ing to historical sources , the city expanded towards the end of the fifth century BCE due to a migration movement ( known as the anoikismos of 432 BCE ) with the intention to form a larger and better defendable city in preparation for the rebellion against Athens ( Thuc . 1.58 ; Diod . Sic . 12.34.2 ). However , it is unclear which cities or specific populations joined this movement and how large the flow of migrants must have been . The archaeological record of Olynthos shows how a sudden growth in population size around this time resulted in the construction of a new residential area on the North Hill , laid out in Hippodamian fashion , with avenues running northsouth and streets following an east-west direction ( Robinson and Graham 1938 , 13-14 ). 1 The newly constructed area was , however , short-lived , as the army of Philip II brought about the city ’ s destruction in 348 BCE . By then some inhabitants might have already left the city with their belongings in advance . Those who remained must have been undoubtedly affected by the siege .
Methodology and theoretical framework The archaeological dataset available for the spatial analyses are the streets and the houses on the North Hill of Olynthos as published by Robinson ( 1929- 1952 ) and Cahill ( 2002 ). The theoretical framework at the basis of space syntax was formulated by architectural and urban morphologists Hillier and Hanson ( 1984 ). The spatial boundaries for the analysis of movement flows are the natural contours of the North Hill in the west , north and east ( fig . 1 2 ). These coincide with the possible circuit of the city ’ s fortification walls ( Robinson and Graham 1938 , 39- 40 ). The southern boundary is Street – I . To reconstruct the movement flows within Olynthos ’ street network , an axial analysis is conducted . In this analysis the street network is divided into the least and longest straight lines that connect the entire street space . Depthmap software , 3 subsequently , calculates the level of accessibility and integration of all the lines within the network . The results are shown in a colour-coded map with a spectral range from red , for the highest integrated lines , to blue , for the lowest integrated lines . The integration value correlates to the potential amount of pedestrians moving
1 In this article a distinction is made between the terms ‘ street ( s )’ as opposed to avenues ( in italics or with a capital when it refers to a specific street , e . g . Street V ), and ‘ street ( s )’ for the collection of all avenues and streets in general . 2 The small alleys , stenopoi , running from east to west in the middle of each house block are excluded from the analysis , as they were mainly used for drainage and not for passage ( Robinson and Graham 1938 , 33-39 ; Nevett 1999 , 55-56 ). Another hypothesis is that the alleys serve as light sources , allowing light to enter the houses through windows placed along the walls ( Graham 1958 , 322 ). 3 Depthmap software is created by UCL ’ s Bartlett School of Architecture . along each line ( natural movement flows ) ( Hillier 1996 , 119 ). A visibility graph based on visual integration and a positive correlation between visibility and movement potential complements the findings ; the higher the visual integration of spaces , the more movement they attract ( Stöger 2011 , 194 ). Again the spectral range goes from red , for the most visible areas , to blue , for the visually most segregated spaces .
Next , the attention turns to an analysis of doorways . Doorways are physical and symbolic points of transition where private life opens up to , closes of from , and intersects with public life ( Laurence 2007 ; van Nes 2011 , 101 ). The location of doors , therefore , reflects how the urban environment was experienced and where potential activity areas of human interaction between the private and public sphere could have occurred . The spatial boundaries for this analysis are House A1 and Street IX , Avenue B , House A12 and the southern edge of Block A IV , and the western edge of Row A ( fig . 3 ). These boundaries enclose a completely excavated area with sufficient data on the architectural remains necessary for the analysis .
The analysis begins with looking at the connection of building entrances to streets . The entrance can be either directly accessible to a street or separated from it by a fence or front yard , providing some form of seclusion and privacy . An examination of which street ( s ) a building connects to can also give information on its relation to public space . The number of doorways opening onto a street , then , directly reflects the amount of social activity and chances for interaction that might have occurred in the streets ( van Nes 2011 , 111 ). For a comparative survey of these activity areas across the entire street network all streets have to be divided in segments of similar lengths . By measuring the occurrence of doorways ( number of doors per x meters ) per segment , then , the potential activity areas are representatively revealed ( Laurence 2007 , 103 ).
Lastly , the way that entrances are positioned to each other influences the probabilities for social control ( van Nes and Lopéz 2007 , 23.7-23.8 ; van Nes 2011 , 111-113 ). This article , therefore , reports the number of times that entrances are placed directly across from each other , enjoying an intervisibility rate of a 100 percent . This implies that the inhabitants might have experienced high levels of social control .