frames of the façade, the pilastres, the hidden doors. the sub-foundation, the mouldings, the décor on the walls, could all be examined.

54Be it in khmer art or any other art, when we try to study the growth, the inflexions of the evolution of the motifs, we are overwhelmed by the richness of the decorative elements which one single art can possess.