Syntactic Step Depth
- Introduced by
- Bill Hillier & Julienne Hanson
- Year
- 1984
- Type
- Topological measure of depth and integration between spaces
- Family
- Space syntax / configurational analysis
What does it do?
Syntactic step depth is the space-syntax measure of how topologically far apart spaces are — how many turns, transitions or moves separate one space from another, regardless of metric distance.
Formalised by Bill Hillier and Julienne Hanson in The Social Logic of Space (1984), it is computed from a justified graph in which every space is a node and every direct adjacency an edge, and a single step is one move between connected spaces. Aggregated into mean depth and normalised into an integration value, step depth becomes the workhorse of configurational analysis, predicting which spaces will be most used, most accessible and most central in a building or city.
The intuition behind the method
Forget metres for a moment and count moves instead. From your front door, the hallway is one step away, the room off the hallway is two, the cupboard inside that room is three. A space that is only a few steps from everywhere else is shallow and integrated — easy to reach, likely to be busy and public. A space buried many steps deep is segregated — private, hard to find, lightly used. Syntactic step depth captures exactly this: it redraws a building or street network as a graph of who-connects-to-whom and asks, from each space, how many steps on average it takes to reach all the others. Shallow spaces become high streets and lobbies; deep spaces become bedrooms and dead ends.
History
Step depth and integration were set out by Bill Hillier and Julienne Hanson in The Social Logic of Space (1984), which proposed that the configuration of space — captured as a graph of permeable spaces and analysed through depth — encodes and reproduces social relations. The justified graph, mean depth, relative asymmetry and integration formed the core of the space-syntax method developed at the Bartlett, University College London, and were applied first to building plans and settlement layouts. Hillier's 1996 Space is the Machine extended the theory to whole cities and the 'natural movement' thesis, and later work shifted from axial integration toward angular segment analysis and choice (betweenness), broadening the original step-depth idea into a richer configurational toolkit.
When to use it
Use syntactic step depth and integration when you want to understand the configurational logic of a building or street network — which spaces are accessible and likely busy, which are private or segregated, and how the parts relate as a whole — independently of metric distance. It is well suited to comparing design alternatives, explaining and predicting pedestrian movement and co-presence, and analysing how layout encodes social relations of privacy, control and encounter. It is less appropriate when metric distance, gradient or travel time genuinely drive behaviour, where angular or metric segment analysis and network methods such as sDNA are preferable, and care is needed because results depend heavily on how the spatial model (axial map, segment map, convex map) is drawn.
Limitations and considerations
- Pure topological steps ignore metric distance, slope and travel time that often matter for real movement.
- Results depend strongly on the (partly subjective) choice and drawing of the axial, segment or convex model.
- Traditional axial integration can be unstable to minor changes in how lines are extended or broken.
- Edge effects make integration of peripheral spaces in a truncated map unreliable near the boundary.
Source conditions do not replace independent scientific review or validation by an analysis engine.
Strengths
- Captures the purely topological, configurational structure of space that metric distance misses.
- Integration values predict observed pedestrian movement and land-use intensity remarkably well.
- Normalisation (relative asymmetry, integration) allows comparison across systems of different sizes.
- The justified graph makes depth, privacy and accessibility structure visually and intuitively legible.
Applications
- Predicting pedestrian movement and footfall from the integration of streets and spaces.
- Analysing building plans for privacy, accessibility and the social logic of layout.
- Comparing urban design and masterplan alternatives by their integration structure.
- Studying how settlement and housing layouts encode social relations of encounter and control.
- Locating land uses and active frontages where integration predicts high natural movement.
How it works
How it works
The configuration is represented as a graph whose nodes are spaces — convex spaces, axial lines, or street segments — and whose edges join spaces that are directly connected. The step depth between two nodes is the number of edges on the shortest topological path between them, counting permeability moves rather than physical distance. Redrawing the graph 'justified' from a chosen root, with all nodes arranged in horizontal layers by their depth from that root, makes the structure visible at a glance.
Formula
d(i,j) = \text{shortest path length in topological steps between nodes } i \text{ and } jThe mean depth of a space i is the average step depth from i to every other space in the system, summing all shortest-path step counts and dividing by the number of other nodes. A low mean depth means the space sits, on average, only a few moves from everywhere else (shallow, accessible); a high mean depth means it is many moves from the rest of the system (deep, segregated).
Formula
MD_i = \frac{1}{n-1} \sum_{j \neq i} d(i,j)Because mean depth grows with the size of the system, it is normalised into relative asymmetry, which rescales mean depth against the theoretical minimum (a fully shallow star) and maximum (a deep chain) for a system of n nodes, yielding a value between 0 and 1. Low relative asymmetry marks integrated spaces, high values mark segregated ones, on a scale partly comparable across systems of the same size.
Formula
RA_i = \frac{2(MD_i - 1)}{n - 2}To compare systems of different sizes, relative asymmetry is further divided by the value for a standardised 'diamond-shaped' graph of the same size, giving real relative asymmetry; its reciprocal is the integration value, where higher means more integrated. Integration is the central space-syntax measure: it correlates strongly with observed movement and land-use intensity, and the difference between local (limited-radius) and global integration distinguishes neighbourhood from city-wide structure.
Formula
i_i = \frac{1}{RRA_i} = \frac{D_n}{RA_i}Common pitfalls
- Comparing integration values between maps of different size or extent without proper normalisation.
- Drawing the axial or segment map inconsistently, so step counts and integration shift with the modeller's hand.
- Ignoring edge effects and reading low integration at the map boundary as genuine segregation.
- Treating topological step depth as if it were walking distance when reporting accessibility to non-specialists.
Frequently asked questions
What is the difference between step depth and integration?
Step depth is the raw count of topological moves between two spaces, and mean depth is the average of that count from one space to all others. Integration is mean depth normalised so it can be compared across spaces and across systems of different sizes: mean depth is converted to relative asymmetry and then to real relative asymmetry, whose reciprocal is integration. In short, step depth is the basic ingredient and integration is the standardised, comparable measure of how central or accessible a space is within the configuration.
Why measure topological steps instead of metric distance?
Space syntax argues that much of how people perceive, navigate and use space depends on configuration — how spaces connect and how many transitions separate them — rather than on exact metres. Counting topological steps captures this relational structure and, empirically, integration computed from step depth predicts movement and land-use patterns very well. That said, modern practice often complements topological depth with angular and metric measures, since gradient, distance and travel time clearly matter too.
How does syntactic step depth relate to the wider space-syntax method?
Step depth is the foundational calculation of space syntax. From it come mean depth, relative asymmetry and integration, which are the standard configurational measures applied to axial maps, segment maps and convex maps of buildings and cities. The broader space-syntax method adds related measures such as connectivity, choice (betweenness) and angular segment analysis, plus visibility-based tools like isovists and visibility graph analysis, but they all build on or sit alongside the depth-based core introduced by Hillier and Hanson.
Related methods
Sources
Hillier, B., & Hanson, J. (1984). The Social Logic of Space. Cambridge University Press.
ISBN: 9780521367844Copied.