A Reply to Carlson and Gorman

Author(s): Bruno Latour

Source: Social Studies of Science, Vol. 22, No. 1 (Feb., 1992), pp. 91-95
Published by: Sage Publications, Ltd.
Stable URL: http://www.jstor.org/stable/370226 .
Accessed: 25/06/2014 02:55

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Sage Publications, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to Social Studies of
Science.

http://www.jstor.org

This content downloaded from 185.44.77.89 on Wed, 25 Jun 2014 02:55:29 AM

All use subject to JSTOR Terms and Conditions
Responses & Replies: Latour: Reply to Carlson & Gorman 91

approaches to science-technology studies, see Michael E. Gorman and W. Bernard

Carlson, 'Can Experiments Be Used to Study Science?', Social Epistemology, Vol. 3
(1989), 89-106.
19. One limitation of this approach, of course, is that it reflects only the inventor's
perspective. Bell, for example, was both recruiter and recruitee; when he recruited a
backer like Gardiner Hubbard, he was also recruited into Hubbard's network. As
noted earlier, Latour's answer is to create multiple STGs. An alternative might be to
represent the network as a set of links among individuals, a possibility we are now
exploring.
20. We have been using our cognitive mapping scheme to teach engineering
students about the heterogeneous nature of technological problem-solving. First, a
team of students has worked closely with us analyzing inventors' sketches and
developing our mapping techniques. Second, we have begun working with several
computer science students to create a computer simulation of inventing a telephone. It
is our hope that activities such as these will permit new concepts in science-technology
studies to enrich engineering education, and we welcome any advice or suggestions
about how this aim might be furthered.

Authors' address: Division of Humanities,

School of Engineering and Applied
Science, University of Virginia,
Charlottesville, Virginia 22901, USA.

Responses and Replies (continued)

A Reply to Carlson and Gorman

Bruno Latour

In working through the many drafts of our paper, and in answering

the many comments and criticisms of the (now not so anonymous)
referees,2 we wanted to interest our colleagues in developing quanti-
tative and visualizing tools that are in keeping with the qualitative

Social Studies of Science (SAGE, London, Newbury Parkand New Delhi), Vol. 22
(1992), 91-95

This content downloaded from 185.44.77.89 on Wed, 25 Jun 2014 02:55:29 AM

All use subject to JSTOR Terms and Conditions
92 Social Studies of Science

work we have all been doing in the last two decades. This search
for'quali-quantitative' tools that could reconcile the detailed case
studies of scientific practice with the more traditional goals of history,
sociology, science policy and economics is the basis of my agreement
with Carlson and Gorman, and is at the heart of the so-called 'actor-
network' theory.3Our own Note is only a small step in this direction,
and I accept most of their reservations about the tricky character of
data treatment, limiting here my Reply to the substantive issues.
One of the many reasons why, after having disputed quantitative
work, it is now possible to develop 'quali-quantitative' tools, has to
do with new computer software. It is now possible to keep data in
several intermediary stages without having to choose, as was the case
in the recent past, between reducing them to meaningless numbers or
getting bogged down in endless details. The excellent method
developed by Carlson and Gorman has this interesting property of
allowing the scholar to circulate between data at different levels of
aggregation, and to zoom in for details or to zoom out for outlines. In
this sense, I agree with their criticism of our paper. It should indeed be
possible not to lose too many data while creating the map of the
territory. In addition, as they rightly say, it is now easy to keep visual
as well as literary documents in the same loosely connected data-
banks, thanks to the many 'hypermedia' which are being developed.
Still, for those of us trained in the careful interpretation of archives,
interviews, settings and accounts, and who have spent so much time
avoiding the reduction of sociology and economics, any map will
appear to be a pale representation of the case study. This is why we
insisted, maybe too much, on the necessary difference between maps
and territories.
This being said, there is a clear difference of emphasis between their
specification and ours. We clearly stated that our goal was not to
compete with the 'thick narrative' that may be obtained after years of
study of one specific case. As the authors say, we simply wish to
provide the 'contours of an innovation episode' (86) in order to allow
a quick and easy comparison with other cases. I must confess that,
having studied for two years the automated subway Aramis, I am
able to write a book about it but still unable to draw its STG! Their
goal is different, since they wish to develop a research tool that could
give access to the 'mental processes' of the innovators.
And this is where I start to disagree with their criticisms. The main
advantage of a mapping process is to force the cartographers to go to
the limit of their premisses. When we claim that the number of actants

This content downloaded from 185.44.77.89 on Wed, 25 Jun 2014 02:55:29 AM

All use subject to JSTOR Terms and Conditions
Responses & Replies: Latour: Response to Pickering 93

is the only source of information to be kept on the map, we simply

push all explanations to their limit exactly as a cartographerwould do
by coding beautiful landscapes as longitude, latitude and altitude.
Several times Carlson (under the guise of an anonymous referee), and
again Carlson and Gorman, claim to show the limit of this position by
offering an example where an innovator with 'feweractants' wins over
another one who has more. How to answer this repetitive criticism
that 'might makes right'? I can only state again that when Carlson
and Gorman utter the very sentence 'a weak innovator like Bell may
be stronger than a stronger one like Edison', they immediately set up
another measurement instrument, a balance of some sort. It is this
balance which allows them to register the 'stronger force of the
weaker', and it is this 'meta' balance that we are defining. We simply
claim that such an instrument is always calibrated through some sort
of metrology. This can be done unconsciously as when the authors
multiply words like 'Bell's strength came from having a strong patent
and carefully linking his patent with his scientific social role, his
father-in-law'sdeterminationto prosecute all infringers, and the skills
of his backers in building an effective company' (my emphasis) (86).
But it can also be done consciously. We do not know in advance if
having many patents is a source of strength or not. The syntagm
should be written in which few patents, plus a father-in-law, plus
backers is better than many patents without father-in-law and
without backers! This syntagm, in part, defines what can be expected
from patents, in-laws and backers. In the tired old slogan 'might
makes right', the first word is as empty as the second one. It is because
we do not know if an element is strong or weak that we have to deploy
the historical network of its allies. Should I add that in science studies
the surprise often comes from seeing the weak becoming stronger
than all the others? It is Plutarch after all, not us, who invented this
notion of an Archimedian point that will reverse the balance of force.
This leads me to my second disagreement. The authors claim that
our graphs might be good for mapping the social aspects, but theirs is
better for the cognitive or technical aspects. This claim is rather
disheartening for both mapping devices, since our goal was to make
this very distinction not only obsolete, but also unthinkable! We were
careful to give human and non-human actants a symmetric treat-
ment, and to replace the social/technical divide by another more
operational difference between association and substitution. If
Carlson and Gorman could show a breach of symmetry in the
encoding of the narratives I could accept the criticism, but what they

This content downloaded from 185.44.77.89 on Wed, 25 Jun 2014 02:55:29 AM

All use subject to JSTOR Terms and Conditions
94 Social Studies of Science

show is simply a lack of detail on the social and on the technical

aspects. As long as the symmetry is maintained, I do not take this as a
critique of the method. The 'zoom' property that is common to both
mappings should provide the details needed on human, collective and
non-human actants.
I am much more worried by the authors' counterclaim. In order to
contrast their approach with ours, they offer the following research
programme:

We would begin by trying to map the mental processes of the Berlin Homeowner
Association, tracing their various efforts to keep apartment buildings secure and
their decision to either design or purchase the special lock ... We would do the
same for Manfred, detailing both his goals and the steps he took to modify the key
and lock. (87)

Either the authors have access to completely new types of documents

unknown in France which allow them to directly observe mental
processes, goals and details, or they have simply misunderstood the
whole argument of our method. If there exist documents about the
lock, Manfred or the Homeowner Association, then they can be
encoded according to the method we described and they may differ
wildly about the attribution of goals, the attribution of mental
processes, or the attribution of technical efficiency. If, on the other
hand, there exists no other document than the one we showed in our
Note, then there is no way that Carlson and Gorman could 'try to
map the mental processes' or 'detail goals'. This is what we meant
when we insisted not enough after all that an instrument is
always a re-representation of data. No instrument can provide direct
access to everything. They add new mediations. This is what many of
us have shown about the inscription devices of the hard sciences, and
it holds as well for the one we will invent. There is no way to diminish
the number of mediations. The only possibility is to align some of
them in a more or less stable and principled manner. This resurrection
of cognitive dimensions, interests and goals in Carlson and Gorman's
comments is an indication that they are still trying to recapture
Edison and Bell in their 'real essence'. Our goal is much more modest:
we simply try to triangulate accounts about Edison and Bell, and to
keep track of their diversity.
Whatever the disagreements I have with the commentators, I agree
with them that mappings like the one we offered and those they have
developed are ideally suited for a collaborative effort. The final result
will certainly be immensely different from the somewhat clumsy

This content downloaded from 185.44.77.89 on Wed, 25 Jun 2014 02:55:29 AM

All use subject to JSTOR Terms and Conditions
Responses & Replies: Giere: Response to Pickering 95

attempts described in our Note. Then the chequered evolution of the

method will itself be mapped through an STG - or its descendants-
offering the reflexivists another occasion to exercise their talents.

* NOTES

1. Bruno Latour, Philippe Mauguin and Genevieve Teil, 'A Note on Socio-
Technical Graphs', Social Studies of Science, Vol. 22, No. 1 (February 1992), 33-57.
2. Culminating in W. Bernard Carlson and Michael E. Gorman, 'Socio-Technical
Graphs and Cognitive Maps: A Response to Latour, Mauguin and Teil', Social Studies
of Science, Vol. 22, No. 1 (February 1992), 81-91. (Page numbers of quotations from
this Response are given in the text.)
3. See M. Callon, J. Law and A. Rip (eds), Mapping the Dynamics of Science and
Technology (London: Macmillan, 1986).

Author's address: Centre de Sociologie de

l'Innovation, 62, Boulevard Saint- Michel,
75006 Paris, France; e-mail LATOUR at
FREMPl 1 .bitnet.

Responses and Replies (continued)

The Cognitive Construction of Scientific

Knowledge (Response to Pickering)

Ronald N. Giere

Andy Pickering (AP) has posed some seriousquestions'for the account

of science I presented in Explaining Science (ES).2 He deserves an
equally serious reply.

Social Studies of Science (SAGE, London, Newbury Parkand New Delhi), Vol. 22
(1992), 95-107

This content downloaded from 185.44.77.89 on Wed, 25 Jun 2014 02:55:29 AM

All use subject to JSTOR Terms and Conditions