is also a powerful element ofidentity polit-
ics. Race, class, gender, sexuality and other
differences can provide the basis for solidar-
ities andresistance. Arguments that point
out the constructed nature of such categories
and that highlight the divisionswithingroups
(such as racial differences within feminist
movements, or class differences within racial
politics) can seem to threaten the basis of
political struggles forged around identities
such as ‘Black’ or ‘Woman’. In response, some
feminists have adopted ‘strategic essential-
ism’, choosing to emphasize the commonal-
ities across women’s experiences in order to
unite women for political purposes.
Difference poses political problems for
those who wish both to value diversity and
to dismantle the structures of discrimination
and oppression that scaffold ideas about dif-
ference. The question of how to recognize and
make space for the very real effects of differ-
ence, orpositionality, and at the same time
to dismantle hierarchies based on difference is
of ongoing concern to political theorists and
geographers. Iris Marion Young (1990a), for
example, has advocated for the celebration of
diversity within an overarching political unity.
Such a project is not, however, without its
difficulties in the context of ever shifting axes
of difference and alliance. ajs
Suggested reading
Fincher and Jacobs (1998); Women and Geog-
raphy Study Group (1997).
diffusion The spread of a phenomenon
(including ideas, objects and living beings)
overspaceand throughtime. There is a long
tradition of diffusion studies in American
cultural geography, most closely associated
with the work of Carl Ortwin Sauer (1889–
1975) and Fred B. Kniffen (1900–93).
According to Sauer (1941), it was Friedrich
Ratzel (1844–1904) who ‘founded the study
of the diffusion of cultural traits, presented
in the nearly forgotten second volume of his
Anthropogeographie’ published in 1891 (see
anthropogeography). In Sauer’s view, diffu-
sion – ‘the filling of the space of the earth’ –
was a ‘general problem of social science’:
‘A new crop, craft or technique is introduced
to a culture area. Does it spread, or diffuse
vigorously or does its acceptance meet resist-
ance?’ The specific contribution ofgeography
was to reconstruct diffusion pathways and to
evaluate the influence of physical barriers
(Sauer, 1952; Wagner and Mikesell, 1962).
Both tasks were pursued by various members
of theberkeley school, but they reappeared
in a starkly different guise in the much more
formal study of innovation diffusion inaugu-
rated by Torsten Ha ̈gerstrand (1916–2004).
One of Sauer’s closest associates introduced
Ha ̈gerstrand’s Swedish monograph to Anglo-
American geography: ‘No one who essays
in the future to interpret the distribution of
culture elements in the process of diffusion
can afford to ignore Ha ̈gerstrand’s methods
and conclusions’ (Leighly, 1954). Even so, it
was some fourteen years before an English
translation of Ha ̈gerstrand’sInnovation diffu-
sion as a spatial processappeared (Ha ̈gerstrand,
1967; see Duncan, 1974). Ha ́gerstrand’s work
had two catalytic consequences: it set in
motion the frozen worlds ofspatial science,
and it opened the door to sophisticated
computer modelling of spatial processes. The
theoretical structure of his originalmodelis
shown in the figure. An interaction matrix
provides the contours of a generalized ormean
information field, which structures the way
in which information circulates through the
population in a regional system. These flows
are modulated by physical barriers and indi-
vidual resistances, which together check the
transformation of information intoinnovation
and so shape the successive diffusion waves
that break on to the final adoption surface.
Most immediate discussion focused on the
operationalization of the model – on the use
ofsimulation methods, the comparison of
‘observed’ and ‘predicted’ patterns of adop-
tion, and the detection of a localizedneigh-
bourhood effect. Within this modelling
tradition, the most important developments
included the following:. A formalization of the mathematical rela-
tionships between the structure of the
mean information field and the form and
velocity of diffusion waves, revealing the
connections between differentdistance-
decaycurves and the classic neighbour-
hood effect (although it is scarcely sur-
prising that a distance-bound interaction
matrix should generate a contagious pattern
of adoptions).
. A demonstration that the Ha ̈gerstrand
model is only a special instance of the
simple epidemic model, and the subsequent
derivation of more complex epidemic
models, particularly through the remark-
able contributions of Cliff, Haggett, Ord
andVersey(1981), whose replication of a
range of ‘spatial processes’ (seeprocess)
confirmed:
Gregory / The Dictionary of Human Geography 9781405132879_4_D Final Proof page 160 1.4.2009 3:15pmDIFFUSION