As I continue to working on my PhD, considering how to conduct archaeological survey and excavation within digital built environments (software generally, and more specifically games), I wanted to turn to primary texts authored by experienced field archaeologists who then summarized their work in handbooks for future archaeologists to use. My Ur-text (as recommended to me by my PhD committee) was Martin Carver’s Archaeological Investigation (2009), a well-regarded how-to guide for designing, implementing, and publishing archaeological projects. When I approached Carver for additional suggestions in 2017, he replied and referenced his own book as a starting point, and to contact him with follow-up questions specifically as they pertained to adapting his work to my plan of doing fieldwork within a digital space. I supplemented Carver’s book with Steve Roskams’ Excavation (2001), to see if I could apply his methods to “excavating” digitally. The danger of course is in trying to put techniques from the natural world to work in the digital, that this might cause me to overlook other ways for doing digital work digitally based on my research questions and the digital environment in which those questions would be asked and hopefully answered. As is shown below as I bounce between Carver’s work in the natural world and my work in the digital, there are a number of times when digital archaeology breaks with its natural world counterpart. There are, however, many, many similarities shared between these two spaces and how to conduct archaeological investigation in them. Note that throughout this text I will refer to the natural and synthetic worlds, the difference being that synthetic worlds are created digitally by people (or by algorithms created by people). Both the natural and synthetic worlds are real. There is no “virtual” here.
Why Survey/Excavate Digital Built Environments?
Carver begins with an explanation of why we should conduct archaeological investigation at all. It starts with “pieces of the past, life’s disjecta membra, the stuff. This is what we study . . . . Our first task is to appreciate why we have what we have. All these cultural remains belong to people who deserve a history, but they do not equally leave us one” (Carver 2009:7).
Digital history, however, is even more elusive, because outside of the hardware (most of which gets deposited in dumps or recycling facilities or ultimately African “disposal” sites), the software engaged with, and the history of use by communities is completely invisible, a kind of intangible heritage. We have what we have now because users have documented what they do, but this is done informally, often on ephemeral community message boards, groups, and chatrooms, with no guarantee of preservation.
When we consider surveying or excavating a site, those sites are often either forgotten or deserted, either buried some distance under the earth or hidden/ incorporated into the modern landscape. Carver writes that “even when archaeological sites are deserted, they do not entirely die. They have a long and varied afterlife . . . (Carver 2009:7).
This statement is equally true with video games (the main class of digital built environments that I am focusing on in my own research). Players can certainly return to old games, abandoned games, and also synthetic, shared worlds that are now devoid of players. The sites largely remain in the digital, much like they do in the natural world. When dealing with sites in the natural world, one must consider the pre-deposition of a site (subsoil, topography, culture) and post-deposition (natural and human attrition). Human activity after the abandonment of a site can include curate behavior, vandalism, stone-robbing, cultivation, digging; natural activity includes bacterial, chemical, vegetation, burrowing animals, frost, flood (Carver 2009:8).
We don’t really see these in abandoned digital spaces, but post-deposition activities do occur. These can include bitrot (degradation of interior code) and/or degradation of physical materials exposed to the elements (or just to time). Games are reliant on hardware on which to run, and older games also relied on physical media, which served as a catalyst between the player and what is played. Just as with any other material artifact, these materials change over time. Post-deposition, the digital is just as susceptible to the whims of the environment as the natural.
Carver identifies (Carver 2009:9, fig. 1.6) five stages for what happens to a human settlement in the natural world, each with its own factors and properties: before deposition, during occupation, at abandonment, the site gets buried, after burial. It is easy to see how these apply to traditional sites of human occupation, but how can they be mapped to the digital?
Use depends on landscape and environmental factors. A place has to be able to be occupied (or to communicate the ability that it can be occupied prior to people arriving and making use of that space). For games (and other software) one could consider this to be either the marketplace (demand creates supply, e.g., people want to play online with their friends, so companies such as Blizzard Entertainment create Massively Multiplayer Online Role-Playing Games (MMOs)), or people (or companies, which are groups of people) feel the need to carve out a niche for themselves, creating their own space(s) to inhabit. Once the space is occupied and developed, culture follows. Consider the popular MMO World of Warcraft, which has millions of players (4 million active accounts according to Blizzard’s February 2017 figures), has its own annual convention (BlizzCon), has fostered books and a feature film adaptation, and has licensed its in-game material culture to real-world companies to produce for players to buy in order to signal their love of the game to other kindred spirits.
The game is the structure inhabited by the players online, their activities preserved in massive databases hosted by the game’s producers. Over time, most games are abandoned in favor of newer titles, or players find other ways to use their discretionary time and income. Populations dwindle until the game is but a shell of what it used to be. Unlike sites of human settlement in the natural world, these game-spaces to not erode, are not “robbed” of resources, and do not weather. They remain in a kind of stasis, abandoned yet timeless, ready to be enlivened at any moment by hordes of returning players who never arrive. Following a game’s abandonment, the game can be buried. This happened literally with the case of E.T.: The Extraterrestrial (Atari 1982), where unsold/returned copies of the game were trucked to a landfill in Alamogordo, New Mexico, to be dumped to make room for new merchandise. In other instances, companies retire the games they produce. They no longer sell the games, and after a period of years, cease technical and community support. The games disappear, maintained in players’ memories, or archived by game developers or player communities. After a game gets abandoned and then buried, it can suffer a few fates: bitrot (see above definition); disappearance, where the game and its code are lost for all time; and archaeological excavation, where recovered games are studied through either play or deconstruction.
Designing an Archaeological Project Plan for a Synthetic World
Once a game is identified as a candidate for archaeological investigation, the researcher(s) must follow protocol in creating a publicly proposed project design. “The project design must be published before work starts, and not just because this results in a better managed programme, but for ethical reasons. . . . The project design itself must contain a programme of long- and short-term conservation as well as programme of research” (Carver 2009:33).
This statement is as true for synthetic sites as it is for natural ones. The project plan should be shared online publicly, especially with those user communities who actively play (or played) the game to be studied. This opens it up to community critique, in effect becoming public archaeology and engaging the games “indigenous” population, many of which will spot mistakes and pitfalls, or who might be willing to help in the research as volunteers. The plan should also identify conservation and preservation efforts for the game-site as well as for the research both pre- and post-publication. When I was organizing the No Man’s Sky Archaeological Survey (NMSAS) in 2016, I publicly broadcast the team’s reasons for investigating a synthetic universe, and how we proposed to do it. Within days of posting the research plan, I was contacted by several community groups interested in conducting citizen science within the gameplay. We were able to work with these groups, and to follow their own discoveries on various community bulletin boards online, and through social media. These communities also helped the NMSAS team revise its project methods during our survey period, correcting serious misconceptions about measuring distance on synthetic worlds. As Carver says, “archaeological project design is mandatory . . . . Researchers are bound by a contract with society (Carver 2009:33). In the case of games, the project design is a contract with society as well as with the player community.
Carver breaks down a field archaeologist’s research agenda into three parts (Carver 2009:47, fig 3.8): feildwork, objective, and outcome. What are we trying to accomplish with whatever it is we’re doing on-site? This is universal for both natural and synthetic sites. We excavate to see a sequence of use in order to confirm or change what we know about the site and its occupants. We survey in the site area to create a map of settlement and other features in order to determine where to dig (if we need to dig at all). We survey the area surrounding a site to map and identify other settlements in order to note changing land-use or to recognize settlement/cultural patterns. We study other areas in order to compare and contrast them with what was found on-site. We can follow these same procedures in any digital built environment. Software can be mapped and even excavated (as will be shown later in this thesis), compared to earlier or later versions of itself, and compared to similar software designed to match the needs of the same user community.
At the start of any archaeological project planning, preliminary reconnaissance is key. In the natural landscape, one must visit that space to experience it first-hand, to identify its challenges, and to begin to inform yourself of how to proceed based on operating in that environment. For the video game archaeologist, this necessitates ample gameplay. Playing a game (or using any kind of software) familiarizes you with the landscape. You play where others played, and the more time spent in that synthetic environment, the better prepared you will be to conduct an archaeological survey and/or excavation of it.
Surveying the Synthetic
Before any kind of excavation can occur, a survey must be conducted, either landscape or site (or both). There are three techniques of landscape survey: cartography, surface inspection, aerial photography (Carver 2009:65). In synthetic worlds, cartography can be done either with in-game maps, or by hand-mapping during exploration. Surface inspection can be done by the player on foot in much the same way a field-walking across natural landscapes. Many games feature flying drones, ships, or “mounts”, which allow players to see the landscape from above, and even to hover at various altitudes over the surface, using the computer or console’s native screen-grab features for aerial photography. Anything digital shown on a display (computer monitor, flat-screen, television, etc.) can also be considered as a frame for a map or plan, a top-down view that can be captured and measured.
The European Landscape Convention defines a landscape is an “area as perceived by people, whose character is the result of the action and intersection of natural and/or human factors” (http://www.coe.int/en/web/landscape). By this definition, games are landscapes. Players have agency in games and their actions mark the intersection of their decisions upon the game-world, and upon other players in a shared space. See my section on landscape archaeology in synthetic worlds for a more in-depth look at how games are landscapes that can be studied archaeologically.
Landscape survey combines geography, environment, and archaeology to explore the unknown world in deep time (Carver 2009:86). The same is largely true in digital built environments, but the question of time (specifically deep time) is tricky. In games, perhaps deep time can be measured by version numbers, but it will not appear graphically in a visual layer on a screen. Time in games is strange, with people playing in real-time, even though that game features the rapid passing of days and seasons while no one (or thing) ages, erodes, grows, decays.
Following landscape survey comes site survey. To Carver, a site is “an area of ground in need of investigation” (Carver 2009:89). Games also fit this definition. To Carver (Carver 2009:89), a site survey is simply a landscape survey on a smaller scale: the area is smaller, but the focus is finer. Surveyors sense archaeological features but do not damage them. A site survey can be invasive in its exploration (e.g., shovel testing), but these surveys attempt to leave everything intact. The objective of a site survey is to know as much about a site as possible before deciding what to do about it. The same works in digital spaces: observe, focus, and then decide next-steps.
Site survey techniques in the natural world include using maps and documents, topographical mapping, surface collection, geophysical survey, and sample excavation (Carver 2009:89). In a digital space, all games will allow for surface surveys and mapping, and some will allow for surface collection. Some might even allow traditional excavation. It all depends on the game’s mechanics, or what a game allows players to do. If actual digging in a game is not possible, the video game archaeologist must consider other ways of conducting synthetic shovel tests, should those be considered necessary. This is especially true when documenting material culture created by non-human actors/agents within synthetic worlds, placed there by algorithms. What can be determined through observation alone, and later through interaction (if ethics permit)? What do we as archaeologists introduce into a game-world through physically interacting with that space? How might that affect future experiences in that world and with its digital residents?
Part of a site survey might include a sample excavation, which can reveal strata and other data. While some games might exhibit stratigraphy during play, this is exceedingly rare if not unique. Strata instead can be visualized between different versions of the same game, changes mapped in a software matrix, a variant of the Harris matrix, which will be demonstrated later in the thesis. If possible, the site survey should conduct some kind of remote-sensing (which some games allow as part of their mechanic). It may become apparent that digging is not necessary based on the survey/sensing results and what they say about the overarching research questions decided upon at the beginning of the project. What is appropriate in the natural world might need to be adapted for use in the synthetic world. These methods will grow and change over time, but it is imperative for all video game archaeologists to document what they did and why and what the results were. This will enable the discipline to advance and grow.
“Excavating” the Synthetic
If excavation is indeed necessary, Carver cautions that “excavation is not an unvarying ritual, but a creative study, carefully redesigned every time it’s done. The methods used depend on what you want to know, the site and the social context in which you work. These are always different, everywhere” (Carver 2009:123). So once spade is put to soil, what happens?
We can turn to Roskam’s Excavation (2001) for details on how to conduct a successful excavation. I have flagged several details with an asterisk (*) that can be applied to excavation in a synthetic world. Pre-excavation strategies include aerial photography*, field-walking*, shovel-testing*, reading documentary material*, studying previous excavations of the same site*, performing ground-based remote sensing*, chemical mapping, coring and auguring, evaluation trenches*.
Background preparation includes: defining finance/administration*, identifying staff/support facilities*, and planning for a safe excavation*.
Site preparation incluldes: site clearance*, site grid*, spoil removal*, shoring, de-watering, finds retrieval*.
Recording includes: defining the stratigraphic unit, creating and using numbering systems*, creating a recording process and related sheets*.
The photographic record includes: determining the reasons to photograph something*, photographic preparation*, and technique*.
The spatial record includes: techniques*, equipment*, and drawing conventions*, types of plan*, techniques of measurement*, types of section*, piece-plotting finds*.
The stratigraphic record includes: types of stratigraphic relationships*, representing these relationships*, calculating stratigraphic relationships*.
Deposit descriptions include: who records and when*, computer storage of records*, deposit descriptions in relation to sedimentology and pedology, deposit color, soil particle size, compaction or consistency of deposits, inclusions within deposits, thickness and surface characteristics.
Non-deposit descriptions include: masonry and brick features, timbers, inhumations, cuts, finds groups.
Excavating the stratigraphic unit includes: sampling strategies for finds*, methods of collection*, troweling methods, making stratigraphic distinctions, completing the record*, checking the record*.
Stratigraphic analysis includes: tidying the record*, on-site interpretations*, correlating between units*, stratigraphic nodes and critical paths.
Prior to excavation, we need to determine what to do with the things we find. Carver stresses that a site’s Recovery Levels are always variable, changing from site to site (Carver 2009:124, fig. 6.10). The table is organized from general to specific, from large to small, defined in advance by how they might be discovered and how they will be collected and described. Surface finds (not collected) are followed by large finds (examples recorded and kept), visible finds (all recorded, examples kept), then sieving of samples (kept), total samples (all visible finds), and micro-sieving (done in the lab). For synthetic worlds, sieving is likely not an option (nor necessary), but in many instances examples can be documented, the smaller artifacts kept in player (or team) inventory for later analysis, everything from architecture to small finds, whatever those might be.
Regardless of the environment in which excavation occurs, stratigraphic excavation records multi-concept: Component (grain of sand) amongst which are Finds (anything kept) and these belong to the following: a Context (any defined set of components—a layer, a surface—recorded with a Context card, plan, section; a Feature—any defined set of Contexts recorded with a Feature card plan section photographs; a Structure defined as a set of features recorded with a Structure card plan photographs; a Horizon is a defined interface between truncated contexts recorded with a survey, plan, photographs (Carver 2009:139). These terms can be defined on a game-by-game basis. The terms are scalable.
A major component of both survey and excavation is recording the work done, and what the work recovers. Each project in a synthetic world needs to determine in advance what will be recorded in writing, what will be drawn, and what will be photographed. For games and other synthetic worlds, it may also be possible/advisable to record video. All should be done before, during, and after survey and excavation to create a complete visual record. The record must include notable elements on sites.
For Carver (Carver 2009:198, fig. 8.2), notable elements on sites are both reassessed and managed. Assemblages (artifacts and biota), chronology (stratification and dating), and spatial (contexts to features, features to structures, and structures to site) are reassessed. Management includes records (digital, context/feature cards, maps/plans/sections, photographs, and program of analysis), objects (conservation, packing/storing, program of analysis), and samples (storage and program of analysis). These elements exist equally (reassessment and management) for the archaeology of synthetic worlds, but differ in their execution. For management of archaeological content from a synthetic world, any digital assets/artifacts recovered will likely include digital records (i.e., database entries), short- and long-term storage of other digital media (photos and captured video). It is unclear at this writing what digital conservation might entail for items recovered in a digital space.
Documentation, Chronology, and Location
Depending on the nature of the archaeological investigation to be completed within a synthetic world, archaeologists can follow the example Carver sets out in Table 8.1 (Carver 2009:199). Different kinds of investigations require different levels of documentation, some requiring a fine grain. For example, if one is performing a reconnaissance of a possible site, the archaeologist will keep a notebook of observations, draw a map, create digital maps from GIS points, photograph the area generally, and will note surface finds. A landscape survey differs slightly, including artifact photos. Excavation, however, requires a written notebook as well as context sheets and notes on features, drawn plans/maps/ sections/3D plots, digital context records, coordinates, and maps, photographs that are overviews and portraits, with assemblages noted. As above, this same methodology, this same way of organizing what to do depending on what you are doing, is scalable to digital built environments. The location of these sites is different, but the method is the same.
The purposes of documentation fall into two categories: 1) preservation of the archaeological record, and 2) analyses of recovered data. Documentation can help establish a site’s chronology (absolute dating as well as a site’s sequence of strata and events). For games/software, stratification is easier to read because it is identified by version and build numbers. Version 1.0 appears earlier in the record than version 2.0. The documentation allows the archaeologist to produce synthetic text on assemblage, space, and chronology—what happened, where, when. These all interrelate. So what gets collected from a site and why?
As stated above, site chronology can be divided into absolute and relative dating. Absolute dates are arrived at through scientific testing. Relative dating comes from the ordering of artifacts and ordering of their context. Dating is extremely difficult to accomplish within synthetic worlds when considering artifacts found within, especially when one knows the year in which the world was created (in real-time). Something might appear to be ancient, yet is only (really) a few years old. Relative chronology is perhaps more important to the video game archaeologist rather than anything absolute. Or perhaps one can create a chronology based on absolute dates of discovery. In a procedurally generated game, it might be interesting to compare what is found in one site one day against another site found a few days later. Age would seem to be immaterial.
To get to any kind of dating, one needs sample artifacts/features/materials. Carver identifies five purposes for collecting samples from a site (Carver 2009:223, fig. 9.7), and where these samples ultimately go (e.g., geologist, botanist, lab other specialist): identification, dating, plant use, ambient conditions, chemical mapping. Depending on the nature of the synthetic world and the research question(s) being asked of it, these purposes might change. Identification remains a constant, however, asking the universal questions of “what is this” and “to what purpose what this used.”
In both the natural and synthetic worlds, a variety of materials from sites are present. For Carver (Carver 2009:224, fig. 9.8), each material (e.g., stone, pottery, metal, etc.) has an identifiable fabric (the kind of stone, pottery, etc.), a specific form/type/style, and a history of use (function/symbolism/discard). In a digital built environment, it is up to the archaeologist to determine the general kinds of material available for collection, and to assign types, forms, and styles to them either based on personal observation and experience within the space, or in adopting the vernacular of the player community for that specific game. Contemporary games have shown seriation within types, and also diversify in style between cultures within a game. Identifying in-world artifacts through context and then through type/style can assist in dating them, or creating a biography for a particular digital artifact. It is likely that with examples of machine-created culture that new materials, fabrics, forms, styles, and history will manifest, alien to the human experience. The archaeologist can then create a typology, updating as future discoveries yield new data.
Artifacts typically comprise assemblages, a collection of objects—either objects of the same kind spread across the site or, more typically, objects of different kinds of materials found in the same place. Assemblages are common within games. In fact, an individual game might itself be considered an assemblage of code, the sum of its parts. But most games collect disparate items together in a single space to be discovered by a player, whether it is the remains of a fallen warrior, or an abandoned building. The assemblage gives the player-archaeologist context and understanding, just as it does in the natural world. When artifacts/assemblages are found, they must be documented, and then (if possible/permitted) removed for conservation and additional analysis.
The care of finds from a digital space is perhaps the single major difference between conducting fieldwork in a synthetic world versus the natural one. For a traditional site with traditional artifacts, one must be mindful of foreign substances attached to an item, how to clean the object (if at all), potential risks when handling, packing, or conserving the object, how to apply “first aid”, how to pack, and where to deliver (Carver 2009:226, fig. 9.9). For digital artifacts, the archaeologist is reliant upon photography and videography, plus inventory management both within the game and also within an external database. Any conservation or transportation occurs with cleaning and sharing the digital media that documented the artifacts found within the synthetic world. One additional option, which is finally available at a very low barrier to entry is to create 3D scans of items found in-game, exporting the data to an open source platform such as SketchFab, allowing these scans to proliferate online following the Jeffersonian principle of “lots of copies keep stuff safe.” These 3D scans can also be sent to 3D printers for real-world visualization at various sizes, allowing the archaeologist to manipulate real-world manifestations of something only previously available (and created) in a digital environment.
Prior to their removal, all artifacts in the real world can be tagged with a specific location with X-Y-Z coordinates tied into GIS for the purposes of plotting and mapping. We know the Earth, and have established a universal way to give everything a number, which designates a precise location to anything. At first glance, this is not the case with synthetic worlds. Modern role-playing or adventure games might contain a relative GIS or map, which allows for occasionally specific markers for locations, objects, and players. Most games do not. One workaround to establish at least relative positions for an individual site is to take a screenshot of the site. The siting of artifacts (including structures, environmental objects, etc.) can then be mapped on a grid based on the size and shape of the display on which the screenshot was taken. The screen has an aspect ratio, and also has X and Y axes. Most hardware will also tell the user the absolute pixel dimensions of the screen. Knowing these hardware basics, one can then assign basic Cartesian points to things of interest in the screen capture.
The drawback to this is that there is no Z axis for height/depth. For most games, this will not be an issue, mostly because 3D is an optical illusion for games played on a flat screen. However, with the advances in virtual reality, a more complex system of documentation must be invented in order to place the locations of objects in a truly 3D space. One possible workaround is to note the orientation of the player’s eyes through the use of a compass rose, and the position of the head-tilt, then take a 2D screenshot, and annotate it with regard to how the player was oriented in the real world. I think this method is both complicated and inexact, but one hopes there is a way to extract that data from the middleware that connects the VR headset to the player and also to the game.
In continuing his consideration of the space occupied by sites, Carver distills these neatly into types of field records, which are then defined by tasks, outcomes, and significance (Carver 2009:252, fig. 10.7). For example, records of the locations of monuments can be mapped by monument type to create a plot of dated monuments, which can then show the location of monuments through time. An artifact survey creates a map by type to illustrate dated occupation areas and a sequence of occupation. The data visualization of different kinds of records for different aspects of a site provides the “why” for the survey or excavation, potentially answering the original research questions.
Carver does the same for landscape survey data (Carver 2009:255, fig. 10.10), noting the possibility of creating a linguistic map based on placenames, or the possible location of other sites based on the mapping of artifacts gathered during surface-collection. Both the visualizations of data from site and landscape survey can apply to synthetic worlds, specifically those that use a natural-looking landscape in which players can operate. One might also be able to visualize data via maps in games that have nothing to do with adventurous navigation through perilous lands. One can still draw meaningful conclusions about landscapes and the sites they contain even when recording a digital card game or the elements of a word processor’s graphical user interface (GUI). It all boils down to the fact that we are looking for social relations in the space; function; chronological development (all re: building distribution in a settlement) (Carver 2009:255).
In order to make the most of the artifacts found either on survey or in excavation, their pinpointed locations can help define patterns that can then be interpreted by the archaeologist, pattern-seeking through computation. Analytical routines developed by geographers have helped archaeologists to squeeze more meaning out of their patterns (Carver 2009: 258). Pattern-recognition should actually be easier when applied to synthetic worlds. The natural world is messy, but fractally ordered. Most games are directly designed by one or more people who apply a logic into their coding of a game-space. The algorithms within the code organize the placement of elements in a digital built environment. It seems to be a necessity to apply GIS software to any virtual world in order to view various maps of data, but it will likely be impossible to standardize such a resource because of the plethora and diversity of synthetic worlds. Perhaps the geographers’ “analytical routines” can be applied to the data of any synthetic world, if their formulae are universal and not written for a specific location or even planet.
Patterns retrieved through running locations through georgrapher’s mathematical routines could then be used to create additional visualizations of that data. Carver suggests, for example, the use of Thiessen polygons (Carver 2009:260), which are polygons whose boundaries define the area that is closest to each point relative to all other points. They are mathematically defined by the perpendicular bisectors of the lines between all points. Carver also suggests that archaeologists can utilize Local Density Analysis and Presence Absence Analysis in order to identify patterns of settlement or of artifact creation and use (Carver 2009:261).
For Presence Absence Analysis, archaeologists can take a page from ecologists who are mapping presence or absence of species within a particular environment, creating models for predicting the distribution of organisms from environmental data. Perhaps there are similar models for archaeology that can use the mathematics of ecological distribution, which can be applied to both the natural and synthetic worlds.
For local density analysis, this looks at things such as concentrations of bones or flakes to identify sites of use or production. We could possibly use this within synthetic worlds, too, based on what we find. The mathematics/statistics should be scalable, and I will test this in my video game case studies.
One other source of pattern recognition can take a look at how human and non-human agents in synthetic worlds move. Carver notes that exploring the way that centres of population interacted with each other assumes that there will be preferred pathways rather than random connections, and that these pathways will themselves promote the establishment of new settlements on the route (Carver 2009:262). This is true in the real world, but is it true in synthetic ones? In a game such as Fallout 4, settlements are found on roads, but also in the wilderness. In many role-playing games (RPGs), wandering monsters are more frequent off the road, but major settlements are on the routes. It depends on the game and its design, but by and large things of interest happen on pre-established routes. These routes can be created by the game’s developers or, in the case of No Man’s Sky, through procedural generation (ProcGen) where the software determines a) if there will be creatures present, and b) if there are, what paths they should follow based on the parameters of movement assigned to the creature based on the kind of animal that it is, and on the landscape in which that creature is situated.
During analysis of site data there are always two analytical programs working in parallel: one working on objects, the other on contexts (Carver 2009:272, fig. 11.4). In the natural world, one might find a timbered house built 100 years ago, yet made from wood hundreds of years old. There are two timelines at work: materials and the objects made from those materials. In games, sometimes this will be the same date. In ProcGen games, the material and the object made from the material are exactly the same, created mere seconds ago (even if they have the appearance of being old). In other cases one might record when a material was mined, and when it was later used to craft an object. One could also consider that a material or artifact was made on the date the game was released, or when a game was played by a player. One can argue if these dates are even important. In most cases, probably not. But in some case, dates might be necessary if only to form a relative chronology from which to order items and to sequence elements within a game.
Sequencing of a site includes stratigraphy. This is often done through the creation of a Harris matrix. The matrix includes all the contexts, so constitutes a total account of every stratigraphic event that occurred, or rather, those that could be observed and recorded (Carver 2009:278). I have been able to create a Harris-style software matrix to record the stratigraphic sequence of the game No Man’s Sky, which appears later in this thesis. While stratigraphy is practically impossible to determine within a synthetic world, it is quite easy (if laborious) to identify and document when looking at the software from the outside in.
Seriation, however, can (and should) be conducted within digital built environments. Sir Flinders Petrie was able to use seriation to determine that graves with the most similar pots were nearest to each other in date (Carver 2009:280). Seriation may or may not be as Petrie described when dealing with synthetic worlds. It is dependent on the game’s logic. For games that contain objects to find, one might find artifacts in many locations, which might make sense in the game, but would make no sense at all in the natural world.
Continuing considerations on how to order sites, Carver shows in fig. 11.11 that different types of sites (e.g., deep urban, rural settlement) have varying levels of stratification and ways to order context (Carver 2009:281). Most games will have none of this within the space of play. There is no need for it. Almost everything is superficial in the game-space.
Synthesis and Publication
Once the survey and excavation is complete, and the features have been analyzed, it is time to synthesize the material in order to draw conclusions from the data and to create a site model. For this synthesis, Carver starts with the outcome of primary analysis, followed by a literature search, looking for ethno-parallels, then experimentation in order to create models (Carver 2009:311, fig. 12.12). For the identity of features, for example, he recommends searching the literature about features identified on other sites, looking for features used by comparable cultures, constructing/using/disuse of replica features in experiments, in order to model the activities of the site being synthesized. Synthetic worlds perhaps provide an easier way of synthesizing data as conditions can be recreated (much of the time) and reproduced by others who also have access to copies of the same synthetic worlds. It is possible in some games to tweak a variable or condition in order to experiment, and to rerun those experiments to test hypothesis, which falls under the rubric of agent-based modeling (ABM), which will be described later in this thesis.
Following synthesis, it is time to publish the results. Publication is key to any archaeological project, and is especially so for new, possibly provocative research into non-traditional areas, such as the archaeology of synthetic worlds, video games, digital built environments. Carver lists eight kinds of publications, how they are presented, and for whom: field reports, lab reports, client reports, research reports, popular books, media (magazines, site guides, television), displays, and presentations (Carver 2009:316, fig. 13.1). The latter half of these publications is intended for the public, those who visit sites/museums, and have an interest in popular historical non-fiction. The other reports fulfill the responsibility of writing up the data for other researchers and for the project’s sponsors. The bottom line is that the ethical project must produce similar work for distribution to colleague and to the public as a conclusion to the original, public research plan.
The publication plan for an archaeological project set in a synthetic world would include a research report describing the investigation, findings, interpretation, and context, authored by all of the principal participants in that project. The report should be published as Open Access for maximum discoverability and use, and should be published digitally in order for the report to link to published data sets, and to related digital media, including photos and video, and, if possible, a digital copy (or links to accessing) the digital site that was investigated. This will allow other archaeologists and researchers to return to the site for additional research, or to test hypotheses and results as provided by the report’s authors. A preliminary report should precede the final report of research, to outline methods, findings, and interpretations, published immediately online.
Popular books and other public media should come second, but should not be ignored by the archaeological team. In the case of synthetic worlds, many of these are popular spaces, and conducting archaeology within them generates enormous public interest. Public reports, interviews, etc., can draw archaeologists closer to the communities who occupy game-worlds or who actively use software, and can lead to additional support, be it financial or, perhaps more importantly, for networking purposes. The public can also raise questions overlooked by the researchers, and can critique the work, allowing the archaeologists to revise theories and reconsider results.
As he stated in the beginning of Archaeological Investigation, Carver restates and the conclusion that “without a pre-released project design, a field archaeology project must be judged to be inept, at worst unethical . . . . A project design is therefore a consultation document that is prepared and circulated widely before serious fieldwork begins” (Carver 2009:335). This sentiment is shared in both the natural and synthetic worlds. Both professional and community input are important prior to the start of the actual project. For synthetic worlds, archaeologists who are not directly involved in the project can critique its archaeological plan of attack, while the community can assist be providing expert/native-level intelligence on what to expect within the digital environment to be studied.
In the natural world, archaeology can appear as an inconvenience when it comes to land use. One must speak with property owners and government representatives, acquiring permission and permits, as well as the public. Blocking off a section of land for an excavation can cost others money and opportunity, and the archaeologist must be sensitive to those people affected by the project. In synthetic spaces, the issues are perhaps less sensitive, but must still be considered, especially if work is being done in a game-space populated by other human players. Consideration of their gameplay and use of the site must be respected. Some games also have non-human actors/agents that should also be considered prior to embarking upon a project. How will the archaeological work interfere with their operation in the shared digital space, and are such interruptions ethical?
Carver takes a very common-sense approach to creating a project plan, taking time to consider what needs to be done and why as based on the research questions being asked (Carver 2009:337, fig. 14.1). We need to know how big we want our survey to be prior to conducting site reconnaissance; we need to know the location, problem, and project scoping before evaluating the site; we need to evaluate what these early stages have told us prior to designing the project; we need to continually update the project design during fieldwork to adapt to the site and its environment; we need to assess the data collected prior to creating a formal analysis; we need to understand the results of that analysis prior to publication. This should be applied to any archaeological project regardless of where it takes place.
Carver offers a few “starter” questions for archaeological research projects, and these can be applied to those synthetic worlds that contain ready-made (or algorithmically generated) cultures that are alien to all prior human experience (Carver 2009:342). How are communities and territories defined, and how are societies organized? What are the “people” like, and how are they organized? What are the gender roles (if gender is present)? What does this new culture thing, and what is their world view? What do they eat; how do they make and use tools; what contacts do they have? What is their environment like? Why (or how) did things change? These are universal questions that can be put to any culture on Earth, but can also be used on Mars or within a synthetic world. These are questions for which archaeologists seek answers, deriving those answers by following archaeological method as determined by their research plans.
Carver lays out the contents of a project design in Table 14.1, breaking it down into an introduction, evaluation, research options, conservation options, and recommended integrated program (Carver 2009:353). Completing each of these sections is essential and once done must be shared publicly for comment/critique. The original research plan should also be included as part of the final report as a way to document if things changed during the project, and how and why.
The purpose of this section was to see if archaeological methods and research plans design for real-world survey and excavation could be applied to similar investigations in digital spaces. I hope to have demonstrated that this is overwhelmingly the case. The questions and planning needed to conduct a successful project are the same, even if there are subtle differences in what can be done in synthetic worlds when compared to their natural counterparts. The next step is to create research questions and a project plan for three digital games and synthetic worlds, publish these publicly for comment, and then conduct archaeological research based on the above methods described initially by Carver, and then adapted by me for use in digital space.
—Andrew Reinhard, Archaeogaming