Is Emergence a Metaphysical Property?

by tommy.rousse

Yesterday, for Digital Game Theory, we had an excellent guest lecture by Espen Aarseth, the very fellow that fired the starting pistol that began the academic race1 commonly referred to as Game Studies.  The topic was “emergence,” and with a reading list consisting of Jesper Juul’s 2002 framework of opposition between progress and emergence in games, “The Open and the Closed: Games of Emergence and Games of Progression,” 2 & a fairly technical overview, Jochen Fromm’s “Types and Elements of Emergence.”  I came in partial to the concept: I remember reading interviews in PC Gamer as a teenager with Warren Spector & Harvey Smith describing Deus Ex as having “emergent gameplay” and it sounded like a cool idea that I was pretty sold on at the time.

Color me surprised that Espen, largely by way of a thoughtful Socratic method, convinced me that almost all uses of the term “emergent” in contemporary gaming talk are at the very least incomplete & most are largely inaccurate.  In fact, when Espen was making a list of fields in which emergence was misused, and a column of warrants beside it, I hazarded adding philosophy, and gave a short justification based on Kant’s Prolegomena to Any Future Metaphysics that I thought was fairly clever at the time.  Given the fact that my shamelessly interdisciplinary adventure has left me somewhat  of a dumb-dumb when it comes to philosophy, I can only plead that it was my sleep-addled state3 that induced me to clamber out over thin philosophical ice.  I have since reconsidered.  But there were some damn nifty ideas in that emergence stuff that’s made me a lot more excited about the study of complexity than I have been in a long time.  Let me tell you a story—

The problem with “emergence,” like many terms with ambiguous meaning, is one of difference and boundary-drawing.  The term’s appearance in wider culture is rarely accompanied by a suitable definition, and it is applied to a wide variety of phenomena which a  careful examination proves to be dissimilar.  To critique the usage of emergence in the buzzword-happy games industry, we can begin by asking “What is structurally different about game X, dubbed emergent, and game Y, which is not considered emergent?”

Is it simply that the players have done something that the developers didn’t anticipate?  Given a suitably flexible and complex system, we should dismiss the arrogance implied by the feigned surprise inherent in the well-worn “Gee golly, look at what those players have done!  Inconceivable!”  genre of developer commentary.  We might more charitably expect that the actions of a community of thousands or millions of players would outstrip the set of hypothetical situations imagined by a small team of designers with some regularity.

Many popular uses of the word “emergent” could be more accurately replaced with words such as complex, surprising, or unpredictable.  On closer analysis, it seems clear that it isn’t that games themselves have emergent properties, but that some games provide a rich medium for expression or communication for the most emergent beings in the known universe: humans.

That’s because we humans are the culmination of two incredibly complex processes which Fromm states are the best examples of “strong emergence”—life and culture.  In cases of strong emergence, the cumulative effect of the phenomenon cannot be derived by extrapolating from a robust but microscopic understanding of its constituent parts.  That is, even if our understanding of the origin of DNA and genetic systems were perfect, it would be impossible for us to trace the causal train from nucleotide to frog, for example.  Genes and memes are replicators which can be combined in a staggering number of possibilities.  So many, in fact, that they quickly approach 10120. Fromm calls this mind-bogglingly large number the Landauer-Wheeler-Lloyd limit,4  which is more or less “the number of bits of information that have been processed by all the matter in the universe.”  That means we lack even the conceivable computing power to sort something that big out.

I had never thought about the limits of computational power before.  It reminded me immediately of Kant’s argument in the Prolegomena, at least as much of my original, limited understanding as I retained.  To doubtlessly grossly oversimplify and misstate, Kant’s slim little book was an incredibly intelligent rant saying to all of philosophy, “Look, until you figure out my god damn book and either accept it or figure out a better way to fix the problems, talking about metaphysics is stupid!”  And the problem with metaphysics is that we lack an obvious faculty to directly experience the metaphysical, and no compelling evidence that we are capable of gaining knowledge outside of the realm of physical sensation.

At first glance, the incomprehensible nature of the Landauer-Wheeler-Lloyd limit seemed to fit the bill.  But after giving it some thought, I realized that the two are very different: while the L-W-L limit might be quantitatively impossible to explicate, we can understand small segments of it, while to my understanding we lack the sensory apparatus to experience Kant’s metaphysical—it is qualitatively inaccessible.

To return to the title of this post, then—


  1. A race to establish an ego-centric academic citation path dependency; to have one’s terminology widely adopted as the foundations for a new field.  Until the dust settles, why not throw one’s hat into the ring?
  2. A small nit-pick: Juul lists “working together in a group” in EverQuest as an “emergent strategy” not immediately deductible from the game rules.  However, class interdependence is an explicitly designed feature of the game (Taylor, 2006)  which is fairly obvious to the player & almost required by its design. Applying that definition by analogous standards, we could argue that running and jumping simultaneously in Mario could also be classified as an emergent strategy, though plausible interactions might also be made for rule interaction and combination, the other proposed levels of emergence.  There is an obvious strain on the severability of these categories implied in this analysis.  As an aside, it seems class interdependence mimics specialization in everyday life—Durkheim might consider Norrath an organic society.  Coyly, we might claim EverQuest creates an organic society mechanically.
  3. Joon & I were up late putting the finishing touches on a poster submission about our alpha jam game development methodology. We may have also been putting the inaugural touches on said poster, between you and me.
  4. I wasn’t able to find it anywhere else, actually. Turns out the name for that number is  novemtrigintillion, though.