Interpretation of Shields


Body, message and empowerment: shields in Island Southeast Asia and Melanesia

Interpretation of shields

Back to main shields page

Back to main 'thinking about objects' page

Extract from M. O'Hanlon's Medusa's Art: Interpreting Melanesian Shields, in Tavarelli 1995.

Pp 74 - 104. References cited by O'Hanlon are given in the bibliography.

... [N]ot all groups in Melanesia possessed shields, and where shields were found they varied in shape, size, weight and in the materials from which they were made (although wood is the base material in most cases). These variations reflect both differences in the weapons against which the shields were used (which included arrows, spears, axes and clubs) and the way in which the shields were deployed. ... Nor do all shields serve purely defensive purposes. Trobriand 'dance shields' were just that: ornamental items which were spun and twirled in dances. But as will emerge in more detail, 'practical' and 'ceremonial' aspects should not necessarily be contrasted, for the powerful visual impact a shield makes may be thought of as providing as much protection as its overtly 'practical' dimensions. Some shields have multiple uses ... as stretchers to carry wounded warriors from the battlefield, while equally large New Guinea Highland shields were often re-used as beds or as house-doors.

Nor, among those Melanesia groups which did possess shields, is there any easy correlations of the 'one tribe, one shield-type' variety. Some groups produced two or more types of shield (Sillitoe 1980); indeed differences between the types of shield used by a single group may calibrate local distinctions between degrees of severity of fighting. For example, the Wahgi people of the Papua New Guinea Highlands possessed narrow stick-like parrying shields in addition to their massive, oblong battle shields. In intra-clan disputes, Wahgi clansmen restrict themselves to clubs and to parry shields; battle shields were only for use in serious fighting and their adoption in clashes between fellow clansmen indicated that the clan itself was in danger of splitting (O'Hanlon 1993:34). Shield designs equally resist being pigeon-holed. While in some instances, they reflect the cosmological dimensions to warfare alluded to earlier, in others it is far less easy to detect a semantically meaningful link between shield motifs and cultural themes. Moreover, motifs are also borrowed and travel between different groups. Equally, the rules which often segregate the sexes in warfare may extend to shields, though they do not necessarily do so. Thus Hagen men were meant to segregate themselves from women and to avoid sexual intercourse while making their shields if the weapons were not to be weakened (Strathern and Strathern 1971:104). Their Wahgi neighbors, on the other hand, who have very similar shields, deny that they exercised such restrictions, saying that their own taboos related rather to the ritually-treated charcoal in which Wahgi warriors, like many in Melanesia, traditionally decorate themselves.

But this introductory overview of Melanesian shields and their variation disguises two obstacles to a more systematic account. The first is simply that colonial contact was, in many instances, too early and too disruptive for rounded knowledge about the context and meaning of shields to be recorded. The long-term fieldwork that might have gathered such information did not become an established anthropological practice until the early twentieth century, by which time massive and irreversible changes were already in train in parts of Melanesia, so the full significance of shields .. . is probably irrecoverable... The second obstacle to any systematic comparison is that what anthropologists and others did record depended, to some degree, upon the intellectual climate prevailing at the time the knowledge was gathered. The kinds of questions that were asked about shields by observers a century ago are often rather different from those asked now...

Some of the earliest systematic work on Melanesian art was done by Alfred Haddon, who spent considerable time in the Torres Strait at the end of the nineteenth century and also visited the Papuan Gulf. As a man of his time, Haddon (1895:2ff) was influenced by ideas of social evolution and regarded the study of the art of 'laggard peoples' such as Melanesians, as a necessary preliminary to understanding 'civilized art'. As a biologist turned ethnologist, he proposed that designs had 'life histories' which evolved from 'birth' through 'growth' to 'death'... Haddon arranged the shields in a sequence according to his perception of the evolution of their designs. Haddon's careful attention to form had the merit of showing that apparently abstract designs were sometimes transformations of more figurative forms (for example, the human face) which had been re-arranged, doubled, simplified etc. What is less clear, however, is whether arranging designs in such sequences made much sense in terms of local culture: indeed, when faced with a shield whose design did not fit into his sequence, Haddon (1894:94) rather unhelpfully suggests that 'it looked as if the artist occasionally did not quite understand what he was doing'.

Partly because art has been misused in this way as evidence of dubious evolutionary speculations, its analysis was temporarily discredited, and no study which might have brought out the full local significance of Papuan Gulf shield design was made before the area was changed irreversibly by colonial impact. Broadly, with the exception of stylistic analyses carried out by museumists, the study of art remained out of anthropological fashion until the middle of the twentieth century. One early sign that this situation was changing was the debate over another Melanesian shield type, the unique, painted, ovoid shields of the Trobriand Islanders. In this instance, much more local information on shields and their context was available, for the Trobriand Islands were the site of what is generally acknowledged to be the first modern anthropological fieldwork, that carried out by Bronislaw Malinowski between 1915 and 1918. Although Trobriand warfare had ended before he arrived. Malinowski (1920) did provide a brief account of the shields. Among other points, he described how most shields were left undecorated: only the most courageous warriors painted their shields, for to do so was a challenge and a painted shield attracted many more spears than did a plain one. However, partly because he was writing at a time when studies of art were unfashionable (indeed, his own work helped discredit the conjectural history into whose cause art had been recruited as evidence) Malinowski published nothing on the meaning of the shields' designs.

After Malinowski's death, however, Edmund Leach (1954) , one of Malinowski's most brilliant pupils, drew upon his teacher's many writings on the Trobriands to propose an interpretation. Briefly, Leach suggested that the design was a representation, folded up vertically, of a female anthropomorphic figure, the lower half showing the pubic area. The figure, Leach argued, represented one of the feared, flying witches that appear in Trobriand mythology, and are believed to give off a poisonous emanation from anus and vulva. The design, Leach proposed, would thus employ the same logic as did Perseus's shield, with its transfixing image of Medusa. Leach's interpretation was then challenged by Ronald Berndt (1958), who argued that the figure was unlikely to represent a witch, since witchcraft - in Trobriand thinking - is a female property while warfare was exclusively a male concern. Instead Berndt reminds us that Trobriand men were supposed to abstain from intercourse in time of war. The shield design, he suggested, represented a male and female about to engage in intercourse ... In Berndt's view, the logic of putting such an image on a shield was not that of Perseus, as Leach had proposed, but rather as a form of visual abuse of the enemy.

So Leach and Berndt take here a rather different approach from Haddon's to the analysis of design. Where Haddon had been concerned with the evolution of design and with delineating stylistic regions. Leach and Berndt pass over the first as unknowable and take the second as given, and focus instead on the significance of the design in terms of what is known about the details of Trobriand culture itself. But, as we have seen, they reached mutually contradictory conclusions. Both, however, acknowledged that their arguments were conjectural, since the information necessary to confirm them had not been recorded from Trobrianders at the time. Subsequently, though, it emerged that a Trobriand gloss on shield design had been published earlier in an overlooked report by the missionary S.B. Fellows, and, moreover, that Malinowski had confirmed its accuracy in an unpublished letter.

But this additional information proved inconclusive, for the Trobriand gloss turned out merely to take the form of a list of terms for the individual motifs from which the design as a whole is made up. The terms were those for various natural phenomena: birds, fish, stars, snakes, rainbows etc. None referred, at least overtly, to witches, anatomical organs or sexuality, as Leach and Berndt had variously proposed. In his earlier letter confirming the accuracy of Fellows' report, Malinowski had observed that 'the native 'theory' of their art consists simply in a series of names given to the various individual motives, but there is no sense, no meaning to the whole design ... No deeper magico-religious meaning, absolutely none - and this applies to their art throughout'. However, ... Glass (1986) argued, the design on war shields was a code operating on three levels, referring to Trobriand mythology, to copulation and to Trobriand sacred geography.

The question of 'the meaning' of shield (and other) designs can thus be very complex. The Trobriand case, to date, offers no conclusive answer, although the general appearance of the design does suggest that it once represented something fairly elaborate, even if its identity is now lost. In the case of the Asmat, who live on the other side of New Guinea in Irian Jaya, the position is a little clearer. Asmat shields, of which there are a number of regional subtypes, are produced by specialists who carve them in relief designs which are enhanced by differential coloring. In Asmat thought, life and death are regarded as mutually implicated, so that the continuance of life and fertility is seen as dependent upon death. Traditionally, this belief system manifested itself in headhunting, a practice which the Asmat also saw as analogous to the behaviour of various local creatures: flying foxes, hornbills, and cockatoos, all of which eat fruit (which is likened to heads), and the praying mantis (which bites off the head of its mate). It thus makes sense that these are among the dominant motifs in Asmat art, with flying-fox motifs (often in very schematised form) occurring on shields in particular.

The small figure atop [an Asmat shield] demonstrates a slightly different way in which shield design may be significant: not simply in representing some external being ('witch', 'flying fox' etc) but through establishing a link between a shield and its owner. Smidt (1993:22ff; 71) reports that many Asmat carvings, including shields, were names after dead relatives. In the case of shields from the central Asmat, an image of the dead person was often carved on top, and other dead relatives (including, interestingly, women) were commemorated in the pattern carved on the front. These relatives were thought to offer support to their living kinsman in the fighting in which the shield was used, and might be avenged by him. On the death of the shield owner, the small figure atop the shield was sometimes cut off. An alternative manner in which the shield may refer to the identity of its owner is given by the rare shell-inlaid shields from the Solomon Islands. Though the early date of colonial contact in the Solomons means that the evidence is fragmentary, it seems likely that these elaborate shields, in which designs in pearl shell have been superimposed on the wicker wood or bark base form which ordinary Solomon shields are made, were social markers of some kind, referring perhaps to the status of their owner (Waite 1983: 129).

Of course, these two modes of references are not mutually exclusive. Thus, Trobriand shields also adverted to the status of their owners (in that only the bravest warriors were said to paint their shields), as well as variously representing witches, sexual congress etc, depending upon whose arguments are found to be convincing. But here we should also be aware of our predisposition, stemming from our own specifically western experience, to search for meaning, iconographic or otherwise. Conditioned by literacy (not a traditional practice in Melanesia) and by our daily exposure to its spin-offs in the form of advertisements which we are invited to 'decode', we tend to expect designs and their constituent motifs to be meaningful, often in a specifically graphical way (see Ong 1982). This can lead us to over-interrogate Melanesian forms for their 'meaning' and to overlook other modes in terms of which shield design may be significant.

This is implicitly recognized by Cherry Lowman in her study (1973) of Maring shields from the Papua New Guinea Highlands. These large, oblong shields are used not only in formal warfare, but also in what Lowman (1973:13) appositely calls 'persuasion displays', in which warriors from the opposing sides massed opposite each other, maneuvering and chanting, gauging each others' numbers and strength. The shields themselves were mostly decorated in large, bold, and overtly geometric forms, executed in contrasting colors. When pressed, Maring would give names to their designs - 'lizard', 'axe blades', 'frog's legs' etc - but Lowman (1973:26, 29) found that 'What is represented ... cannot be considered important for there is rarely consensus as to what object is actually being represented .... One gets the impression that the iconic forms in the designs are projected or interpreted, rather than intended.' Instead, Lowman argued, the significance of shields and their designs resided in their contribution to the 'multi-sensory' encounters in which shields were deployed. Their large bright designs would have stood out at a distance, visible against the forested environment in which the Maring live. The designs were intended to intimidate, not through what they represented, but through their dazzling appearance, just as the massive shields increased the warrior's apparent size (particularly since, in use, whippy canes decorated with cassowary plumes are mounted above the shield)...

However, what has undoubtedly been neglected in the literature's preoccupation with shield designs (whether examined from the point of view of their 'life history', their representational significance or their affective impact) is the provision of equivalent material on how shields were actually used. In some cases, we know that there were specialist shield men. For example, the Telefomin shield carrier was unarmed himself and directed the tactics of the bowmen who accompanied him - although a notable warrior might sometimes rush forward and pin down an enemy with his shield until the bowman could shoot him (Cranstone 1968:612). Papuan Gulf shields, on the other hand, were (as we have seen) devised to be used by an archer. But in other instances, even in the case of well know shield types, we often do not know how many men carried them in battle and how, tactically, they were deployed in relation to other weaponry. And while museum records often provide the dimensions of shields, they seldom give their weight: yet this is likely to have been as significant a factor as size, for the capacity of a shield to withstand a blow depends, in part, upon its inertia and thus its weight.

The literature's relative neglect of shields as objects of use as opposed to purveyors of meaning is especially apparent in the lack of attention given to the reverse side of shields. Of course, some shields (such as those made by the Arawe ... ) are almost as elaborately decorated on the reverse as on the front, which itself implies a slightly different role for the shield design since in such cases it would seem to be oriented as much to the shield bearer as to his opponent. But what is more remarkable is the inattention to shield support-mechanisms: the handles, straps and loops by which they are variously carried and manipulated when in use. The omission is odd, not only because the type of shield support mechanism must, in turn, influence how a shield design is seen by opponents and spectators, but also because at a gross level, the type of support mechanism offers clues as to weapons' use more broadly. Thus, we could probably suggest as a general rule that heavier shields, which could not easily be held in one hand, have straps allowing them to be supported over the shoulder... Shoulder straps potentially free both hands. so that the bearer can also act as an archer, whereas if one hand is occupied in shield-holding, the weaponry held in the other is more likely to take the form of spear or club.

But shield support mechanisms can vary enormously, even within so limited an area as the West Sepik regions (Craig 1988:62) and even in otherwise well-documented cases it may still not be clear how a shield was actually carried when in use. Lowman (1973:24), for example, describes Maring shields as being supported over the shoulder by a long vine fixed vertically just short of the shield's right hand edge. However, it would seem from her diagram that donning the shield in the way she describes would actually leave it slung over the warrior's back, and thus in no position either to protect him, or for the design to have its intended intimidatory effect upon his opponents. This neglect of the pragmatics of shield use is likely to be due not just to a preoccupation with their designs, but also to the fact, that, until recently, few anthropologists had seen Melanesian warfare at first hand. In the last few years, however, this situation has changed in the Highlands of Papua New Guinea where, for a variety of reasons (Strathern 1977), a destructive recrudescence of inter-group fighting has taken place, and for the first time in fifty years it has been possible to see shields in use. One such area of the Highlands is the Wahgi, where John Muke (himself a native Wahgi speaker) has made a study which brings out in unique detail the complexities, both cosmological and tactical, of shield use in p