Visual style hermeneutics: from style to context

This essay re-examines the once promising idea that style analysis can provide an independent source of insight into an artifact's non-stylistic context. The essay makes explicit the consequences of treating collective style as such a source in archaeology and anthropology of art, and further develops a new framing for the idea that avoids the criticisms largely responsible for the decline in theoretical interest in the epistemic import of visual style analysis since World War II. This re-framing proposes that inference from style to context is permissible on those occasions when a collective style signals by its morphology its suitability to serve a certain function. And it does so because it prescribes publicly certain modes of behavior or spectatorship. Furthermore, the public nature of the signaling may be such that it allows even uninitiated spectators to get a sense of it and thus to gain access to some of the motivations and norms informing the collective's form of life.


Introduction
Is it ever permissible to infer from style to context? Namely, is it ever justifiable to treat the character of a collective visual style as revelatory of the collective's norms, institutions, or attitudes? The question is both alluring and controversial. It is alluring because to answer it in the positive is to suggest that analyzing visual styles of material cultures may provide independent insight into these cultures' circumstances, even when little is known about them otherwise. It is controversial because the story of efforts at such a stylistic analysis is rather unwholesome; what it reveals is a reliance on holistic, essentializing, and often racialized notions of collective style as expressive of culture. To put it bluntly, it often reveals a whole lot of ill-motivated wishful thinking mixed with reliance on pseudo-scientific conceptions of physiognomy and race-based anthropology (Gombrich 1963;Kasfir 1984;Summers 1989;Gell 1998, 216;Michaud 2019). This helps explain why theoretical interest in the epistemic import of visual style analysis has been in steady decline post-World War II in social and historical sciences, the trend intensifying at least since the 1980s. There has been no recent discussion of style comparable in its prominence to, for example, debates on 'isochrestic' variation in archaeology (Sackett 1985(Sackett , 1986(Sackett , 1990Wiessner 1985Wiessner , 1990 or on the expressive power of Oceanic styles in anthropology (Forge 1979;O'Hanlon 1992;Roscoe 1995;O'Hanlon and Roscoe 1995). Perhaps most striking is the virtual disappearance of theorizations of style in art history, 1 where it has been the central concept for much of the twentieth century, but 'the king has been dead' 'since the revolution of the seventies and eighties' (Elsner 2003, 98). 2 It would seem prudent, then, to be sceptical about the possibility of there being a reliable general (let alone objective) way of correlating collective styles and extra-stylistic context based on stylistic properties alone; that is, a way which would 'on pain of circularity […] separate out those aspects of the impressions works give of their sources which merely reflect what we already know or believe about them' (Walton 2008, 248). Instead of defending the prospects of somehow rescuing the largely discredited line of reasoning, I offer a different framing of the problem that will dispense with the unwanted ideological baggage. This re-framing proposes that inference from style to context is permissible on those occasions when a collective style signals by its morphology its suitability to serve a certain function. And it does so because it prescribes publicly certain modes of behavior or spectatorship. Furthermore, the public nature of the signaling may be such that it allows even uninitiated spectators to get a sense of it and thus to gain access to some of the motivations and norms informing the collective's form of life.
This strategy promises both to keep the alluring prospects of inferring from style to context and to avoid the major pitfalls of the various forms of holistic expressivism. But it does so at a price: first, it goes against the grain of the received wisdom about style in archaeology, anthropology, and art history, namely, that there is no such thing as an objective demarcation of collective style (Ackerman 1962;Alpers 1987;Hodder 1990;Wollheim 1987Wollheim , 1995Shanks 1996;Elsner 2003). And second, it treats collective styles as instrumental structures, thus endangering the distinctness of classification by style, which has traditionally been treated as different from classification by function (Sackett 1990; Davis 1996, 173-174). At the same time, the notion that collective styles provide access to their context by standing in some non-instrumental relation to it creates space for precisely those vague invocations of collective expression or unconscious communication that have made collective style analysis a deeply suspicious tool of social or historical enquiry. In shutting that space down by tying style to instrumental structure, my proposal provides solid immunity against similar specters. The crucial question facing us then becomes: how can collective styles be thought of on the model of instrumental structures without collapsing the difference between them?
After demonstrating what is at stake on the example of the Hittite Sword God relief (Part 2), I review three ways in which collective visual style has been understood to reveal information about context (Part 3): First, stylistic accidence and variation have been treated as tracking social stratification; second, stylistic characteristics have been used to corroborate or lend support to non-stylistic evidence; and third, styles have been perceived as evidence sui generis, implying context by their very morphology. It is only the third kind that presents a clear case of inferring from style to context. I then discuss what the required parameters of a corresponding notion of style should be, namely, that it be (in a specific, weak sense) universal and non-conventional. And I show how these parameters are applied in instrumental structure analysis that infers from morphology to function (Part 4) to argue that the universal non-conventional link between style and context should be framed in terms of instrumentality (Part 5).

Sword God
In hilly Central Anatolia, just a kilometer outside of what was once the Hittite Empire's capital Hattusa, lies Yazılıkaya, the home of the Empire's central sanctuary. Among the reliefs and friezes dating back to the 13th century BCE and covering the walls of the open-air rock shrine one finds a remarkable relief sculpture. Carved into a rock face of a narrow cleft called Chamber B and 3.38 m in height, it is the tallest of the large reliefs in the sanctuary as well as the only one of them that implements explicit vertical bilateral symmetry as its organizing principle ( Figure 1). Its singular appearance has not gone unnoticed. It has been described as 'highly unusual' (Seeher 2011, 114), 'remarkable' (Bittel 1978, 21), and 'unique' (Burney 2018, 308). Referred to as the 'Sword God,' the relief depicts a figure composed of a profile of a human head with a peaked and horned capa sign of deityand a vertically split mirror image of two lion profiles. The composition suggests both a human-like figure and a sword grip or hilt, the latter aspect supported by the depiction of what looks like a double-edged blade of a dagger or a sword plunged into the ground ( Figure 2).
The leading expert on the sanctuary's archaeology, Jürgen Seeher, relies mainly on two facts to interpret the Sword God's appearance. A Hittite sword hilt survives that uses the lion motive and bears an inscription dedicating it to what is assumed to be one of the gods of the underworld. Yet the hilt, while symmetrical, looks nothing like the Sword God and does not employ bilateral symmetry to compose an anthropomorphic bodynot to mention that it may predate the relief by half a millennium ( Figure 3). The second piece of information is that there also exists a Hittite inscription describing the acts of a priest who models out of clay god figures in the shape of swords and, depending on interpretation, either sticks them in or spreads them on the ground (Seeher 2011, 114-115).
If one wants to learn what motivates the unusual appearance of the Sword God and what the appearance suggests about its context and the environment within which the image was commissioned, produced, and observed, the interpretation that relies on the two contextual facts and that identifies the Sword God as one of the lords of the underworld is rather frustrating. Granted, it does make an effort at explaining why the relief employs bilateral symmetry and depicts lionsby speculating about a stylistic tradition exemplified by an actual Hittite sword hilt. But it does not infer from the presence of explicit vertical bilateral symmetry any observations about its possible context.
To infer from style to context is different from letting context explain style or from corroborating context with style. It is to treat stylistic configuration as providing an independent access point into circumstances of production, circulation, and use. To regard similar inferences as legitimate modes of research is to assume that isolating principles of stylization may potentially contribute to social or historical knowledge wherever these principles obtain. In other words, it is to presume that stylistic analysis may provide original insight into the nature of a collective based on the presence of a style: The style is no more an explanandum, it is an explanans. This does not mean that stylistic analysis can 'substitute for history' (Sauerländer 1983, 267). It is a rare breed of a historian who would share the excessive optimism captured in the modernist architect Adolf Loos's claim that serves as this essay's motto. By contrast, the ambition is not to filter out everything one might know about the Hittites and rely purely on the Sword God's style to learn of its context. The claim would rather be that stylistic analysis may legitimately enrich our knowledge of the Sword God's context by providing access to information that is encoded in the style itself. The aim of this paper is to present the claim in as intellectually coherent a way as possible.

Collective style and socio-historical enquiry
The notion of visual style as it is used here is confined to objects of material culture. It amounts to a mode or type of visible configuration or delivery that is recognized as instantiated by a particular object and therefore as at least potentially not unique to this objecteven if it may be the only extant realization of the style (comp. Schapiro 1953;Ackerman 1962;Gombrich 1968;Sauerländer 1983;Hodder 1990;Davis 1996;Lang 1998;Elsner 2003;. 'Context' refers to any non-stylistic historical variables that may co-determine the appearance of a style such as functionality, norms of making or behaving, social structure, or information exchange. Let us review three ways in which collective visual style has been understood to reveal information about context. (1) When visual style analysis is used in historical or social inquiry, it is often assumed that there are certain limits to who can legitimately replicate stylistic properties (Neer 2005, 11-12). So, for example, a genuinely individual or personal style is often taken as a signature style in that it can only be delivered by one particular individual (Wollheim 1987, 184). Or an artifact may be treated as truly done in a collective style such as a period, tribal, clan, or household style only when its history of production places it squarely within that social unit. In such cases, visual style analysis establishes replication patterns within a class of artifacts with the same history of production. On its own, such an analysis may help ascribe an artifact of unknown origin to its proper context of production or identify a common pedigree of a set of stylistically similar artifacts, with the proviso that the classification is incomplete unless and until corroborated by extra-stylistic facts to avoid misidentifying a fake, a stylistic borrowing, or an object with an altogether different origin (Davis 1996;Neer 2010, 6-11). But nothing about the stylistic configuration as such helps clarify the motivations behind its implementation.
One comes across such a use of stylistic analysis whenever style is treated as 'diagnostic or idiomatic of ethnicity' (Sackett 1990, 33) or as 'localizing social units in time and space' (Bettinger, Boyd, and Richerson 1996, 133) in the sense that its presence does no more than signal a producing community. No further claims about the collective style's expressive or signifying potential are made, nor is the signaling treated as following general style-context-matching principles. In processual archaeology (particularly in 'ceramic sociology'), fluctuations in stylistic variability across time or space were often explained as correlating with shifts in social stratifications: the greater or more rigid the hierarchical stratification of a society, the greater the stylistic complexity (e.g. Fischer 1961;Merrill 1987, revived more recently by Peregrine 2007). The stylistic variations were claimed to follow universal deductive principles, but the principles did not address the particular configurations of the decoration (for decoration was what 'style' meant in this scholarship) and what they may have revealed of their context (e.g. Plog 1980;Pollock 1983). 3 (2) A more ambitious use of visual style analysis in social or historical inquiry treats it as complementing the analysis of an individual or a community within which the stylized artifacts are produced and circulated. In identifying correctly an artifact's style, one can corroborate or build on the body of knowledge about the values or motivations of an individual or collective agency responsible for the artifact's production. An individual style is treated as the product of the psychological outlook of the individual and along with other behavioral patterns contributes to a larger picture of said individual's psychological profile. Similarly, a collective style is informed by and contributes to the form of life shared by the collective, meaning stylistic analysis can provide an access point to understanding it. Typically, one's reading of a collective style relies on endemic ethnographic or historical circumstances of production that tie the style uniquely to a particular time and place. Because of its role as complementing or corroborating non-stylistic evidence, this stylistic analysis is not an independent source of information underived from the context. It is thus perhaps better understood as a method of letting the context explain the collective style rather than the other way around.
In his explanation of traditional Maori designs and decoration, which is predicated on a structuralist homology between visual collective style and underlying social structure, Hanson (1983, 78) observes that Maori designs are 'studies of nearly mathematical precision in bilateral symmetry' and argues that these patterns correspond to 'large segments of Maori myth, religion, traditional history, and social, economic, and political behaviour' which are also 'organized in terms of duality' (79). Maori people make sense of the world in dualistic fashion, Hanson claims, and this structure underlies all their cultural institutions, including their visual art ( Figure 4). In his discussion of the Marquesan art style and its relationship to community, Gell (1998, 216) dismisses what he takes to be Hanson's reification of culture into a '"head office" which decrees […] what artifacts will look like'. Instead, he sketches an explanation of the influence of culture on style that consists in isolating an 'axis of coherence' governing the modus operandi of a collective style (effectively a description of its general formal procedures). This axis is then to be traced to the modus operandi in other areas of production in the given society. If these axes of coherence or structural principles match each other, as Gell argues they do in the case of Marquesans, then one is justified in viewing the stylistic system as motivated by the principles organizing its producing culture (Figure 5;Gell 1998, 219-220). Style is explained by reference to the specific forms of comportment in a given society: artifacts take on stylistic features because certain norms of production and use obtain across the various branches of culture and not because these  features issue from or somehow 'express' a culture. Gell's criticism notwithstanding, both Hanson and Gell engage in a similar project: they treat visual collective style as informed by a recognized underlying structure to which it in turn provides an access (for another example, see Washburn 1999).
(3) An even more ambitious socio-historical application of visual style analysis generalizes from the findings of context-based style analysis to produce universal principles of style-context correlation. It treats certain stylistic features as universally indicative of extra-stylistic context informing the style. This kind of visual style analysislet's call it 'universal visual style analysis'relies on the possibility of isolating instances of a universal style. The claim to universality of a collective style does not rest in its universal visibility to an 'innocent eye' uninformed by the historical circumstances, however. Rather, its universality is established by its potential universal availability: the style materializes wherever its constitutive visible features are replicated, regardless of the context (Wollheim 1995, 46;Davis 1996, 173). Any style, no matter how locally restricted its incidence (including if only one individual producer has been involved or if only one specimen of the style survives), is effectively treated as universal, when it implies a universal community of producers and observers. Because its principles of visual configuration can be abstracted from its historical context, anyone applying or recognizing the principles is a member of that community. It follows that universal style is collective not because its instantiations are identified with a historical community of producers, but because, in principle, anyone can work in it.
Universal visual style analysis strives to infer from the instantiation of visual styles their non-stylistic context. For such an inference to be possible, it is not enough that universal visual styles be potentially universally visible; they must also be non-conventional. 4 This characteristic means that a universal style's implementation is not based on purely conventional grounds, but rather creates, or is motivated by, its relation to context. So, while there may be universal styles that are conventionaltheir instances do not indicate in themselves anything about their contextthe only kind of universal style that carries an explanatory force in a socio-historical inquiry is the non-conventional kind.
It would be difficult to find anyone post-World War II explicitly maintaining, in line with Loos, that wherever a certain visual style obtains, it necessarily correlates with particular collective constellation. 5 This would amount to defending a version of universal visual style analysis that sources extra-stylistic information strictly from style in a deterministic fashion. As Gell (1998, 216) puts it forcefully, 'one certainly cannot argue […] directly from stylistic properties of artworks (such as bifold symmetry) to properties of socio-cultural systems'. The most obvious reason why one cannot infer context from the style in this fashion is that replications of style do not necessarily correlate with reproductions of context: a style may survive its original context or spread to other, radically different contexts (Davis 1996, 187;Descola 2006, 170-171). In order to remain immune to this criticism, a viable explanation would be needed for how and why a universal style necessarily issues from or signals a particular non-stylistic context. Lévi-Strauss's (2006) interpretation of split representation probably comes closest to the Loosian position. As originally described by Boas (1927, 221-231) for Northwest Coast groups of North America, in split representation the subject of depiction is split and projected onto a flat surface so that the front view is represented as composed of two identical profiles ( Figure 6). Split representation thus represents a sub-class of vertical bilateral symmetry. Lévi-Strauss understands split representation more broadly as a dislocation of a three-dimensional subject into elements that are put together again on a plane following arbitrary rules. 6 He claims that the violence of disfiguring and then reconfiguring the subject of depiction in split representation visually representsor visually negates (Lévi-Strauss 1961)strict supernaturally sanctioned delimitations of human culture from nature, but also of various social stratifications within the human world. In short, split representation gives visual expression to a socially rigid caste structureor, on his alternative account (Lévi-Strauss 1961), it provides an imaginary overcoming of the caste structure.
In contrast to Loos, Lévi-Strauss does not infer such universal social implications of split representation simply from the stylistic properties and in total disregard of any ethnography. He relies, rather, on his field research among the Caduveo of the Amazon, and the work of other scholars studying visual cultures employing split representation (Lévi-Strauss 2006, 56). He then looks for common aspects of their social structure and finds it in a 'chain of privileges, emblems, and degrees of prestige which, by means of masks, validate social hierarchy through the primacy of genealogies' (69). Lévi-Strauss's hope is that a valid style analysis will contribute to, and in cases when little other information is available, even provide a major means of social analysis: Even if we knew nothing about archaic Chinese society, an inspection of its art would be sufficient to enable us to recognize prestige struggles, rivalry between hierarchies, and competition between social and economic privilegesshowing through the function of masks and the veneration of lineages. (69) This passage is, in fact, as close as Lévi-Strauss's rhetoric comes to Loos's. The details of his argument, however, suggest a more moderate position. He first reads from the ethnography of the Caduveo to split representation, just as Hanson does with respect to Maori style and Gell with respect to Marquesan style; second, he abstracts from the local context universal principles of split representation; third, he takes its manifestations as pointing to a particular social structure which it is a functional part of. The claim is not that the presence of split representation signals necessarily the presence of a rigid caste society, but rather that its presence increases the likelihood of such a structure, especially when combined with other evidence such as the apparent reliance on masks in ritual.
Neither of the two first strategies described above relies on inference from style's character to its context. The first strategy does not purport to draw from stylistic configuration anything apart from establishing matching visual taxonomy for objects with the same history of production. The second strategy builds on non-stylistic evidence and looks to style's character for its corroboration or extension. It is only the third strategy that passes muster: stylistic configuration (not its accidence or variation) is to provide an independent access point to context.

Universal style and instrumental structure
Universal style is this essay's prime subject: if stylistic features are to provide an independent source of insight into context, not one derived from or corroborating it, they will have to be universally recognizable regardless of context. Universal style differs from typical notions of collective style, such as the style of a school, a nation, or a period (Wölfflin 2015, 88-90), as long as the identification of these depends on tracing a common history of production. The typical approach to establishing a non-conventional relation between style and context follows a general principle that whatever historical or ethnographic knowledge such analysis claims to provide, it does not concern any general purposefulness artifacts (of a particular style) have. In other words, while one may learn much about a community starting from the appearance of its toolsbased on what these tools seem to be used forthis needs to be distinguished from learning about that community based on the style of the tools. When the nineteenth-century adherents of the expressivist paradigm like John Ruskin and Hyppolite Taine looked with self-assuredness to the exterior of the Doge's Palace or the interior of the Church of the Gesù for genuine and reliable expressions of the changing mores of the Venetian Republic or the Jesuit Order respectively, they were not interested in what the architectonic features revealed about the buildings' prescribed functionality; they believed, rather, that their appearance somehow tracked the moral substance of their times (Haskell 1995, 304-362).
On some accounts, driving a wedge between instrumental structure and style is untenable because the very instantiation of such a structure is unavoidably stylistic and distinguishing a stylistic from an instrumental feature is therefore impossible (Hodder 1990, 45;Neer 2005, 5). But one can grant that there is no clear-cut distinction between style and instrumental structure and still maintain that describing, say, a clay vessel as a knee-high curved container that can hold liquids without capsizing and be manipulated by a single grown-up human tells us little about its style, although it arguably tells us a lot (sometimes all we need to know) about its visible instrumental structure. Compare this description with one that focuses more on, say, its particular curvature, color shade, texture, decoration, and its mode of delivery. Even if some of these aspects affect its functionality, a taxonomy that will group the clay vessel with others based on such aspects will readily be recognized as stylistic. Furthermore, an interpretation that will use this taxonomy to infer historical knowledge about the producing communityeven if only to identify it as this-style-of-vessel-peoplewill be distinguishable from an interpretation that draws inferences from the instrumental structure about, for example, the agriculture or economy of a society in which such a structure occurs.
To classify objects based on their instrumental structure is to be involved in what has been described as 'artifact hermeneutics' (Dennett 1990). Some artifact hermeneuticians claim that the function of an artifact is determined by makers' intentions (Vaesen and Amerongen 2008). Others argue that the function is determined by what the artifact is optimally suited for (Dennett 1990). Yet others think that it is determined by what the artifact type has been selected for (Eaton 2020). Regardless of the interpretation, artifact hermeneuticians stress analyzing artifacts' appearances for clues to their general purpose (say, to hold liquids or solids, to be used for cutting or stabbing, as a garment, as a blanket, and so on). This explains why in their accounts imagined or real archaeological scenarios of inferring from structure to function receive such a prominent place as paradigm cases (Dennett 1990, 182-184;Vaesen and Amerongen 2008, 787-791;Eaton 2020, 36-37). Indispensable for archaeological field work but little theorized (Neer 2005, 3-8), the practice of inferring from structure to function is predicated on the sound assumption that artifacts often publicly signal their function through their structural properties (Thomasson 2014): by being so structured, objects both become functional and signal their functionality publicly, that is, in ways familiar to target users. 7 The archaeological heuristic requires also that they signal their functionality in a way that is accessible to those who may not be fully initiated to the specific cultural norms of use (for example, archaeologists). Arguably, it is accessible because their morphology is perceived as analogous to the instrumental structure of artifacts of known or universal purpose (as is generally the case with the vessels).
But such inferences from analogy cannot provide logical certainty and are always under suspicion for anachronic projection. They can, however, be strengthened, as Wylie (2002, 150) has argued, by incorporating the form of the artifact under investigation into a broader cluster of analogous co-occurrences. If successful, such a move increases the likelihood of similar purposes. One may also stress the extensive similarity between the structure of the artifact under investigation and the structure of the artifacts with established function (Wylie 2002, 150). The tentative and partly speculative nature of such endeavors just comes with the territory of artifact hermeneutics: the involvement of local ontologies 8 and external observer's biases 9 (of the kind Wylie has sought to neutralize) creates a constellation in which any claim to universalitysuch as 'this instrumental structure signals universally this proper functionality'can only aspire to what I want to call a weak universality. Weak universality in effect relies on pragmatic abductive reasoning about what tends to be the case when certain conditions cooccur rather than on iron-clad deductive inferences. Any artifact hermeneutics identifying universal instrumental structures should be content with such a weak universality lest it fail to take into account that reproductions of structure often survive the prescribed functionality motivating their shape. 10 It may be illuminating to ask why the assumption that public artifacts tend to provide clues of their general function by their structure and thus provide a (weak) universal access point for artifact hermeneutics fares better in this regard than the suggestion that a collective visual stylea collective mode or type of visible configuration or deliverycan be universally and non-conventionally indicative of its context, that is, can be the subject of a stylistic hermeneutics, or, of what I have called universal visual style analysis. Why can't we model such an analysis on artifact hermeneutics?
A major difference between the use of stylistic analysis and instrumental structure analysis in socio-historical inquiry is that it is presumed that whatever purpose an artifact is mandated to serve should play no part in explaining what its style can tell us about its context. Classifying something as having been done in a collective style is assumed to be a different cognitive act than classifying it as having been done to serve a general purpose. While it is recognized that an important role of visual style is to individuate instrumental structure, this stylistic individuation is often treated as arbitrary, random, or 'isochrestic'it is subject to individual and locally sanctioned ways of making, more or less equivalent as to their utility value (Dunnell 1978;Sackett 1990). Others have correctly objected that such a local style may become recognized as typical both within and outside the community of makers and perhaps be elevated to a distinguishing mark of the community and consequently acquire various symbolic meaningsa 'passive' style turns 'active' or 'emblemic' (Wobst 1977;Wiessner 1983, 257-258;Sackett 1990;Summers 2003, 64-66). Recovering these conventional meanings, however, cannot fall within the purview of universal visual style analysis as long as nothing in the visual style itself establishes the symbolismas long as it is taken for granted that 'the precise form of a style has no function' (Richerson and Boyd 2005, 249). 11 From this view, universal visual style analysis cannot relate style to context in instrumental terms, as it is only within the purview of artifact hermeneutics to infer from structure ('precise form') to function, even when very little is known about its context beforehand.
What remains for universal style analysis is to rely on some kind of universal stylistic semiotic: an object rendered in a universal non-conventional style signals to the analyst its particular role and place in its social environment independent of whatever instrumental structure it may possess. To explain the correlation between universal style and context in both non-instrumental and non-circular terms, however, has been a tall order. Critics have observed that what underlies or motivates the mapping of a style onto a context has never been convincingly explained or made clear (Gombrich 1963;Wollheim 1995). The criticism certainly applies to Lévi-Strauss's writings on split representation; Gell's observations on Hanson noted above are also a thinly veiled attack on this aspect of Lévi-Strauss's structuralism.
It may at first not be clear why the idea that style analysis can provide an independent source of insight into an artifact's non-stylistic context must always take the form of identifying non-instrumental links between collective styles and underlying social relations. Why not simply ground the non-conventionality of a universal style in its making visible the universal instrumental structure of artifacts? An object rendered in a universally visible, non-conventional style would draw its universality and non-conventionality from its role in mediating instrumental structure. This answer, however, comes dangerously close to collapsing the difference between the social analysis of instrumental structure and the social analysis of collective style. To come back to my example, most of the vessels in the world can be recognized thanks to their common instrumental structure (a curved container that can hold liquids without capsizing and can be manipulated by a single grown-up human). But it would be counter-intuitive to call this instrumental structure their universal style. If anything, a style is a mode of realizing instrumental structure. And since the ambition of isolating universal style was to provide sociohistorical inquiry with an independent source of insight relying on stylistic properties, this approach serves to disqualify the ambition more than anything else.
To sum up this section, a public artifact often makes its instrumental value universally visible by its non-conventional structure and this structure's particular instantiation is subject to stylistic variation. If universal style analysis is to contribute to socio-historical inquiry, the stylistic variation cannot be always completely conventional, random, or arbitrary. Either the source of a style's non-conventionality is non-instrumental and then something like a universal principle of style-context correlation appliesa contention that has been regarded with great suspicionor its non-conventionality is grounded in instrumental structure and the difference between instrumental structure and universal style collapses.

Universal style as instrumental structure
For a universal visual style hermeneutician there are two potential ways out of this impasse. One can pursue the path of a universal stylistic semiotics, arguing for the existence of stylistic principles whose application counts towards co-occurrence of certain contexts. This would bring up the need to explain the co-occurrence. The other path is to associate universal styles with instrumental structures, which creates the problem of keeping universal visual style analysis different from artifact hermeneutics. It has been noted already that criticisms of collective style analysis in socio-historical enquiry affect primarily the first option. Whatever its chance of succeeding (and I take it to be slim), we need not explore this here. Instead, what remains of this essay will be devoted to the second pathway, that is, the possibility of grounding universal style's non-conventional character in instrumentality.
For social or historical enquiry, what I propose is to treat universal style as universal instrumental structure (structure likely to have been developed/selected/intended for certain general functions or purposes). Just like an instrumental structure, a non-conventional universal visual style is a means of satisfying a need or desire while signaling publicly the mandate to be so used.
The advantage of this approach becomes apparent when confronted with what may be labeled as the epistemological problem of universal visual style analysis. Social enquiry relying on universal visual style analysis is predicated on the possibility that there are collective styles open to analysis from an observer perspective. But what cognitive capabilities does the hermeneutist employ to recover a style's context? A collective style may be 'active', that is, epistemically effective within a participant ('emic') perspective: it is intended to show or prescribe certain patterns of thought or behavior to those with the mandate to observe the artifacts. Or it may be 'passive', meaning, it is not thematized by the participants, but it may provide epistemic access from an observer ('etic') perspective: showing to the uninitiated certain ways of the participants' world (Sackett 1990). Sometimes, a style is taken to reveal different things to participants and different things to observers, or to be effective on different levels of awareness. For example, a collective style could be claimed to be effective precisely because its proper role is mis-recognized by participants; it is for this very reason more susceptible to analysis by those immune to its effects and experiencing them sideways-on, so to speak (e.g. Forge 1973, xviii-xix;1979, 284). On the non-instrumental understanding of the style-context relation, the stylistic hermeneutist ought to be clear about the channels of stylistic communicationto whom style reveals the context and how the hermeneutist tunes in on it. For the instrumental approach, the problem does not press itself with any great urgency, as the solution is more or less self-evident. To whom is style as instrumental structure supposed to reveal or prescribe contextual information? To potential users, it reveals or prescribes norms of use because it makes an artifact's instrumentality visually salient to them, in other words, it makes it public. And because it is publicly visible, the instrumental structure potentially provides social scientists with an entry point into practices of artifact production, use, and distribution.
The main disadvantage of the instrumental approach is that the difference between style and instrumental structure is in danger of collapsing. Luckily, the difference does not need to evaporate. A difference between style and structure comes in handy when one needs to address non-conventional shifts in morphological variation within an identified instrumental structure. When a universal instrumental structure is identified, universal styles will be sub-classes of the instrumental structure tracking morphological variations. Such universal styles will themselves be nonconventional, if their variations are deemed as affecting instrumentality they comprise instrumental styles, in other words. For example, some Acheulean bifaces, produced by our hominin ancestors, are symmetrical beyond need; some are either too big or too small, or just too oddly shaped to be used as hand-axes (Figure 7;Mithen 2003;Currie 2011). Such morphological variations may affect instrumentality and invite speculations about what the structures could have been instrumental for. Whatever the answer to that question, for my argument here, the shifts in morphology within the instrumental structure are properly called stylistic. There is no universal, non-conventional style without instrumental structure. Furthermore, each such style becomes effectively an instrumental structure, when further non-conventional stylistic variations are identified within it. The distinction between instrumental structure and a universal non-conventional style is thus a matter of scale; it is relative to the taxonomist's perspective, but with the proviso that it is a perspective that can be allowed only to zoom in or out within a taxonomic system tracking instrumentality of structure/style. In other words, the taxonomist cannot introduce arbitrary parameters as to what counts as a structural/stylistic feature and what does not. This suggestion has two major advantages. First, by recognizing those features that vary within an instrumental structure class as stylistic, it honors the intuition that instrumental structure analysis differs from stylistic analysis because it works on a different scale. Second, by associating universal non-conventional styles with instrumental structures, it steers clear of the serious problems plaguing non-instrumental explanations of style-context correlations. To show the advantages of this approach, I will apply it to the case of the Sword God. The discussion will be brief for reasons of space and fuller elaboration is planned elsewhere. What I hope to demonstrate, however, is a new direction for style-based research and to reveal what the dominance of non-instrumental approaches to visual styles has pushed to the margins.
The most evident instrumental feature of the Sword God is its figurative content. Imageswhether two-or three-dimensionalserve the general purpose of conveying their figurative content (e.g. Hyman 2012); and this general purpose is inscribed in their very configuration. We may not know what the Sword God refers to, but we recognize it as a bearer of figurative content. In other words, we recognize its universal instrumental structure of an image.
To discuss its universal structural/stylistic features, I will therefore focus on its pictorial nature. My notion of its pictorial style is based on the general characterization of visual style provided earlier, and it corresponds to a general mode of pictorial configuration instantiated by a particular image (the Sword God). We want to know how the Sword God image sheds light on its context, and arguably, to already characterize something as an image is to identify its instrumental structure: the nonconventional nature of its morphology is rooted in its instrumentality for conveying figurative content. To describe the condition of having figurative content as a pictorial style is, however, strongly counter-intuitive, just as it was counter-intuitive to treat the condition of being a vessel as enough to classify its visual style. Stylistic variation is a variation within an instrumental structure. And when this variation is also instrumental, we have identified a universal non-conventional style. In the case of the Sword God, a good candidate for such a variation is the vertical bilateral symmetry as its organizing principle. The central question then is, how its implementation affects its instrumentality.
For this purpose, I want to briefly make use of Summers's (2003, 349-353) treatment of what he terms 'planarity'. I turn to Summers because his brand of art-historical postformalism is developed in explicit opposition to the expressivist paradigm of inferring from style to context (32-34; see also Summers 1989), yet he does not shy away from speculating about context based on artifacts' purposeful configurations. 12 According to Summers, the precondition of any image-making is the ability to perceive a facing surface of an object as notionally planar (roughly: as a geometric plane). Some images make this condition explicitas is the case with images employing vertical bilateral symmetry (Figure 8). Explicitly planar images have their figurative content ordered in such a way that it conforms to the maximum with the planar uniformity of the image surface: the content is schematically spread out onto the surface (or virtual plane). Summers argues that explicitly planar images, such as those following vertical bilateral symmetry, have historically and globally opened the way for making images semantically charged beyond conveying resemblances. As opposed to mere tracing of outline shapes (as in Paleolithic cave paintings of animals), explicit planarity invites the schematization and hierarchical structuring of the figurative content as well as the conventionalization of meaning, such as when a pubic triangle of a Paleolithic Venus lends itself to be abstracted as a symbol of fertility (Summers 2003, 346-349). An explicitly planar ordering is also, according to Summers, equally suitableand has been developed globallyto enhance the sense of authoritative, effective presence of the depicted subject. As each and every part of an explicitly planar image addresses a point of view perpendicular to the surface, it demands a humanly impossible viewing position (Summers 2003, 350-353; see also Hagen 1986, 116-176). Such an explicit subordination of figurative content to a geometric planar ordering would serve the manifestation of the image's independence from a particular viewing angle and help create a sense of overbearing presence.
Vertical bilateral symmetry of the Sword God is a clear example of explicit planar imagery; it is a 'double planar representation' (Summers 2003, 372). It explicitly schematizes figurative content around a vertical axis to render the mirrored design values mutually equivalent. Applying Summers's proposal to the interpretation of the Sword God, vertical bilateral symmetry would be introduced to stress the symbolic features of the figurative content (see also Morphy 1977) or to make its figurative content's presence more commanding. The former, symbolic interpretation finds indirect support in the fact that the Sword God lacks any hieroglyphic identification, making it 'unusual among the large figures of Yazılıkaya' (Alexander 1986, 62). 'Apparently,' conjectures Alexander, 'he names himself', suggesting that the relief itself might have served as a pictographic symbol. This conjecture is in line with Summers's reading, for the Sword God's bilateral symmetry increases its potential for symbolic use. The latter interpretation is consistent with the spatial aspects of the relief's format. With its 3.38 m, the Sword God towers above the spectator who is confined to a narrow cleft just about three meters wide. The socially sanctioned desire the relief would meet could thus be described as one of creating a sense of a fuller, efficacious presence of the figurative content, translating the social authority of depicted subjects into the visual authority of images.
Here thus is the promise of a universal visual style analysis: by identifying an artifact's appearance as exemplifying a universal style, it opens access to insights about its socio-historical context. I have proposed that this identification be understood as isolating a non-conventional, instrumental variation within an instrumental structure. A style's instrumentality links it to context: vertical bilateral symmetry aligns figurative content explicitly around a vertical axis in a process of schematization that makes the image prone to symbolization as well as helps establish the figurative content's commanding presence. Identifying this non-conventional variation of an instrumental structure (image) in the Sword God does not determine beyond any doubt that the relief actually served either or both of the functions. In accordance with the principle of weak universality, the variation's implementation counts towards a particular context of use, especially when combined with other evidence such as the relief's spatial coordinates and the absence of hieroglyphic designation. This other evidence does provide some further justification as to why one would want to implement vertical bilateral symmetry in this particular instance. But what made this line of interpretation possible in the first place was inferring from the Sword God's style to its context. 13 Funding This work was supported by Deutsche Forschungsgemeinschaft [grant number STE 2612/1-1]; European Regional Development Fund [grant number CZ.02.1.01/0.0/0.0/16_019/0000734]; NOMIS Stiftung.