I’ve been thinking a lot about landscapes recently, especially after a spectacular flyover in July above the southern tip of Greenland. I looked down and could instantly see small ice floes, rocky beaches, extinct volcanoes, blue-green lakes, and glaciers. There was no evidence of humanity, at least in that small part I could see of the country (pop. ca. 56,000). I wondered where people might live, where they might fish, hunt, build a place to live. I also thought about where they might shop, where they might work, and scanned for clues far below.
I feel similarly when I play No Man’s Sky (Hello Games) or other games where I can cruise over the landscape in a shape or on a mount, watching it unfold as I make decisions on where to land based on the quests/missions that I have. It’s landscape archaeology even though it’s in a synthetic world. Any landscape archaeologist must create a research plan, must then select an appropriate site in which to do the work, and then conduct a variety of activities ranging from aerial reconnaissance, remote sensing, shovel tests, fieldwalking, finds analysis, and more.
In order to better understand how landscape archaeology works in synthetic worlds (largely digital games, but this can be applied to all digital built environments), we need to understand the definition of a landscape and the varieties of landscape archaeologies, all of which shift neatly between the natural and the synthetic. I’m using the terms “natural” and “synthetic”, following the lead of Edward Castronova, an economist and pioneer in understanding the relationships between people and digital spaces. Video games do not present a virtual space. They are very real environments that evoke real emotional responses, and contain robust, mature economies and societies both human and non-human. “Synthetic” acknowledges the reality of these spaces while also noting that they are not “natural.” Synthetic worlds are designed, the creators and their blueprints/code known. For the digital landscape archaeologist then, we can explore these spaces asking exactly the same questions as our natural-world counterparts. I’m willing to bet that what we find will not be that dissimilar.
In the Preface to 2008’s Handbook of Landscape Archaeology. (Left Coast Press), editors Bruno David and Julian Thomas identify three broad themes of landscape archaeology (p. 20):
- Landscapes are fields of human engagement as in Heidegger’s notion of dwelling. These include both explorations on conceptual ways of approaching, and experiences of landscapes as fields of engagement, as the “in” of “being-in-the-world.”
- Landscapes are physical environmental contexts of human behavior (such as investigations of the tree cover or topography of the environments).
- We should also reflect on representations of landscapes, such as in landscape art, or the identification of colonial tropes in landscape analysis of textual preconceptions.
Both natural and synthetic landscapes share these three themes: 1) For synthetic worlds, players are very much “in-the-world”, completely engaged with their surroundings as they explore and play; 2) synthetic worlds do provide physical environmental contexts, which can include understanding vegetation and topography that can (and often do) dictate player behavior; and 3) synthetic worlds very much provide representations of landscapes, typically created by one or more designers, based on what they conceive of landscapes appear, as well as how they advance narrative or drive player behavior.
When we engage with synthetic worlds, we occupy a place within a place within a place, inhabiting multiple landscapes at the same time. We sit in the natural world before engaging with the synthetic. And once inside the synthetic landscape, we visit multitudes of places dotting our field of vision and beyond. The landscape beckons us in, and affords us new frontiers.
Landscape archaeology seems to be preoccupied with human presence and engagement within the spaces they occupy over time. It doesn’t matter if that space is natural or synthetic. Synthetic worlds have nearly infinite space with which players can engage, creating lived-in environments where players can spend more time than in the natural world. Players inhabit these spaces, invest time, money, and resources, and create their own culture side-by-side with any synthetic culture pre-built by developers and/or algorithms. Some synthetic worlds (e.g., Minecraft and No Man’s Sky) offer players the chance to manipulate the landscape themselves for whatever reason, be it for entertainment, or for something more practical. Create a real or synthetic garden. Dig a hole to see where it goes and what resources might lie beneath your feet.
Every landscape (natural or synthetic) can contain material remains and networks connecting agents, materials, and places. A goal of the landscape archaeologist is to connect the two. While this can be easy in designed games (synthetic worlds where everything has been crafted by the software developer), it becomes increasingly more difficult (and arguably more “real”) in worlds where everything is procedurally generated. This is especially true when algorithms are entrusted with the task of building a world and populating it with things for the archaeologist to discover. This placement is almost random, but bears some logic based on the landscapes discoverable within the synthetic world. The landscape shows the archaeologist where to look.
A landscape, either natural or synthetic, can be interpreted in any number of ways. A synthetic world’s designers interpret their digital built environments from the perspective of narrative, player engagement, and aesthetics. From the player’s perspective, the landscape exists as something to explore, as something to be traversed, and as a provider of materials for quests and recipes. In the natural world, one must also consider the landscape from the indigenous perspective(s), understanding how the landscape is a permanent resident, a host of memories, and perhaps a space of spiritual resonance. How do non-human denizens of synthetic worlds “perceive” their landscapes? For now, automatons unquestioningly operate within the constraints of a world’s rules. They “see” the landscape through software instructions. They engage with that space based on parameters, acting when acted upon. It’s little different than a human commuter in a rut.
Inhabiting synthetic landscapes can, over time, create a material record of that occupation through in-world crafting and building, through rubbish, but also through documentation of what came before. Players record and photograph their creations, share them, save them. It’s a rich record of space and time that the natural world will never see. It is perhaps easier to illustrate synthetic landscapes because they can be recorded during exploration, and can even “instanced” for exploration by others on their own hardware, helpful if one wishes to have results independently confirmed by others, something not normally possible in field archaeology, which can be a destructive process leaving only data but not the primary source(s) of that data (i.e., strata).
Speaking of visualization, how does the archaeologist-player map the landscape of a given synthetic world? Some games come with built-in maps, while others tease out their mapping only through a player’s progress. Other synthetic worlds have no maps at all, so players may either hand-draw their explorations, or rely on memory, the latter of which most non-human entities do. In many games, players rarely see their world from above, but frequently at eye-level. We remember how to get from one part of the forest to another through experience, or from following something, curious to see where it goes. We create a mental map, which is visual, at times aural, occasionally punctuated with memories of events that happen along the trail or traverse. We learn the game’s rules for wayfinding in the synthetic wilderness, and then lead others to what we’ve seen, or we write about it so others can follow in our footsteps. Some of these writings are literally called “walk-throughs.” With an active player community (or a preserved collective memory), we also have what can equate to “tavern archaeology” in the natural world. Yannis Lolos wrote about this Land of Sikyon, about approaching local Greeks in taverns and other gathering spots to share a drink and listen to stories. At times these stories would provide clues to where to look for previously undocumented archaeological sites, which were then compiled into a gazeteer for future archaeologists to use as they continue to understand the history of Sikyonia. Video game archaeologists should do well to spend time with player communities on reddit and elsewhere for the same reasons as Lolos. We can learn from indigenous populations, can work together with them, and can share what we find with them.
When an archaeologist explores the landscapes within a synthetic world, they notice both presence and absence in the archaeological record. The archaeologist applies filters, scanning for vegetation, animal life, water and other topographical features, and evidence of current or past use and occupation. One landscape flows into the next, synthetic biomes bleeding into each other. The archaeologist looks for boundaries, for patterns, whether confronting the landscape head-on, or floating above it in a flying vehicle or in a disembodied “photo mode” provided by the game. The archaeologist can ask why some spaces are neglected and others used, and if used, how frequently and why. It is a search for an underlying algorithm or a set of landscape behaviors as designed. In most synthetic worlds, landscapes are very much characters and have their own rules on how to act, which includes weather-making and resource-production. Landscapes in synthetic worlds have personalities, or at least that is how players perceive the spaces in which they occupy.
There is a haptic necessity to understanding landscapes. In order to write about landscape in an archaeological way, it must be experienced first-hand by the archaeologist. What applies in the natural world is also applicable in the synthetic. Christopher Tilley articulates this in “Phenomenological Approaches to Landscape Archaeology.” The more we experience a landscape in person, the more we understand its rules, which include formation processes and use by both human and non-human actors.
In “Object Fragmentation and Past Landscapes,” John Chapman writes that “landscapes consist of a network of places, some natural, some culturally constituted, some created by human manipulation of the landscape. It is this network of places that gives human lives their meaning, through an identification of past activities and present embodiment.”
This also describes synthetic landscapes and the networks that run atop them: some created by human manipulation (players create networks through their own agency in a synthetic world), some culturally constituted (these networks are created in advance and by design prior to player-arrival), and some natural (emergent behavior independent of player or designer agency).
So what’s in a landscape then that immediately grabs the archaeologist’s eye? Architecture is a start, and is one of the things most easily recognizable in synthetic worlds. Players expect architecture, and most games deliver. As part of the synthetic landscape, architecture behaves in the same way as its natural counterpart: it is recognizable, is used/re-used, assists in wayfinding, and contributes to the understanding of a place’s present and past. This raises an interesting point regarding determining the age of something in a synthetic world. How do we do this?
Dating landscapes in synthetic worlds varies greatly from those that are natural. This is perhaps the greatest difference between the two. In synthetic worlds, there is no real stratigraphy, and no real superposition of layers of earth. For most synthetic worlds, there is no geologic time. The spaces just sit there not eroding, unless you count “bit rot,” the slow decay of underlying data. The archaeologist is left with software version and build numbers, which are tied to absolute dates in the natural world. But inside the game, landscapes can be made to look old (or to imply age), even though the game itself has only been playable for a few days. Even playing older games results in landscapes that look exactly the same now as they did in 1983. Little, if anything, changes.
Regardless of when they were created/accessed, these synthetic landscapes in games can provide insight into how people (and their things) move from place to place, offering glimpses into what players (or non-human agents) carry with them, use, sell, or discard. In observing in-game habits of players as well as coded entities, especially in games where the landscape can vary (or can be varied), the archaeologist can begin to experiment with agent-based modeling (ABM), making modest (or radical) changes in order to run experiments to see what happens. In some cases, landscapes could be modeled after those in the natural world to assist in answering archaeological questions about earthbound places and the people who used them.
How does an archaeologist conduct landscape archaeology within a synthetic world? It’s not terribly different from working in the natural world, as articulated by Thomas Richards in his article, “Suvey Strategies in Landscape Archaeology”: first develop a research question. Once a research question is developed, a suitable landscape must be selected. Research problems often require a landscape to be representative of a larger area, so the distributional patterns of the surface record and associated human behavioral interpretations can be widely extrapolated. Following the selection of a regional landscape study area, the formation of its present surface characteristics needs to be considered in developing a survey strategy: geomorphology and vegetation (which affects visibility of the ground surface). Try to determine human land-use history, particularly large-scale ground disturbance activities. Determine the intensity and coverage of your initial survey. Decide on the scale of the survey and what (and how much) to sample. Decide on whether or not to conduct remote sensing or invasive subsurface testing. How will you record this data? Will there be follow-up surveys allowing for more detailed collection and analysis? At the survey’s conclusion, can you develop any kind of predictive model for where one might find other artifacts or settlements?
The survey and research plan points are applicable to the archaeological investigation of synthetic worlds. Any archeological fieldwork much be driven by research questions and a plan on how to answer those questions. For my online archaeological survey, I needed to find a site that extensively used procedural content generation, which would perhaps exhibit evidence of “machine-created culture,” material remains created and organized by algorithms that would populate a synthetic landscape. No Man’s Sky seemed to be the best fit for this need. I wanted to check distribution of sites and artifacts, to see how buildings and landscapes interacted, to look at animals and vegetation and their relationship to the landscape and these structures, and to see how these manifested in the nearly infinite worlds that could be explored. Barring a multi-player option (which would materialize in a basic way one year after the game’s initial release), I had to conduct my surveys as a lone fieldwalker and as a lone pilot. I used different levels of granularity in my surveys, less detail when flying (counting structures and times between them), and more detail when on the ground (noting artifacts, brush density). Future surveys can be undertaken in synthetic worlds that actively encourage guilds (groups of players) who can survey together in the same landscape on the same server. This will also allow for specialization. Depending on the synthetic world, we can sample and “shovel test.”
I mentioned non-invasive survey techniques above. Paul Cheetham covers this for natural worlds in “Noninvasive Subsurface Mapping Techniques, Satellite and Aerial Imagery in Landscape Archaeology.” Prior to starting a non-invasive survey, consider: survey objectives, archaeological questions, previously remotely sensed evidence and results, current land-use, former land-use, underlying solid and drift geology, other local geomorphological and topographic factors, degree of access to the land, time, money, personnel, and equipment available for the survey.
Depending on how the synthetic world is constructed and what it allows players to do, it can be possible to conduct a variety of remote sensing tasks. For example, in No Man’s Sky, it is possible to orbit planets like a satellite prior to penetrating atmospheric cover for high-altitude overflight, followed by high- or low-speed travel a few meters above the deck, giving the surveyer different kinds of information about what’s on the surface of the landscape, and if there are any patterns to be observed. Once on the ground with the proper tool equipped, one can conduct a rudimentary form of remote sensing, which produces icons showing players where to dig, and occasionally revealing underground caverns and other geologic formations. Each synthetic world has its own rules of engagement, which must be followed by players in order to successfully survey these landscapes. Depending on the hardware and platform used to host these worlds, however, it may be possible to create modifications (mods) of tools and other equipment, which allows for bending the rules of a world in order to conduct aerial imagery or remote sensing. This raises ethical questions about whether modifying a world affects how archaeologists see the world, and how non-human agents operate within that world once it has been changed from its original state. Mods inject an additional level of complexity into a synthetic world, which might have unexpected/unintended results. At the same time, however, the archaeologist might be able to create a more useful digital toolkit for exploring these digital built environments, instead of relying on more traditional methods that might not be the best-suited for the task. To mod or not to mod recalls the decision to dig or not to dig, to use an invasive approach on the landscape in order to retrieve information about it.
Throughout this post, you have likely been thinking about landscapes in games such as the ones featured in Elder Scrolls, Dragon Age, World of Warcraft synthetic worlds or similar, landscapes that are fantastical yet feel real containing mountains, trees, lakes, things anyone would consider to be part of the landscape. I propose, though, that in order to have a landscape archaeology of digital games and other synthetic worlds, that the archaeologist must apply definitions and methods to any game or world, not just those that are facsimiles of recognizable natural environments.
Every game is a landscape. There is an architecture of space on the plateau of the screen, which can be measured and engaged with, that utilizes time and location in order to function. Every definition, every principle of landscape archaeology applies to games specifically, and more generally to software—anything with a graphical user interface (GUI). We are working on a micro-scale, but it is easy to tell when an application’s GUI nudges users down a path of activity, reacts to user-agency, offers up artifacts (glitches/bugs), and is the subject of communities and cultures (e.g., Apple v. Microsoft). Most of us belong to at least one digital tribe, and we interact with that software as an indigenous population, as developers, or as users who made their pilgrimage late.
One final thought: all of the above about conducting landscape archaeology (or archaeology of any kind) within a synthetic world might sound daft, especially when we know that these are all designed environments. Think of these, however, as a proving ground for ideas on method and theory, testing on software we know that is well documented. I predict that by 2020 we will finally see video games set in completely procedurally generated worlds where the cultures that players encounter have never been considered by the game’s designer(s), instead created from a complex set of rules that, when mixed together, create emergent cultures distinct from one another. We’re already getting glimpses of these “machine-created cultures” (MCCs) in games such as Mark Johnson’s Ultima Ratio Regum, and more are coming. One day we’ll have a Turing test for cultures to determine what is real. How will we determine that level of reality, and if a new, born-digital culture thrives, what obligations do we have to interacting with it, and ultimately to preserving it?
—Andrew Reinhard, Archaeogaming