Print

---

 In
> trying to get a grasp on Space Syntax through its methodology I
> feel I have been peeling an onion - always another method but
> never a central core theory which these methods are intended to explain.
>
Well, the central core theory is well set out in the Social Logic of Space.
Of course the theory has moved on since then and are also in Bill's and
Julienne's more recent books, in various papers produced by people doing
syntax research all over the world, most of which I know you are aware of
from what I have seen you write to the mailbase. But when you come on to
'methodology' it is bound to be like an onion - that is because one tests
hunches and devises methods to do so. There is no single method. These is
even no single method for devising 'correct' methods. Science isnt like
that. It is a creative activity a bit like architecture. Imagine the
following dialogue:

Alan:  Tom, you are an architect?
Tom:   Yes.
Alan:  Tell me how to design a house..
Tom:   Well, tell me a bit more. Who is it for? Where is it?
Alan:  No, just tell me how to design a house. You must have a method.....

Now in the email thread you started here I get the feeling that we keep
going in circles, and that they dont get much tighter. The reason is that
science - and I would claim that syntax is a bit of that - is like an onion.
Take physics, Newton, Relativity, Quantum theory, unified theories,
superstrings, twistors and the rest. All are stabs at understanding the
world, but the process of how our understanding of physics has advanced over
the last 300 years is not a neat and simple story. It is complexly
interrelated - actually an onion is far too simple a shape to describe what
goes on.

> Alan suggests I am "trying to neaten things up too much", but for
> my part I have found throughout my readings a pervasive feeling
> that there is a central subject which is never explicitly stated,
> and it is this that I am trying to find.  I was not clear that I
> felt like that until Alan wrote that he is "not holding back a
> deeper understanding"; well it may be absurd, but that is JUST
> the impression I got (and perhaps it is what has frustrated other
> outsiders).
>

Perhaps that is because you are looking for a single firm answer and dont
like it when someone says that we dont know all the answers. Let me say it
again - there is no single method any more than there is a single method to
design a house.

> I should say right now that all this is just the way it has
> struck me - I'm not saying that this is how it is, I'm trying to
> find out how it is.
>
> So where did I get that idea from?  The writing on space syntax,
> both in the press and specialist papers, seems to concentrate on
> techniques and on related issues in society, but seems to miss
> out the essential link between them.  It gives the impression
> that space syntax is a branch of mathematics which just happens
> to be applied to maps of towns.
>
> Alan writes that "words axial lines represent linear properties
> of a spatial pattern etc. they happen also to represent lines of
> sight along which people may move".  It seems to me quite the
> other way round.  The study of linear properties of spatial
> patterns IS a branch of mathematics - graph theory I believe.

well not quite, but never mind. Graph theory is about elements and
relations.

> Graph theory is an important tool used by spacesyntax, but not
> its purpose (is it?).
>
Exactly right.

> Surely the reason for drawing the lines in the first place is
> what Jake said; to explore how space may influence people's
> lives.  Perhaps this blurring of purpose dates back to the
> setting up of a Space Syntax Lab., instead of a Social Logic of
> Space Lab.?

Not exactly. I would say that a lot of the 'mathematical' concern with
syntax representations came more recently in attempts to understand 'why the
axial line', as opposed to any other of a multitude of representations,
seemed to make sense of empirical (social) data. This is different to the
attempts to automate the production of axial and convex representations
which came earlier, about 1985 (before the SSL was established in 1987). The
syntax lab as it is now does several things, amongst which are acting like a
'clinical' research unit where we deal with live cases, and it now does most
of the resaerch into the underlying social theory as an aspect of this. One
thing about live cases is that they raise questions about society in a very
precise form.  Most of the more mathematical and comupter software
development is carried out through Research Council funded projects.

>
> It seems to me that the 'core theory' is that there is a Social
> Logic implicit in the Syntax of Space.  That is not the only
> thing you can analyze with the syntax of space, but it is the
> purpose of (most) space syntax papers as far as I can see.  And
> purpose is important, it is the purpose of a study which lets you
> determine exactly what tools to use to test your ideas, and to
> judge your success.
>
> To be more explicit about the social logic as I understand it, it
> is a 4-stage theory that;
>
> 1) Spatial configuration affects what people can see and where
> they can go (pretty obvious).
>
> 2) This affects how people can find their way through a spatial
> complex, AND affects the possible interactions between different
> people & groups of people (would anyone argue with that?).
>
> 3) The combined effects of navigation & interaction factors
> influence where people go in fact:  global configuration affects
> local movement patterns (all other things being equal, a change
> in configuration changes patterns of movement)
>
> 4) Movement of people is a causal factor in many important social
> issues.  (What is burglary rate but 'where the burglar went'?
> What is interaction but "where two people went"? and other
> indicators like shop rents or shopping centre failure are again
> directly caused by whether people go past / into the shop).
>
> I am not suggesting it is simple, of course. There is plenty of
> complexity in working out what aspects of configuration have
> particular effects - lots of room for diverse methods of
> measurement & correlation - which I guess is what is going on
> with all the work at UCL and elsewhere.
>
> Neither am I suggesting that this theory is the only thing you
> can investigate about buildings using graph theory. Hillier gives
> an interesting account of using a sort of finite element
> integration to measure the perceptual centre of irregular solids
> (if I remember rightly) in Space is the Machine.  But this isn't
> Space syntax is it?  Or should I say, it is not 'Social logic?
> Is that different?
>
> I am going on about this because this seems to me to be lying
> unsaid within much of the writing.  Alan writes that "it would be
> very unwise to postulate 'mechanisms' that say how space leads to
> behaviour. This doesn't mean that we don't have some hunches
> about what might be happening - these are the hunches that guide
> the research "
>
> But how do you carry out research if you are not FIRST clear
> about what you think might be happening?  Doesn't research
> consist of having a hunch and then testing it out to see if it
> works consistently?

That is why I said : Alan writes that "it would be
> very unwise to postulate 'mechanisms' that say how space leads to
> behaviour. This doesn't mean that we don't have some hunches
> about what might be happening - these are the hunches that guide
> the research "
>

And isn't that then a testable theory?
> Besides which, if I am even roughly right in what I just wrote,
> the theory is a lot more than a hunch.  It may not have the
> certainty of mathematics, but hey! this is social science!  It
> has a fairly convincing logic to it just in layman's terms (which
> is about my limit), and it must surely be fairly well evidenced by now.

Yes, but I was talking about mechanisms that might explain 'why the axial
line' as opposed to somthing else seems to be related to behaviour. We know
that it is related - the empirical evidence is robust - but through what
individual level mechanism? About that we have a number of (often
incompatible) hunches. These are not yet near to the status of hypotheses.

>
> Which brings me right back to my first question - why draw
> straight lines?  Why not make a map with curvy lines connecting
> nodes?

Well, it should be clear that there is nothing to stop you mapping it any
way you like.
The only plea is to be consistent, and able to pass on to someone else your
'way of doing it' so that they can reproduce your results. It could be that
a curvy line map will work even better - try it, test it, and then once you
have shown that I look forward to a long email discussion on the subject of
'why the curvy line?' :-)

> I just cannot believe that it happened by chance and
> then was found to make a good correlation with movement patterns.

Well, now you are asking a history of science type question - I believe that
the answer is that Bill was jogging through the Marquess estate (where
attempts to explain interesting things using convex spaces had only got a
little way) and it came to him that some of the convex spaces were lined up
and others not... It was not by chance, but as an attempt to represent
somthing about what one can do with space (line it up or maki it go round
corners) in a simple way. The history is that the finding about correlations
with movement came well after axial maps. Look back at the Social Logic.
Axial maps are there, but correlations with movement do not appear except as
an addendum just before publiaction.

>  Isn't it because it is thought that the number of intermediate
> points of visual fixation between start & finish of a journey
> corresponds to how difficult the journey seems?  There may or may
> not be evidence for this, but it could be collected couldn t it?

That's one of the hunches. Ruth Conroy has done a PhD partly concerned with
this.
>
> For me it is also important because it gives some humanity to the
> study.  It is what is going on for people, not for statistics.
> It tells us about otherwise unmentioned properties of (built)
> space which have not been accessed before.  It makes space syntax
> worth bothering with.
>
> Am I getting warm yet?
>

That's the telling question. 'Getting warm' presupposes that we have a
'true' answer hidden away somewhere. We havent hidden it because we havent
found it yet (and when we do we wont hide it!)

All the best,
Alan