Emergence

Starling Murmuration - an emergent phenomena

When we see flocks of birds or schools of fish, they appear to operate as integrated wholes, yet the whole is somehow produced without any specific bird or fish being 'in-charge'. The processes leading to such phenomena are driven by networks of interactions that, because of feedback mechanisms,  gradually impose constraints or limits upon the members of the system (see Degrees of Freedom).  Recursive feedback between these members (or 'agents') take what was initially 'free' behavior, and gradually constrain or enslaves the behavior into coordinated regimes.

These coordinated, emergent regimes generally feature new behavioral or operational capacities that are not available to the individual element of the system. In addition, emergent systems often exhibit mathematical pattern regularities (in the form of Power Laws ) pertaining to the intensity of the emergent phenomena. These intensities tend to be observed in aspects such as spatial, topological or temporal distributions of the emergent features. For example, there are pattern regularities associated with earthquake magnitudes (across time) city sizes (across space), and website popularity (across links (or 'topologically')).

Quite a lot of research in complexity is interested in the emergence of these mathematical regularities, and sometimes it is difficult to decipher which feature of complexity is more important - what the emergent phenomena do (in and of themselves), versus the structural patterns or regularities that these emergent phenomena manifest.

Relation to Self-Organization:

Closely linked to the idea of emergence is that of self-organization, although there are some instances where emergence and self-organization occur in isolation from one another.

Example:

One interesting case of emergence without self-organization is associated with the so-called 'wisdom of crowds'. A classic example of the phenomena, (described in the book of the same name), involves estimating the weight of a cow at a county fair. Simultaneously, experts as well as non-experts are asked to estimate a cow's weight. Fair attendees are given the chance to guess a weight and put their guess into a box.  None of the attendees are aware of the estimates being made by others. Nonetheless, when all the guesses from the attendees are tallied (and divided by the number of guesses), the weight of the cow that the 'crowd' collectively determined is closer than the weight of the cow estimated by experts. The correct weight of the cow 'emerges' from the collective, but no self-organizing processes are involved - simply independent guesses.

Despite there being examples of emergence without self-organization (as well as self organization without emergence), in the case of Complex Adaptive Systems these two concepts are highly linked, making it is difficult to speak about one without the other. If there is a meaningful distinction, it is that Self-Organization focuses on the character of interactions occurring amongst the Bottom-up Agents of a complex system, whereas Emergence highlights the global phenomena that appear in light of theses interactions.

Enslavement:

At the same time, the concepts are interwoven, since emergent properties of a system tend to constrain the behaviors of the agents forming that system. Hermann Haken frames this through the idea of an Enslaved State, where agents in a system come to be constrained as a result of phenomena they themselves created.

Example:

An interesting illustration of the phenomena of 'enslavement' can be found in ant-trail formation. Ants, that initially explore food sources at random, gradually have their random explorations constrained due to the signals provided by pheromones (which are deployed as ants that randomly discover food). The ants, responding in a bottom-up manner to these signals, gradually self-organize their search and generate a trail. The trail is the emergent phenomena, and self-organization - as a collective dynamic that is distributed across the colony - 'works' to steer individual ant behavior. That said, once a trail emerges, it acts as a kind of 'top-down' device that constrains subsequent ant trajectories.

Emergence poses ontological questions concerning where agency is located - that is, what is acting upon what. The source of agency becomes muddy as phenomena arising from agent behaviors (the local level) gives rise to emergent manifestations (the global level) which subsequently constrains further agent behaviors (and so forth). This is of interest to those interested in the philosophical implications of complexity.

There is a very tight coupling in these dynamics between a system's components and the environment that the components act within. Thus, a specific characteristic of the environment is that it also consists of system elements. Consequently, as elements shift in response to their environmental context, they are, in turn helping produce a new environmental context for themselves. This results in the systems components and the system environment forming a kind of closed loop of interactions. These kinds of loops of behaviors, that lead to forms of self-regulation, were the object of study for early Cybernetics thinkers.

Urban Interpretations:

The concept of Emergence has become increasingly popular in urban discourses. While some urban features come about through top-planning (like, for example, the decision to build a park), other kinds of urban phenomena seem to arise through bottom-up emergent processes (for example a particular park becoming the site of drug deals). It should be noted that not all emergent phenomena are positive! In some cases, we may wish to help steward along emergent characteristics that we deem to be positive for urban health, while in other cases we may wish to try to dismantle the kinds of feedback mechanisms that create spirals of decay or crime.

The concept of emergence can be approached very differently depending on the aims of a particular discourse. For example, Urban Modeling often highlights the emergence of Power Laws in the ratio of different kinds of urban phenomena. A classic example is the presence of power law distributions in city sizes, which looks at how the populations of cities in a country follows a power-law distribution,  but one can also examine power law distributions within rather than between cities, examining such characteristics such as road systems, restaurants, or other civic amenities.

Others, such as those engaged in the field of Evolutionary Economic Geography. (EEG) are intrigued by different kinds of physical patterns of organization.  EEG attunes to how 'clusters' of firms or 'agglomerations' appear in various settings, in the absence of top-down coordination.  They try to unpack the mechanisms whereby firms are able able to self-organize to create these clusters, rather then looking at any particular mathematical regularities or power-law attributes associated with such clusters.

Still other urban discourses, including Relational Geography and Assemblage Geography, are focused on how agents come together to create new structures or agents entities: which might  be buildings, institutions, building plans, etc. These discourses tend to focus on coordination mechanisms and flows that steer how such entities come to emerge.

Accordingly, different discourses attune to very different aspects fo complexity.

Proviso:

While this entry provides a general introduction to emergence (and self-organization), there are other interpretations of these phenomena that disambiguate these concepts with reference to Information theory. These interpretations focus upon the amount of information (in a Shannononian sense) required to describe self-organizing versus emergent dynamics.

While these definitions can be instructive, they remain somewhat controversial. There is no absolute consensus about how complexity can be defined using mathematical measures (for an excellent review on various measures, check the FEED for Ladyman, Lambert and Weisner, 2012). Often, an appeal is made to the idea of 'somewhere between order and randomness'. But this only tells us what complexity is not, rather than what it is. The explanation provided here is intended to outline the terminology in a more intuitive way, that, while not mathematically precise, makes the concepts workable.