### abstract ###
schelling  CITATION  observed that macro-level patterns do not necessarily reflect micro-level intentions  desires or goals
in his classic model on neighborhood segregation which initiated a large and influential literature  individuals with no desire to be segregated from those who belong to other social groups nevertheless wind up clustering with their own type
most extensions of schelling's model have replicated this result
there is an important mismatch  however  between theory and observation  which has received relatively little attention
whereas schelling-inspired models typically predict large degrees of segregation starting from virtually any initial condition  the empirical literature documents considerable heterogeneity in measured levels of segregation
this paper introduces a mechanism that can produce significantly higher levels of integration and  therefore  brings predicted distributions of segregation more in line with real-world observation
as in the classic schelling model  agents in a simulated world want to stay or move to a new location depending on the proportion of neighbors they judge to be acceptable
in contrast to the classic model  agents' classifications of their neighbors as acceptable or not depend lexicographically on recognition first and group type e g   ethnic stereotyping second
the face-recognition model nests classic schelling  when agents have no recognition memory  judgments about the acceptability of a prospective neighbor rely solely on his or her group type as in the schelling model
a very small amount of recognition memory  however  eventually leads to different classifications that  in turn  produce dramatic macro-level effects resulting in significantly higher levels of integration
a novel implication of the face-recognition model concerns the large potential impact of policy interventions that generate modest numbers of face-to-face encounters with members of other social groups
### introduction ###
based on his counterintuitive observation concerning neighborhood segregation  nobel laureate thomas schelling  CITATION  established what would become a large and influential literature connecting various subfields of the social sciences
schelling's observations was this  even in the absence of intrinsic aversion to those who belong to other groups  and without anyone explicitly aiming to locate themselves in a segregated community  high levels of segregation could nevertheless result from a modest desire to avoid being too much of a relative minority
when one observes the sharp ethnic segregation that exists in a regrettably large number of us cities  schelling argued we ought not conclude that this is necessarily the result of anti-ethnic sentiment among either majority or minority group members
schelling's classic segregation model shows  for example  that when people are happy with any location at which up to half their neighbors belong to a different ethnic group  one should nevertheless predict dramatic segregation into nearly homogeneous ethnic blocs that no individual explicitly sought or wished for
the incongruity of macro consequences that do not reflect individual objectives is the overarching theme referred to in the title of schelling's  CITATION  micromotives and macrobehavior
of particular relevance to judgment and decision making researchers  we hope  is this link - or lack of link as was schelling's argument - between individual-level decision-making process and macro spatial dynamics
one might dismiss the relevance of schelling's very simple model to the complexity of real-world neighborhoods and other social communities  such as academic departments  where methodological splits into subgroups sometimes lead to conflict and segregation e g   empirical versus theoretical divides which are common in economics departments  or social psychologists interacting quite separately from other sub-disciplines within psychology departments
yet schelling's model is widely used to inform analyses of policies at virtually all levels of local  state and federal government  as well as among private firms and non-profits such as universities dealing with segregation of many kinds
although schelling's neighborhood segregation model gave rise to a substantial new literature that remains active to this day  there is an important mismatch between theory and empirical observation that has received relatively little attention
schelling's model predicts high levels of segregation starting from virtually anywhere within a very large set of initial conditions and parameter values
yet empirical studies documenting various forms of segregation e g   ethnic types among cities  gender types among work places  or methodological types among academic institutions reveal considerable variation in the extent to which social groups are observed to engage in inter-group mixing
whereas the world presents observers with a rich variety of heterogeneous segregation outcomes  schelling's theory does not easily account for this variation as a systematic function of variables or parameters within the model  which raises interesting questions
can the schelling model be squared with real world data
are there extensions of the schelling model that come closer to reality by predicting heterogeneous segregation levels that vary systematically with observable factors in the environment
this paper presents such an extension
we augment schelling's classic model by endowing agents with recognition memory
this capacity enables simulated agents to apply the face-recognition heuristic
face refers to an evolved capacity that is key for our model  namely recording faces into recognition memory
at the same time  the acronym face for fast acceptance by common experience refers to the insight that shared local experience can facilitate rapid formation of relationships and  thus  transform assessments of others' underlying quality in a process by which a recognized face  and the quality of its associated memory i e   positive or negative  absolutely over-rules the inference that would have been made by stereotyping based on group identity
according to this definition  fast acceptance by common experience refers to rapidly formed recognition-based classifications of others' quality e g   an  acceptable  versus  unacceptable  neighbor without regard to group identity  when classifying those with whom face-to-face experience has taken place in the past
when classifying those whose faces are unrecognized  classification continues strictly according to group identity i e   ethnic stereotyping
when an unrecognized other person is to be classified  the face-recognition heuristic reduces to stereotyping based solely on group identity  exactly the same as in the classic schelling model
however  when there is even a small amount of shared experience  the quality of that shared experience from the past determines how other people are classified
classifications based on recognition memory lexicographically over-rule group identity  which is the basis for classification of unrecognized agents in both face and classic schelling models
given the plausibility of the assumption that context-specific experience from the past can influence the classification of others  it came as a surprise to us that we could not find any previous attempts to extend the schelling model in this direction
the model shows that when agents possess face-recognition that lasts as short as a single period encoding a maximum of only  NUMBER  individual faces out of a substantially larger population  this alone is enough to produce significantly higher levels of integration
the key comparison investigated in this paper concerns this variable degree of recognition memory e g   no recognition memory as in the classic schelling model versus any positive number of periods for which the faces of those one encounters remains coded in memory
by introducing variable recognition memory as a representation of heterogeneity in real-world environments which sometimes have few  sometimes many  opportunities for random face-to-face encounters with other-type agents  the model investigates a novel source of systematic variation into the otherwise classic model of segregation
the motivation for studying the effect of recognition classification on segregation is to better understand why some real-world environments succeed at achieving sustained levels of cross-group interaction i e   high levels of integration while others seem to be locked into a stubbornly unchanging pattern of segregation
the model is intended to contribute substantively and constructively to policy analysis with a simple message  namely  that we can  relatively cheaply  design institutions that produce modest opportunities for face-to-face encounters with members of other groups
then  to the extent that people use an acceptance rule based partially on recognition  random face-to-face inter-group mixing could potentially generate large and stable levels of integration that are too pessimistically ruled out by the vast majority of studies based on schelling's model
the paper is structured as follows
we outline the classic schelling model of neighborhood segregation  review previous research related to our extension of this paradigm  and present the limitations of the classic schelling model mismatch between its predictions and real world segregation data that motivate our extension
we then introduce the face-recognition heuristic and specify the recognition-augmented schelling model  an encompassing model that nests the classic schelling model as a special case
subsequently  we present a series of agent-based simulations demonstrating the effect of agents' recognition memory and decision rules the micro-level on their spatial distribution in the environment the macro-level
