Talk:Space syntax

Refactoring
I would like to completely rewrite this article. I will endeavour to extend rather than to delete what everyone else has written, however, I may lose things people feel are precious in the effort. Obviously, I will move the article to here, so please feel free to comment and to rework in anything you feel is important. The structure I propose is: Introduction, History, Analysis Methods, Criticisms, See Also, External Links, References. Any comments at this stage would be useful. --stochata 10:28, 26 Feb 2005 (UTC)

various
Details of the central theories of Space Syntax should be added. The glossary of terms should be edited down.

It would be helpful if information on the value of Space Syntax and a more open appraisal could be added, as currently the article resembles something that owes more to marketing.

Is there evidence that it is a useful predictor of pedestrian movement? What other models are there? Is the Space Syntax model any better than the judgement of a trained professional? Is there any evidence base that makes a case for Space Syntax? —Preceding unsigned comment added by Evidencebase (talk • contribs) 14:58, 8 May 2010 (UTC)

spacesyntax.org is not available anymore. Should be replaced by http://www.spacesyntax.com/ — Preceding unsigned comment added by Nophaket (talk • contribs) 14:13, 2 September 2011 (UTC)

Strange Article
To a mathematician, this article seems very strange. The "theory" seem to be repackaging of some fairly modest ideas of mathematics in a rather wide-eyed new vocabulary. It expresses the enthusiasm of its creators for the 3-dimensional world and its relationships, but does it have any real importance? For example, did it become standard terminology and get applied by many architects or engineers? Even fluffy design talk is significant if a lot of practitioners got involved -- 84.227.224.164 (talk) 20:42, 17 February 2014 (UTC)

Strange Article really?
I think your missing the fact that they did a lot of work on this in the 1970s before mathematics had invented similar terminology. If this is a 'fairly modest mathematics in a wide-eyed new vocabulary' then perhaps you could supply the name of the mathematical theory which permits the centrality values of one graph of size N to be fairly compared to those of a graph of 2.3N ? In terms of importance I've found 10,000+ references on google scholar to "space syntax" checking the first 30 pages they all look to refer to that subject area. This includes one reference including one to There appears to be, at least one peer reviewed journal of space syntax there also appears to be an on going conference which is bi-annual and has got up to 9 ( so 18 years). I also count at least three books, ,

In terms of practical impact ('Even fluffy design talk...') I found which states that  Fosters and Partners used space syntax on this and a number of other projects.

I think from this evidence you will agree that is a reasonable body of academic knowledge even if the page is in need of a re-write. RebeccaMoreau (talk) 10:31, 16 April 2014 (UTC)

The Problem With And Value Of Space Syntax
Bona fide: I have qualifications and teaching and professional experience in actual syntax (i.e. linguistics), and in architecture and urban design, and I understand computers and computation, and some maths.

Space syntax is problematic because it is neither a science nor a practice, per se.

Hillier brings up many interesting problems and points in his Space Is The Machine (his most grand statement of space syntax), and the concept of the j-graph demonstrates that spatial configuration is a thing in its own right (essentially, how are spaces that are physically contiguous logically or operationally connected, e.g. where are the doors), that doesn't reduce to other spacial concepts, e.g. mere proximity, or shape, or other social concepts, e.g. context, or program.

But the proposals he makes to develop better design on this basis are pretty low-key. The syntactical concept base is either obfuscatory or merely embryonic. The starting point for his ideas (hence the name) is that syntax, being how sentences are put together, i.e. make meaning, is an analogy for well-configured spaces. He would that we saw a potential science of optimal spatial configuration as a 'syntax of spaces'. This is pretty good as an idea, why not: lots of people use language as a premise for arranging components towards some human end.

However, what he concretely puts into this investigatory, um, space is the idea that, cognitively, the straight or axial line is primary to optimal spatial organisation in the urban format, and that, thus, if you have to 'turn' from a straight line to get to somewhere, this adds some kind of cognitive burden or disconnection or suboptimality. At this point, this is neither science nor practice. It seems to be kind of just made up, scientifically, and very very difficult to prove, or even use, operationally. When you elaborate it all out, you get the idea (because, practically, cities are on a grid plan, and to turn means to use another road), that the roads that link to other roads with the fewest turns are 'good' - i.e. you have 'discovered' the High Street (UK) / Main Street (US) and the you know where the suburbs are. That's not much value out of all this speculation.

The other ideas being proposed, from others, e.g. isovist, and convex space, are even worse defined and worse applied, and just super speculative. Here, you get to the point of, as I see it, people cherry picking ideas from maths and computation, and sort of trying to find a spatial analogue for them; which is even less disciplined and application-ready than what Hillier has been trying.

Ultimately, this is not science because the foundational analysis is derived by analogy (or just idea grabbing) from other disciplines: syntax, convex space, etc; and it is not practice, because the human elements don't come from any obvious evidence base: why the fuck is the axial line so primary, who says, far more likely that /some/ turns and /some/ distance is actually optimal not /no/ turns and /any/ distance, but you'd have to /prove/ that yada yada.

Because space syntax is hiding halfway between real analysis and some practical ideas, you can't really argue with SS people: they keep changing the goalposts, and keep hiding the premises on which they are building. Ratti's critical paper is not even very complex: he just points out that a very long straight road would be optimal in SS, but suboptimal in, you know, real city life.

Having said that, SS is /trying/ to bring computability to urban design: it is creating conceptual frameworks that can be decomposed into computable elements, and to this extent, it is a huge advance on the made-up nonsense that counts as most of actual planning theory. So, that's the real summation: at least they are trying.

Also, SS is important because it is /trying/ to bring a level of human relevance to urban planning that is, ironically, historically absent. Planning is starts historically as an extension of government planning (indeed military planning), of the allocation of people and resources, at various scales towards a grand optimum; this is the essence of top-down, birds-eye, control-oriented spatial arrangement - this is the origin of the plan-view! More recently it can be seen as a spatial branch of engineering, or systems engineering at least, where you try gradually try compute a system according to nodes, edges, processes, inputs, outputs, forces, and so forth. Which, again, is essentially conceptually-top down: you can't design or manage a system, unless you know what all the components are, how they (are supposed to) fit together, and how to control this assemblage of parts. Hillier is somewhat passionate and certainly interesting when he says this is not how cities are, and is certainly not how they form: people building houses and facilities next to each others, and then moving among them, is how it actually comes about and then continues, and planning is almost always something overlayed partially on that. Hillier wants urban design to start at least as much with the personal experience of this urban 'organism' as with the planning / engineering of the spatial system: he is advocating for a worms-eye, or individual personal experience, as a primary reference point in the design concept, as much as the whole model and system, for the city. This is what led him to fixate on the axial line: it is his proposal for something that is central to the /experiential/ aspect of the city's performance, not just its /metrical/ qualities. Again, a very good idea, and alongside making urban space computable, this attempts to make it not just humane but human i.e. something relevant to real people.

But, sadly, the concepts are far too weak, speculative and untested for it to be more than a noble and promising contribution. Anyone who understands this much could think of 20 ideas additional to axial line, isovist, convex space - derived from other disciplines - that blend spatial context and configuration with psychological optimality, and would be just as relevant. Such as:

- number of friends and family within a certain distance of your home

- number of native speakers of your language in your daily movement pathways

- size of voting district

- etc

If you want examples of this kind of work, check out Ratti and his visualistions at MIT SCL.

Practically, if you really want to critique the limits of space syntax with regarding to designing urban space, it is depressingly easy: you just observe that it is fixated on some kind of abstract psychological optimum for some kind of abstract personal displacement, largely on foot, which in practice accounts for a vanishingly small fraction of actual urban life, and just can't be used for most design problems. Some examples:

- if you want to know where to place metro or bus stations, SS can't help you (assuming that you are trying to optimise for the /transport/ system when people are /on/ it, not just people' experience of getting /to/ the stops; ask, what is the axial / isovist / convex relevance or data for a subway journey?)

- if you want to know where to put schools, to achieve best catchment socially or for safety or for noise tolerances or other non-spatial-access benefits, SS can't help you

- if you want to know where to put district heating systems, or waste collection depots, or logistics terminals for new service-based lifestyles, or parking spaces, SS can't help you<br /

... try it yourself: think of a real-world urban design challenge, and see if SS tools and concepts help you at all.

The SS salesfolks and researchers will claim they can do all these things, but they can't. So, really, it's not much use, other than a somewhat speculative guide for where to put the main pedestrians streets.

And I have been on design projects, large scales ones, where there is a SS consultant. And I tell you, not just analytically, but professionally, it's not much use.

Bring on real computation for urban optimisation. This is not it.

jmanooch 14:22, 29 October 2015 (UTC)

Complete article rewrite proposal
(I have no affiliation with space syntax. I'm an archaeologist, presently working on a PhD in which I use space syntax methods to analyse an ancient settlement plan - which is just a small part of my research)

In my view the present article does no justice to what space syntax is, neither do the various comments on this talk page. I understand that most people are familiar with the theory through urban planning, and regard space syntax as an applied method for urban design. The first two sentences of the article are telling:

"The term space syntax encompasses a set of theories and techniques for the analysis of spatial configurations. It was conceived by Bill Hillier, Julienne Hanson and colleagues at The Bartlett, University College London in the late 1970s to early 1980s as a tool to help urban planners simulate the likely social effects of their designs."

First, 'a set of theories and techniques for the analysis of spatial configuration' is too narrow and inaccurate as a definition. Second, helping urban planners was not at all the reason for developing space syntax.

As I see it, space syntax is first and foremost a theory about the relation between space and society. This can be read in the introductory chapter of Hillier & Hanson's the social logic of space (1984). Here (and on a multitude of other places) they argue that the current paradigm found in sociology, social geography, and anthropology to name a few, does not succeed in describing nor explaining the relation between people and their built environment accurately. As Hillier in his writings engages with many influential theories in the social sciences (including 'biggies' from Durkheim to Levi-Strauss and Giddens) his aims are clearly much larger than just providing urban planners with some useful tools.

As a matter of fact, the theory being of some relevance to urban planning is just an accident, a by-product of the subsequent testing of several propositions of the theory on real world situations. At some point they discovered that the integration values of the axial map corresponded well with observed aggregated movement flows of pedestrians and certain land uses (first main insights of these investigations published by Hillier et al in 1993 ). Many more cases of such studies followed, testing other variables as well, but generally with the same results. So they concluded that axial lines could more or less predict human presence and movement within urban grids. Initially, they were as puzzled about this as user jmanooch on this talk page: "Why should the axial organization of space be so fundamental?" Hillier et al wonder in their 1993 article (p. 62). In subsequent years, the hypothesis was put forward that this could relate to the way people cognize space and elementary spatial relations, which, against common assumption, is not primarily determined by metric factors. Recent cognitive research in way finding behaviour is starting to confirm this picture.

Also the assertion of user jmanooch that "SS salesfolks and researchers will claim they can do all these things [solving real urban design problems], but they can't" is not really a fair assessment of space syntax in my view. I may of course be wrong here because I have no insight in the workfloor dynamics of the urban planning profession. But this quote of the last sentence of their 1993 article might represent better their view of how space syntax can contribute urban design: "It does not tell designers what to do. It helps them to understand what they are doing" (p. 66).

Anyway, long story short (too late for that..): this article needs a thorough rewrite.


 * 1) It needs to position space syntax theory in relation to other social theories about the relation between space and society
 * 2) It needs to represent the history and original aims of space syntax research more faithfully
 * 3) It needs to discuss its subsequent reception or uses in various (academic and applied) fields, not only urban design/planning (although clearly, here it has had the most impact)
 * 4) And of course the criticism that space syntax theory and method received should be systematically dealt with, including replies by Hillier & others. Now it only contains a very imprecise review of Ratti's 2004 article. There have been many others, and from various angles
 * 5) Last but not least, the method needs to be discussed concisely, without going into too many technical details. Presently the methodological/theoretical part below the vague heading 'thesis' just discusses some random concepts of relevance. Very uninsightful, very pointless.

I hope I can contribute at some point. But I would like to see some feedback first.

--TijmLanjouw (talk) 18:26, 17 December 2015 (UTC)