Comp. by: VPugazhenthi Stage : Revises1 ChapterID: 9781405132879_4_S Date:1/4/
09 Time:15:23:36 Filepath:H:/00_Blackwell/00_3B2/Gregory-9781405132879/appln/3B2/
revises/9781405132879_4_S.3d
statistical problems. There are two broad
types of application: the creation of ‘artificial
worlds’, and the use of simulation as part of
the estimation of quantitative models. An
early application of the former was Torsten
Ha ̈gerstrand’s work on the diffusion of
innovation. He treated this as a spatialsto-
chastic process, in that while it had an
underlying structure (adoption would
decrease with distance as social interaction
decreases) it could also turn out differently
each time. Monte Carlo procedures were used
to draw random numbers and innovation dif-
fused outwards from initial adopters according
to a probabilisticmean information field.
Such simulation has grown rapidly in recent
years and it is at the core ofmicro-simulation
methodsandagent-based modelling. Such
approaches have now developed to the extent
of the SimBritain project, where as part of
evidence-based policy the geographical
impact of government policies can be evalu-
ated (Ballas, Clarke, Dorling and Rossiter,
2007).
The second use of simulation, statistical
model calibration, has also seen major recent
developments, which have been considerably
aided by increased computing power. One
aspect of this is a computer-intensive appro-
ach to confirmatory data analysis and
hypothesis testing, in which the observed
data are re-sampled to characterize the empir-
ical distribution of a test statistic. More
importantly, a new method of estimation
calledmarkov chainMonte Carlo simulation
has been developed that allows the estimation
of very complex models, includingmulti-
level models. This methodology allows a
building block approach to estimation, so that
complex problems can be decomposed into
lots of small ones. It is especially important
forbayesian analysis, for it allows the estima-
tion of previously intractable joint posterior
distribution (based on prior beliefs, data and
assumptions) by iterative simulation from the
much more straightforward marginal distribu-
tions (Davies-Withers, 2002). kjSuggested reading
Gilbert and Troitzsch (2005); Gill (2002);
Noreen (1989).situated knowledge A term coined by the
feminist cultural critic of science, Donna
Haraway (1991, p. 188), to denote ‘a doctrine
of embodied objectivity that accommodates
paradoxical and critical feminist science pro-
jects’. Situated knowledge replaces thetraditional conception ofscienceas the pur-
suit of a disembodied, inviolable and neutral
objectivitywith a formulation of objectivity
that stresses corporeality,social construc-
tionandcultural politics.
Haraway argues that vision or sight is a
guidingmetaphorfor Western scientists in
carrying out their work: they see the world,
they make observations and collect evidence
(from the Latin verb,videre, to see) and they
write down its truths. This conceit carries over
into ordinary speech too when we say, for
example, ‘I see’, meaning ‘I understand’. The
language of knowledge production is saturated
with visual metaphors in both a technical–
scientific and an everyday sense. But
Wittgenstein reminds us that ‘the limits of
my language are the limits of my world’, and
in working within this language, in construing
vision and visualityin this way, scientists
limit their worlds by writing themselves out
of their own stories: failing to recognize their
constitutive role in world-making, they reduce
themselves to the status of ‘modest witness’
(Haraway, 1997, ch. 1). That presumption of
modesty, Haraway argues, is a direct conse-
quence of the starting point of visuality. It
creates the illusory possibility of a disembod-
ied science. She calls this illusion a ‘god trick’,
the idea that it is possible to have ‘vision from
everywhere and nowhere’ (Haraway, 1991,
p. 191). Moreover, it is just such a trick that
is the basis of one of science’s most cherished
ideas, objectivity, the belief in the possibility of
a single, final, detached and unmarked order-
ing of the world. For Haraway (1991, p. 188),
however, the ‘gaze from nowhere’, as she calls
objectivity, is really arhetoricalmove that
hides and protects the interests of those who
propose and benefit most from it, typically
white Western men. ‘Modesty pays off ... in
the coin of epistemological and social power’
(Haraway 1997, p. 23: see alsoepistemol-
ogy). In this sense, then, being a ‘modest
witness’ turns out not to be very modest at
all. It is a strategy to promulgate a particular
kind of knowledge and to guarantee its
unassailable truth, which is often indelibly
marked byheteronormativity,masculinism
andracism(as her case studies from the life
sciences testify).
Haraway (1997, ch. 4) labels scientific
practices that masquerade under the cloak of
objectivity ‘fetishistic’ because such know-
ledge is represented as a thing rather than a
social process. Fetishism would not occur if
it were recognized from the outset that all
knowledge is embodied and partial; that is,Gregory / The Dictionary of Human Geography 9781405132879_4_S Final Proof page 683 1.4.2009 3:23pmSITUATED KNOWLEDGE