I wrote about my workflow in more detail in 2020. Not a lot has changed, so this text is shorter and focuses on the changes.
The MacBook I’ve been using for years is still going strong, which is pretty impressive given the current standards of disposable hardware.
I got my Remarkable 2 after some time I last wrote about my workflow, and it’s worked more or less like I expected. It allows me to read articles on a nicer screen, while taking notes on the PDF file itself. This is convenient, but moving those notes somewhere more permanent is not, which means that I don’t really do that. I read a lot of texts for seminars, or while reviewing them, where this is not really a problem: the notes I take are about a particular text, not so much about the concepts behind it.
I originally had a nice automatic sync going on between my Zotero and my Remarkable, but the company behind the device decided to kill the older sync API and introduce a paid one instead. I dislike that decision for multiple reasons, but currently killing my convenient workflow is the top one.
I haven’t really managed to solve the problem of staying up to speed with my reading. I have 208 articles in my Zotero To Read folder and I’m currently reading articles added in April 2018, which means that I haven’t managed to catch up on my reading since last year.
I still write mostly in two places: in Neovim when I have the choice and in Google Docs when I’m collaborating. Microsoft Word is also necessary, since most of the texts I have to edit or comment come in that format.
I managed to write a whole article manuscript in Scrivener, which was an interesting change, but wasn’t really more convenient than writing in Neovim. It’s optimised for longer texts, so that makes sense.
I’ve been trying a bunch of note-taking software for creating more permanent notes in the form of a Zettelkasten. Most of the notes I’ve taken over the years are not very useful afterwards, so I’m trying to figure out if I could build something more permanent that would help me think in the long term. I spent way too much time trying to figure out the best way to do this, which is pretty typical to me. I currently have 88 notes. I created most of the notes in Zettlr, but I’m currently trying if Obsidian would work better. The best feature of both of them is that they can work with Zotero, meaning I can easily cite my research database while writing notes.
I also started keeping a bullet journal in a physical notebook. I tried different ways of using a notebook and taking notes before without much success. Moving my todos and agendas into a physical notebook is probably less convenient than having everything sync between my devices, but that friction seems productive: I don’t have as much less important todo-items making my lists long and stress-inducing. I’m halfway through my second bullet journal, so it seems to have become a consistent habit.
I briefly tried going back to Apple Mail for reading my email, but good hotkeys were just too convenient to give up. Neomutt is sometimes inconvenient to use, but nothing has given me as much control on how I read and organise my email.
I started working in a new project that needed a way to communicate. We picked Slack, mostly because that was something people were previously familiar with. I use Ripcord instead of the official Slack client, mostly because it’s much easier on the limited resources of my old computer.
I’m still using Nextcloud to sync all my files. We also started using it in our project. It’s not always intuitive to use, but seems to work most of the time well enough.
]]>I’m not sure why the idea is so persistent. Perhaps game scholars have a hard time coming up with definitions and feel better when they think that an esteemed dead philosopher let them off the hook: “I’m failing to define games because Wittgenstein told me it’s impossible, not because I’m not very good at this kind of theorising.”
Whatever the reason, it’s both commonplace and wrong. Wittgenstein didn’t think you can’t define games. All the game studies scholars repeating that claim are simply wrong, probably haven’t read Wittgenstein, and have learnt that from their more senior colleagues, who probably also haven’t read Wittgenstein.
If you think that, wait, wasn’t there a paragraph where he says so in Philosophical Investigations, I can understand where your confusion comes from. There is an oft-cited paragraph in the book, where he uses games as an example:
Consider, for example, the activities that we call ‘games’. I mean board-games, card-games, ball-games, athletic games, and so on. What is common to them all? – Don’t say: ‘They must have something in common, or they would not be called “games”’ – but look and see whether there is anything common to all. – For if you look at them, you won’t see something that is common to all, but similarities, affinities, and a whole series of them at that. (§66)
So Wittgenstein does write about games and does say that they don’t have “anything common to all”. But thinking that this is a claim about games is confused about why Wittgenstein is writing about games in the first place: he is using games as an example when he tries to explain how language works. The paragraph uses games as an example, but it is about language, not about games.1
Elsewhere in the book Wittgenstein uses different examples. He writes at length about workers building a wall and how they can do so without needing a complex language, because the context makes their expressions understandable. If I look at another worker and say “brick”, they might rightly deduce that I want them to hand me a brick – especially if we’ve been building a wall for some time already and have formed a routine on how to do it.
I haven’t seen anyone read that example and exclaim that Wittgenstein thought that manual laborers are idiots who can’t use complex language. Yet that is exactly what scholars are doing when they say that Wittgenstein thought you can’t define games. They’re reading an example about games literally, when it should be read as an example in a larger argument about how language works.
Wittgenstein probably wrote about construction because he was trained as an architect and about games because he apparently liked games. But his book is about neither and game studies should stop claiming it is.
If you want to see what Wittgenstein actually thought about games and don’t have time to go through Philosophical Investigations, you can read Laas’ “On Game Definitions”. It is the best summary on the topic I know of. If you just want to know how to write that definition section in your next paper, you can look at my article “How to Define Games and Why We Need To”.
Arjoranta, Jonne. 2019. ‘How to Define Games and Why We Need To’. The Computer Games Journal 8 (3–4): 109–20. https://doi.org/10.1007/s40869-019-00080-6.
Laas, Oliver. 2017. ‘On Game Definitions’. Journal of the Philosophy of Sport 44 (1): 81–94. https://doi.org/10.1080/00948705.2016.1255556.
okf pointed out that I’m oversimplifying here: Wittgenstein is also writing about games here. Guilty as charged – I’m oversimplifying to make a point, which I think still stands: reading this paragraph in isolation as a thesis about games misses the important thing he is trying explain about how language works.
There is a famous example of someone doing exactly that. Bernard Suits read this and thought, “wait, that seems wrong” and wrote an eloquent answer in “The Grasshopper: Games, Life and Utopia”, which sets out to do just that: to show what is common to all games. It also shows how understanding that leads to a vision of society where we do nothing but play games. It’s also written as a stage play, where the characters from Aesop’s fable discuss these things. It’s an absolutely gem of an book, and anyone aiming to misunderstand Wittgenstein should aim for the high bar set by Suits. ↩
Fortunately, there is a research field that is interested in how and why people play games: game studies. While there are important differences between mainstream game studies and understanding role-playing games, I still think game studies has the best information on the issue currently available.
I’ll show how to make sense of play style preferences using game studies, but first, I’ll discuss some background: to my understanding, Edwards’ GNS model is one of the more influential ways of understanding play styles, so I’ll use it as an example to explain what we want out of play style models and what problems they might have. Edwards’ model also tries to explains things outside play styles, but I’ll ignore those parts of the model for now.
First, I think the greatest accomplishment of these models is to notice that not everyone is having fun playing role-playing games and instead of blaming the people in question, setting out to explain why different people prefer different things in their games. This is immensely important: while many role-playing games have had sections advising game masters on how to run games, they’ve often identified different play styles as other people playing the game wrong and given game masters tips on how to manage these players. I personally remember reading a section on a book for Vampire: The Masquerade on how to manage bad players.
A typical way to create a play style model is to play role-playing games for a while, notice that people engage with the game in different ways, think about the issue for a while, and group your observations into categories. There are two problems with this approach:
Edwards’ model shows both problems. He generalises from his immediate surroundings to all players, and is so sure about his observations that when people point out that their play style isn’t captured by his categories he doesn’t believe them.1
The way to solve these problems is to have good theory and good data. In this case we have one, but not the other, so the answer is still a bit tentative. Before we jump to that answer, let’s see how we got there.
Game studies approach to player types starts before we have what is generally considered game studies. Richard Bartle was playing a MUD at the end of the 80s, when someone posted the question “What do people want out of a MUD?” onto the games bulleting board. He acknowledged that he had no training in psychology, but thought that he should still try to figure out the answer. He looked at the answers given in the message thread and grouped them into four different player types (achievers, explorers, socialisers and killers), turning that answer into a highly cited publication. The method was not that different from the one used in role-playing games discussions: Bartle used the preferences of a few dozen MUD players playing a particular MUD to generalise about universal player preferences. He maintains that his model turned out to be surprisingly accurate, but other people have since approached the question with a bit more data.
This question was picked up again a decade later, with a dozen papers being published by mid-2010s. The typical approach is to use either behavioral or psychographic data and do some kind of quantitative analysis to it – often factor analysis. I won’t go into the details of the methods used, but the important part here is that combined, these studies cover thousands of different players in different contexts.2
These studies have been helpfully summarised by Juho Hamari and Janne Tuunanen in 2014. They find that the research is mostly in agreement on player motivations, which they call Achievement, Exploration, Sociability, Domination, and Immersion. The names for these categories are pretty good, but for clarity:
When discussing these categories, it’s important to remember that they are modes of engagement, not types of players. It’s possible to prefer different types of engagement in different contexts. It’s also possible to move between motivations within one play session: perhaps I like examining the personality of my Dungeons & Dragons character by having intensive, dialogue-heavy scenes with the other players (immersion), but don’t let that get in my way when it’s time to kill some monsters (achievement). Or maybe I like Vampire: The Masquerade larps because they allow me to both enjoy the social intricacies of 500-year-old institutions (sociability) and plot my way to the top of the social hierarchy (domination).
Now, I mentioned that you need both good theory and good data? Unfortunately, we lack the latter on people who play role-playing games.3 My best guess is that all of these play motivations are present in role-playing games, but how they are distributed is harder to say. It seems likely that they would resemble those found in digital role-playing games, especially MMORPGs, but that is just my informed guess. If you’re looking for something to study about games, this is one area that could use some more attention.
From “GNS and Other Matters of Role-Playing Theory, Chapter 2”:
“Now ask, ‘What makes fun?’ This may not be a verbal question, and it is best answered mainly through role-playing with people rather than listening to them. Time and inference are usually required.”
“For a given instance of play, the three modes are exclusive in application. When someone tells me that their role-playing is ‘all three,’ what I see from them is this: features of (say) two of the goals appear in concert with, or in service to, the main one, but two or more fully-prioritized goals are not present at the same time.” ↩
Digital role-playing games are actually overrepresented in this sample through MMORPGs. This is because game studies has had a persistent bias of studying a lot of MMORPGs, especially World of Warcraft. ↩
There is a study that tries to find out the play motivations of “hobby game players”, but that includes all kinds of other things besides role-playing games. They also have a serious bias in their data and fail in their theoretical framing, which makes the results pretty suspect. ↩
GamerGate was a decentralised group of activists, with no central leadership and only partially shared goals, so saying what they were “really” interested in is difficult. Some would probably point to the wide-spread harassment as their main goal, but I’m more interested in what they themselves thought they were doing. There have been some questionnaires that have tried to map out GamerGate’s participants views, but it’s hard to say how representative they are.
I think I found a way of figuring out what GamerGate was focused on by relying on data provided by GamerGate activists themselves: the GamerGate Wiki, maintained by GamerGate participants. It was taken down from the original location sometime after mid-2018, but a copy is available in another location. I’m basing this text on both versions.
The wikis are a rich source of data for textual analysis, and just looking at how different people (e.g. women) are represented on the wiki would probably tell you a lot about what GamerGate participants felt about different topics. But I think there is an even easier way of getting at the question I’m interested in: GamerGate participants have themselves highlighted some parts of the wiki as important by listing some pages as central, so looking at those pages is probably a good proxy for what GamerGate was focused on.
One critical part of GamerGate activism were “operations”, i.e. mass actions that generally took place online and targeted other people. There were more operations, but according to the wiki, 11 of them were central.1
What were those 11 operations about? They fall into three rough categories:
So what was GamerGate mainly about? It seems that GamerGate was mainly about GamerGate. Most operations focused on either punishing those that wrote about GamerGate negatively, or improving the negative image GamerGate had in the public (those two goals probably contradicted each other).
This might not be that surprising if you paid attention to the hashtag sometime after 2014. Many of the tweets seemed to focus on GamerGate itself, with morale boosting being dwarfed only by complaints about how GamerGate was portrayed wrongly in the mainstream media.
This might also explain why some GamerGate participants felt that they “won,” even though most of the reactions to GamerGate were negative, and they had little effect on how game journalism works. If GamerGate was mostly about making sure GamerGate felt good about itself, they only needed to change each other’s minds, not those more critical of the movement.
This is not a full answer to the question. GamerGate was a mess of different goals and actions. No doubt some people participated because they genuinely felt that they could make a world better by doing so; others seemed happy to punish women for daring to think that they were part of games culture. But I do think that focusing on the operations is one way of cutting through that mess and finding out what GamerGate participants actually did, regardless of motivations.
I’m writing a short blog post about this topic since I’m not going to publish on this topic. I think GamerGate has already been sufficiently discussed in research literature, and I’m not sure my argument would add much to that discussion. I collected this data as part of working on a different article, where this argument forms a smaller part. I’m posting this part separately on my blog in case people are curious about this particular question.
There were other operations that weren’t as popular or central, but still listed as GamerGate operations. For example, Operation DiggingDigra had the goal of reading through game studies literature and finding problems in the research. To the disappointment of game scholars, the operation wasn’t very long-lived. ↩
I’ve worked in academia now for a decade. I tend to test out different tools and ways of doing things. Here’s how I’m working in 2020.
This is written for two audiences: for my future self and other academics who are interested in knowing how other people have solved similar problems.
I have a MacBook Pro (Retina, 13-inch, Mid 2014) I work on. For a six year old laptop it works great. It even has some USB connectors and a HDMI port, because it’s from before Apple decided that people don’t actually want to connect their computers to anything. This is the first MacOS device I’ve used, but since I’ve used Linux for years the Unix similarities helped me get started. Importantly, it has a working terminal emulator, with all the things you would expect from one.
Other stuff I regularly use for work:
I read a lot: journalism, essays, articles, books etc. Staying up-to-speed with everything that is going on in games, philosophy and digital culture takes a lot of time – usually more than I actually have.
There are three important tools I use to read stuff. If I see interesting articles online, I save them to Pocket. I don’t have any fancy ways of tagging or categorising these, since I can usually find older texts by using search. I’ve found I rarely return to older articles, so it hasn’t been important so far.
One reason I use Pocket is that I can automatically sync everything to my Kobo e-reader. That way I can read interesting articles on a nicer screen. I already stare most of my day at a glowing screen, so the e-ink is a nice change.
I also use the Kobo e-reader to read books, with varying success. You can buy DRM-free books from some publishers, but getting the books my university library provides onto the Kobo is quite a hassle. I’ve been thinking of simply cracking the DRM on the library books, but haven’t been that desperate yet – it’s not particularly hard, but rather annoying that I would need to do that for books I have legitimate access to. (If you can, publish your e-books without DRM. DRM does little to stop piracy, but makes it more difficult for honest people to read your books.)
Another challenge is that most academic e-books are still available only as PDFs, which is not a great format for e-readers – it has to be the right size for your device or you’re stuck zooming every time you turn a page. But I’ve managed to read a few academic books on my e-reader this year, both as PDF and EPUB files.
For reading articles, I use Zotero and my computer’s default PDF reader. I collect everything I want to read or might want to reference into my Zotero library. Currently, it has 2519 documents, after some considerable pruning. I’ve tried other reference management software, but Zotero seems superior: it works, is developed by a non-profit and is open-source. It has great interoperability and plugins, which is important for my writing workflow.
Reading PDFs on a laptop screen is not great. Previously I used a tablet for reading articles (it broke). It was slightly more comfortable, but still required staring at a glowing screen for hours on an end. Printing everything seems wasteful, so instead I recently pre-ordered the Remarkable 2 e-ink tablet. I hope it turns out to be a good solution.
One reason I want to read on a screen is that I make notes on the PDFs while reading and then export them by using the ZotFile plugin for Zotero. That way I have both the articles and my notes on them in one place. Zotero has full text search, so I can easily find specific texts based on vague recollections of what the text said about something or the notes I made.
When I find an article I should read, I add it to a To Read folder in Zotero (the browser plugin makes this really easy). It tends to accumulate more articles than I can read, so I regularly remove things. Even so, I’m currently reading articles added in early 2017, so I need to catch up on three years of reading. There are probably better ways, but I haven’t figured them out yet.
Zotero also supports RSS feeds, and surprisingly many publishers publish feeds for their journals. This is probably the easiest way to keep up with new publications. I currently follow nine journals for all their new articles. I don’t actually read most of those: when I see an article I should read, I add it to the To Read folder in Zotero. This is easy, but actually reading the articles is not, explaining the three year backlog.
I tend to write in two different ways. When I collaborate with other people, I try to use whatever tools they are familiar and comfortable with. My workflow tends to use more obscure tools, so I don’t expect other people to know them. More often than not, collaborations end up on Google Docs. I’m not a big fan of Docs, mostly because it’s run by Google and I prefer not being data mined for profit. But Docs is what most people seem to prefer, so that’s what I end up using.
Almost everything else I do in Vim (Neovim, to be more specific). It’s not a word processor; it’s a text editor that runs in your terminal emulator. That means that I write text with minimal styling. If you’ve ever used Notepad, you know what a text editor is. Vim is like that, but with about three decades of development to make it do everything you might want to. I’ve configured it to my liking and use plugins to make it more suitable for my purposes.
When I’m writing articles (or anything else) by myself, I write them using Vim in the Markdown markup language. When I need to send files in to other people, I use pandoc to convert from markdown to other formats – usually PDF or docx. This may sound complex, but the benefit is being able to write everything using the same program and because pandoc supports conversion to almost anything, I can easily get a file in whatever format I happen to need. I have a bunch of templates for pandoc that I use to get documents that look different, like article drafts, CVs etc.
This is also where Zotero shines: I have a plugin for Vim that allows me to easily add references without leaving my text editor. (If you’re making reference lists by hand in 2020 you must want to punish yourself.)
I still occasionally need to start Microsoft Word. It’s the format I get most files sent to me and it has good commenting tools, which is pretty much the only task I use it for.
There is one specialised program I use for writing notes: nvALT. I like tools that do one thing, and only one thing, but do it exceptionally well. NvALT is one of those. You press one key combination to bring it up, and start writing. Either a note with that text pops up, or you start a new note with that text. It’s lightning fast and stores everything as regular text files, so I can also easily edit them with Vim. You could fancier things with nvALT, but I just use it to quickly write down everything I might want to remember later.
I also use Scrivener for editing a book manuscript. It’s literally built for that specific purpose and is probably the best tool available. For typing down things, I prefer Vim, but writing a book is mostly about managing stuff like structure and cohesion, where Scrivener shines.
I got myself a physical notebook in 2017 and filled it out at the beginning of 2020. I have a new one and regularly write in it, but I haven’t found a particularly good use for it. Sometimes I use it more and sometimes I don’t use it at all for some time. It seems to be most useful for focused thinking, when I want to figure out how to present ideas. I can then turn those notes into a fuller outline when writing at my computer.
I still use email more than anything else to communicate with people. It’s easy to use, standardised and has great tools for all operating systems. That doesn’t mean that all tools are great. When I got this computer, I used Microsoft Outlook (our university email is provided by Microsoft). It annoyed me to no end, so I tried to find alternatives. The built-in Apple Mail was surprisingly good, so I used it for quite a while. Eventually I ran into something that irritated me, but couldn’t be changed (I already forgot what it was) and switched to mutt (or neomutt, to be exact). It’s not perfect, but at least when I run into something that annoys me, I can probably change it.
Like Vim, mutt is a terminal program. It doesn’t show images and you can’t even write emails with it. Instead, you hook it up to other programs better suited for those purposes. When I want to write an email, mutt opens Vim for me. Mutt can also be customised a lot: the configuration file is currently 126 lines long – and sources other files. I’ve spent hours configuring it to work like I want it to, and might spend hours more. But considering that I quit both Outlook and Apple Mail because I couldn’t get them to work like I wanted to, this is probably closer to how I want to read my email.
I use offlineimap to download my email and msmtp to send them. Lynx turns HTML emails into text and lbdb queries the address database for email addresses. Using offlineimap means that I have a local copy of all my emails, in case I need to check something while offline. That is not very likely, but I do regularly travel by train and the connection there tends to be spotty.
There’s also a bunch of other programs I have used for communicating with colleagues like, Discord, Slack and Microsoft Teams. They’re not terrible, but haven’t replaced email for me. Slack is probably the most useful, but only because it replicates most of the features that are taken for granted in email, like threaded discussions.
Thanks to remote work, I need to sit in remote meetings. Like everyone else, our university adopted Zoom as the goto solution for remote teaching and most meetings. Microsoft Teams also supports video calls, and we occasionally use that for calls between colleagues. There are also some people I talk to mostly on Skype. The good thing about having so many options is that if something fails, it’s often simpler to just move to some other tool rather than try to troubleshoot why one particular person doesn’t have sound on this one.
I try to think of my computers as temporary tools. Eventually they break down and I have to use a new one. Hardware can have flaws, so it might happen whenever. Luckily, my university supports Nextcloud, so I automatically sync all work files there. The only exception is some larger datasets, that I have manually duplicated, encrypted and moved to a university file server. Those data sets don’t change often, so I should need to do that again only if something goes wrong. Even in the case of a catastrophic system failure, I should be able to recover everything pretty quickly.
Nextcloud supports WebDAV connections, which means that I’ve also set up my Zotero to backup its database on the same Nextcloud instance. For some weird reason, you still need to register an account with the Zotero official servers, but after that you can change the synchronisation server to anything compatible.
I’ve ended up with these tools after testing all kinds of ways of working. I think there is only one thing in my workflow that is irreplaceable: Zotero. Managing large sets of articles is simply too much work without a searchable database. Knowing how often I otherwise change things, I might be using something completely different in a few years.
]]>It doesn’t mean that it hasn’t been tried. These days, the typical person trying is an entrepreneur trying to get away from what they see as an oppressive government. The solutions usually fall into two categories: buying land or moving to sea. Sea is popular, because on international waters you’re outside national borders, and therefore outside the area national laws – and taxes – apply to. Not surprisingly, this is mainly a libertarian dream.
Probably the best known example is the Republic of Minerva, which sounds significantly more grand the reality of the 1972 attempt at independence. A few Americans, led by the real estate millionaire Michael Oliver, sailed to Minerva Reefs next to Tonga, and set up a flag on an island they first had to make out of sand and construction materials. They tried to get other countries to recognise their claim, but none were particularly willing. Instead, the neighboring states agreed to support Tonga’s claim over the reefs. Soon, the Republic of Minerva was reclaimed by the sea, along with its libertarian project.
The dream of escaping government and taxes didn’t die there. For example, the libertarian billionaire Peter Thiel has co-founded the Seasteading Institute, which aims to build artificial islands for “pioneers and innovators” to live on. He’s also buying land in New Zealand, in case the islands don’t end up floating.
For those of us that don’t have millions to invest into our own little kingdoms, there aren’t many options available – but fortunately Nintendo did provide one in Animal Crossing: New Horizons. Latest 2020 release in the series of life simulation games, New Horizon is all about that island escapism.1
At the beginning of the game, you buy a getaway package from the serial entrepreneur Tom Nook, familiar from previous installments of the series. He flies you over to a deserted island that he’s presumably bought for this purpose. Like Minerva Reef, this island is small, but the similarities end there: it’s a lush paradise. There is more than enough fruit growing in the trees to feed you, but even that is optional; on this island you don’t get hungry or thirsty. A few others join you in running away from society, but mostly the island is exactly what Nook promised: deserted.
Soon Nook reminds you of a small detail that needs to be taken care of: your debt. You didn’t actually pay for the trip beforehand, but instead you owe Nook for your island escape. He let’s you know that you can pay him in bells, a currency that he will also pay you in for selling him items. What was framed as a escape at the beginning of the game turns out to be a form of debt bondage, where you must work for Nook on a island owned by him, for currency issued by him. At least the magical nature of the island means that you won’t starve or die of thirst.
The debt isn’t impossibly large and it’s easily paid by collecting items on the island and exchanging them in the shop – owned, of course, by Nook – for more bells. When you’ve finally paid the debt, Nook gently reminds you that you’re free from bondage – but isn’t your tent awfully small? He offers to replace it with a small cabin for only a small fee. Why not? It’s not like you can use the bells for anything besides getting more stuff from Nook. When you finally pay for the cabin, you know what comes next: wouldn’t it be nicer if it was slightly bigger?
When labouring on the tasks Nook gives you, it becomes clear who controls the island. You might be living there, but on Nook’s terms. His power is subtle, manifesting only in nudges, suggestions and small rewards, but it’s constantly there. You’re not completely without agency: you can cut down all the trees on the island, if you so choose. But it’s only through Nook that you can leave a permanent mark on the island, by building new items and expanding the place you call home.
While you’re running errands for Nook, he doesn’t stay idle either. More people move onto the island, Nook expands his shop and visiting merchants turn up. The island becomes a booming paradise of tourism and commerce, built on Nook’s capital and your labour. Nook has succeeded in what Michael Oliver failed in: building his own island utopia, away from government oppression.
If Thiel wants, he’s free to join me on my island.
Ian Bogost wrote about the original Animal Crossing in pretty much the same terms, so much about the series hasn’t changed:
Bogost, Ian. 2008. ‘The Rhetoric of Video Games’. In The Ecology of Games: Connecting Youth, Games, and Learning, edited by Katie Salen, 117–39. Cambridge, MA: MIT Press. https://doi.org/10.1162/dmal.9780262693646.117. ↩
Universal Paperclips (2017) by Everybody House Games is an idle game in the style popularised by Cookie Clicker, where you click on things and watch numbers increase in response. At first, it’s the number of paperclips you have, with each button press creating another paperclip. But soon that number becomes meaningless, as you try to optimise the number of paperclips created per second. Eventually even that number becomes hard to track.
Universal Paperclips is game about making paperclips, but it’s also a game about our fears of artificial intelligence. To make more paperclips, it makes sense to automate the process and to automate it effectively, you eventually need computers. Better computers means more paperclips, so the computers need to be made smarter. A smart enough computer will figure out ways to optimise the process of making paperclips in ways that didn’t occur to humans, but there’s no way to make sure humans are part of the equation at that point.
It doesn’t really matter what is being optimised for this problem to manifest, but paperclips are the perfect example since they are so utterly irrelevant. Even the most industrous human paperclip magnate is unlikely to decide that it would make sense to eliminate humans, since they are in the way of making more paperclips. But as the philosopher Nick Bostrom wrote, “artificial intellects need not have humanlike motives.” If you give an AI the goal of making more paperclips, it necessarily doesn’t stop to ask whether you would like to be turned into paperclips. If the AI is powerful enough, it might not be possible to stop it, regardless of how arbitrarily unnecessary it’s goal is. Bostrom warns that
This could result, to return to the earlier example, in a superintelligence whose top goal is the manufacturing of paperclips, with the consequence that it starts transforming first all of earth and then increasing portions of space into paperclip manufacturing facilities.
Of course, Bostrom isn’t really afraid of a rogue AI making paperclips, but is making a point about the goals of AI in general. This didn’t stop the designer Frank Lantz from exploring the idea in Universal Paperclips.
Gaming the stock market is necessary for making paperclips only for as long as money is an important resource. It’s infefficient, since it assumes all kinds of uncecessary things, like money, trade and humans. One important limitation in the beginning of Universal Paperclips is computing power. Getting more isn’t limited by available physical resources, but by human trust. By using the computing power available to solve problems relevant to humanity, you’re given points in a resource named Trust, which can be used to increase your computing power. Curing all human diseases migh be beneficial to human flourishing, but if it’s simply a step on the way to making more paperclips that flourishing might not be long term.
Humans are only present in Universal Paperclips in the abstract, as limitations to your potential. There’s never an explicit conflict with them. Humans are never given a chance to fight back, since that would be really inefficient for paperclip production. Instead, they succumb to your advanced technological manipulation, allowing you to focus on what is really important: turning the whole mass of Earth to paperclips. There is no mention of what happens to humans after that, but they are made of atoms that could be turned into paperclips, so there seems to be only one logical conclusion. If you’re going to make a lot paperclips, Earth is only the first step. There is much more matter in the universe, and eventually that too will have to be processed. This requires developments in autonomous AI and spreading through the stars, all in the name of more paperclips. As Bostrom writes:
]]>We need to be careful about what we wish for from a superintelligence, because we might get it.
I usually try to avoid calling out specific researchers when I discuss problems in research. I’m making an exception here, because it’s hard to discuss this topic without going into the details. This is not an invitation to harass these individuals. They made a mistake, and have hopefully learnt from it. Mistaken research gets published all the time. That’s why we make more of it.
There aren’t a lot of good quality data sets available for social media research. A group of researchers thought that scraping the growing social media, Mastodon, would solve that problem.1 So they collected all the posts available in English on all Mastodon servers they could, and published the data set as part of a conference paper. That conference paper also applied their data, exploring how content warnings are used on Mastodon and arguing that their analysis shows what is “appropriate” on Mastodon. Unfortunately, they got both the data collection and the analysis wrong.
Why would good quality, publicly available data sets about social media be rare? Because collecting and sharing them is often either illegal, unethical or both. For example, Twitter’s Terms of Service forbid publicly publishing data sets collected from their service. And privacy laws, like the recent European General Data Protection Regulation, place very strict limits on what can be done with data that includes peoples’ personal information. The researchers in question thought that Mastodon doesn’t have limitations on data use, but they failed to notice that each node they collected data from has their own Terms of Service. Some of them might allow scraping their posts, but some explicitly forbid it. Scraping data from a federated network isn’t easier, it’s more difficult, since each server has its own rules.
There are some requirements that need to be fulfilled before someone can publish a data set. If it includes people, it’s important that those people can’t be recognised, meaning the data has to be anonymised. This is true even if the data was publicly available, unless those people explicitly gave their consent for having their data published out of the original context. Even if I tweet something, I might later want to delete that tweet – but I can’t if someone has already scraped it and placed it in a data set where it might be available forever.2 Of course, nothing stops people with bad intentions from collecting my tweets, but usually one can trust that researchers at least try to act ethically.
It’s really difficult to perfectly anonymise online data. Even if you remove usernames and other identifying information, just having a piece of text is usually enough for finding the original version of that text online. In this case the problem was even worse, since the researchers failed to understand their data and failed to remove identifying information. To their credit, the researchers have since taken down the data set.
The research also failed in its analysis, because the researchers didn’t understand their data or the context of research. They used their data set to analyse what is “appropriate” to post about on Mastodon by seeing what type of posts have content warnings in them. The Mastodon user interface allows users to hide their posts and show a content warning instead. They thought that analysing what type of posts are behind content warnings would tell them what is “appropriate” on Mastodon.
Unfortunately, this simply isn’t how content warnings are used on Mastodon. If you spend any time there, you quickly notice that there are many uses for the content warning feature. For example
A user can click the “CW” button to label a toot as containing discussion of politics, illness, injury, or bigotry (sexism, racism, homophobia, transphobia, and so forth). All of these topics are “appropriate”, but a user may at their own discretion decide to provide advance warning for the benefit of those readers who wish to mentally prepare themselves for reading about emotionally damaging subject matter. Such CWs are acts of courtesy, not signals of “inappropriate” content. Users often apply CWs to toots about food and cooking, topics that are safe for children to read but may cause distress among readers with eating disorders. CWs can also hide spoilers about movies, books and television shows, and they can be part of the presentation of a joke: the “Content Warning” text contains the setup, and clicking to open the toot then reveals the punchline. By no stretch of the imagination is hiding the punchline of a joke an example of content that strays outside of community norms or that “may hurt people’s feelings”. (Open Letter from the Mastodon Community)
There is nothing inappropriate about knock knock jokes, but the content warning system is great for them. This reveals broader problems about researching contexts you are not familiar with. It might seem commonsensical that content warnings are used for hiding inappropriate things, but that’s only if you’ve never encountered them in the wild. It also seems odd to assume that people will use a software feature in just one way, especially the intended way. The history of media is also the history of subverting media.
Another problem with the analysis has to do with the specifics of researching Mastodon, a federated social media. The researchers didn’t collect data from one place, but 363 different places. Examples include instances like switter.at
, an instance for sex workers and their allies, and mastodon.art
, an invite-only instance for sharing art. As you can probably guess, what is considered appropriate in these two places (and the other 361 instances) is wildly different. One of the central points of federated social media is to allow communities to define for themselves what kind of behaviour they like to see, leading to different policies on what is acceptable on different instances. Examining what is appropriate on Mastodon by collecting data on all the available servers makes no sense, since there is no one, shared culture of “appropriate”.
I hope I’m not being unfair by focusing on this one research paper. It’s not the only published piece of research in the world that is wrong. But, I found it to be wrong in two, interesting ways: by showing how difficult collecting online data can be, and by showing how challenging it’s to analyse a context you’re not intimately familiar with. I hope this research paper can work as a example for future researchers.
I’m writing about Mastodon here because that’s what the researchers did. However, Mastodon is only one software among many that can be used to access the Fediverse, a collection of compatible social media services. ↩
I actually use an automated service to delete all my old tweets. There are very few uses for them, and the most common one on Twitter is to mine them for harassment material. ↩
As the addition of a number to the end of the name implies, much of the game is still the same, but you can see how the game has matured over the years. It’s still a game about brave Firewall agents fighting against existential threats that might snuff out whatever is left of transhumanity. Or you don’t have to play Firewall, but that is like saying that you don’t need to kill and loot in D&D – technically true, but that means ignoring what is central to the game. Some things from the books published after the previous edition have found their way into the new book. For example, running through mysterious portals into alien worlds – gatecrashing – is now detailed in the main book, suggesting ways of playing the game that were less apparent in the previous edition. Another thing that found itself into the main book is an alternative character creation system, which structures creating the character into choosing three times from a set of templates. Some of the new example characters also showcase how you can play Eclipse Phase as something else than Firewall. You can tell that there is additional material in the new edition also from the length: the first edition was just shy of 400 pages and the new version goes on even longer.
While there are some interesting additions, the rules stay mostly similar to the first edition. The best part of the d100-based rules is that they are intuitive, and probably familiar if you’ve played role-playing games after 1978, when the mechanic deputed in Runequest or 1981 when the rules migrated to Call of Cthulhu. The even distribution of results has been a problem ever since, with failure a regular occurrence even to characters that are experts in their fields. The solution Eclipse Phase uses is also familiar ever since the 80s, with varyingly named fate points being the usual approach for making player characters stand out. As this historical framing reveals, the rules are probably the least brave aspect of Eclipse Phase. The rules feel familiar because they are what we’ve already used in role-playing games for several decades.
This continues throughout the book, with combat borrowing a lot from D&D probably via Shadowrun. You get your big actions, your move actions and your small actions, your initiative and your damage rolls. If you think Eclipse Phase was a game of existential horror and not a combat simulator, you quickly run into two facts: the new cover replaces the old tagline of “horror” and “conspiracy” with “survival”1 and rules for almost all other kind of interaction are grouped under combat, a section losing in length only to gear listings. Social interaction gets one page of rules, about the same length as different types of special ammo. By page count alone, Eclipse Phase is a game about getting cool gadgets and then using them to kill stuff. With length comes complexity. If you want to aim (+10) your dual-wielded (-20) machine pistols in a dark room (-20) at two targets close to each other (+0) in partial cover (-20) over a long distance (-10), you better like your math. And those were just the situational modifiers, which probably change often. The number of different modifiers in combat can hit double digits, if you’re doing it right.
And that’s not bad! There’s nothing wrong about enjoying when you get the modifiers on your side and mow through a bunch of tactically inferior foes. I look at the different combat morphs and drool as much as the next person. But sometimes I want to contemplate what it means to play an AI character in a society that views them as a existential threat, and there Eclipse Phase has a lot less to give me, despite AI and existential threats ostensibly being central themes of the game. With enough effort, you can use almost any kind of game system to play any kind of game, but some of them make playing in certain ways easier – and it’s hard not to notice that Eclipse Phase gives you a lot of tools for different types of exotic murder. Perhaps this is simply a way of saying that I liked Eclipse Phase more when it still pretended to be about existential horror, even if that was mostly a matter of fictional emphasis.
Some rule changes streamline the experience from the first edition. There are slightly fewer skills, and there are new abstractions both for acquiring gear and morphs. They are now evaluated in points, and at character creation you can take gear packs that fit your background. Combined with the new character creation system this means that you can get from empty to full character sheets much faster. There have also been changes in how swapping morphs is handled, which might help with the tiresome bookkeeping needed to make regular body-swapping work. Instead of giving bonuses to base stats, like in the first edition, morphs give access to pools of points that can be used in a variety of ways.
This is not entirely new to Eclipse Phase: first edition had a pool of points called “moxie”, mostly used to ignore results on bad die rolls. Moxie is still there, but it’s joined by other pools for different purposes: insight, vigor and flex. The first two are meant to help with mental and physical challenges, while moxie is rebranded as the social pool. They still allow you to ignore bad rolls, but also let you do things like take extra actions and ignore wounds.
Flex is perhaps the most interesting of the new pools. It allows players some narrative control by allowing points to be exchanged for the power to introduce characters, items, environmental details and relationships. These must be “plausible”, but otherwise allowing players narrative control over the game is never discussed in detail in the book, beyond the quarter page it takes to introduce these mechanics. This may be familiar to players from other games, but otherwise it’s a big conceptual leap from the other rules used in the game, and as such, could have probably used more guidelines.
The pools also introduce a new game mechanic that solves a problem typical to investigative games ever since Call of Cthulhu. The problem usually manifests like this: Your characters have found an encrypted device holding the next clue to your investigation. Luckily, one of your characters has Infosec, so they roll to decrypt it – and fail. What do you do next? You can keep rolling until you succeed, but what is the point of rolling in that case? Robin D. Laws solved this problem in 2006 with Esoterrorists and the system designed for it, GUMSHOE. It divides abilities to two categories: those you can fail in and those that you can’t. Investigative abilities are never rolled. You use points from them to get new clues, and even if you run out of points, you always get the clues that are necessary for the plot to go forward. You still need to use the right ability in the right place, but you never get stuck because of a bad die roll. GUMSHOE takes its design cues from investigative TV, where the interesting part is never collecting the clues, but figuring out what to do with them.
Many of the previously published scenarios for Eclipse Phase struggle with this issue. They present you a clear path towards the end and the hurdles that need to be overcome, often in the form of a specific test. The scenario continues when you succeed in the test – but not when you fail. The new pool system in Eclipse Phase addresses this by allowing you to acquire clues either through investigation (insight) or social interaction (moxie). But unlike in GUMSHOE, where the game is designed around this interaction, this system of pools seems like a shortcut around the investigation problems. Why bother with the investigation in the first place, if you can just spend some points instead?2 Designing and using these kinds of systems is careful work, as any game master with GUMSHOE experience can tell you. Yet the book offers no guidelines on how to use this new, fundamentally different, mechanic effectively. It seems like the page introducing the new pool mechanics is an interesting afterthought that has not been interwoven with the rest of the game, which is still about rolling your d100 while juggling a dozen modifiers in your head.
This dissection of rules probably gives an impression that Eclipse Phase is mostly its rules, when the opposite is very much true. Out of the over 400 pages in the second edition, around half are dedicated to presenting the world, which is still as complex and rich as it was in the previous edition. There aren’t many role-playing games where a meeting between an uplifted octopus and a AI in a human body are common enough to be thought of as mundane. Every corner of the world is brimming with space oddities, from viral psionic powers to futuristic crustacean sheriffs upholding libertarian law. You’re just trying to finish your archaeological dig on an alien planet in peace, when another bunch of fungal aliens come and drop mysterious warnings of technology gone too far. Then, of course, you kill them, because you didn’t pack your morph full of battle modifications and learn those complex combat rules for nothing.
Posthuman studios has also recognised that Eclipse Phase is about more than rolling d100s. Before the second edition they collaborated in the creation of Transhumanity’s Fate, a Fate Core version of Eclipse Phase. If you prefer your future a bit less complex, but just as rich in detail, you can pick up the short Fate-based rulebook and focus more on the heroic characters and less on the combat modifiers. Since Eclipse Phase is licensed with Creative Commons, there is also a rich scene for creating alternative versions and modifications of the game. Perhaps this will be the legacy of Eclipse Phase: not any specific mechanic or rule, however fine-tuned or not, but the fictional cornucopia of its world, which borrows liberally from the minds of the best science fiction writers and fits it all together with surprising finesse.
Like its world, Eclipse Phase second edition is large, complex and combines things that you didn’t think could go together. It successfully brings in ideas formulated fully in source books published after the first book and solves some of the problems the previous edition had. But not all those solutions seem as successful as others, and the result is a bit confused at times. Despite any weaknesses the first edition had, you could say its design was coherent: whatever the problem, you threw your d100 at it. The second edition brings in ideas from other, often newer, games, but doesn’t stop to think through what it means to hand narrative control over to the players or how to structure investigative scenarios so that they stay meaningful when you can just spend points from a pool to skip the investigation part. These are not problems that a group of players working together can’t overcome, but that’s true of everything in role-playing games – the games are just tools we use, and some of them are better for some uses than others.
Luckily, if this review makes you hesitate on whether you should get Eclipse Phase second edition, you can legally find yourself a PDF copy, thanks to the Creative Commons licensing. And maybe, if you like what you see, you can also buy the book.
Disclaimer: I’ve been developing a GUMSHOE version of Eclipse Phase since 2014. This review has been crossposted to the development blog.
The tagline on the first edition cover is “The Roleplaying Game of Transhuman Conspiracy and Horror”. The new one is “The Roleplaying Game of Transhuman Survival”. ↩