Brief encounters: Calculation and the interaction order of anonymous electronic markets

Description
Calculation has been recently discussed in relationship to market transactions as: (1) a set of operations, including classifications
and computations, which support decision-making processes by economic actors; (2) action plans or strategies
which can be evaluated against efficacy criteria; (3) broader social processes which induce behavioral modifications and
transformations along (1) and (2). Calculations would appear as situated plans or strategies, bounded by institutional constraints,
and anchored in classifications, computations, and evaluations, strategies which are implemented within trading
interactions. Such plans make use of available resources and adapt to constraints, but are prior with respect to live trading
interactions. Using a conceptual apparatus anchored in the work of Erving Goffman, I argue that calculation is situational
action. Its features are shaped by the interaction order of trading, and it can be conceptualized as emerging from gaming
encounters—i.e., competitive displays of the participants’ socially relevant attributes. These arguments are supported with
empirical data from online, anonymous financial trading. In these markets, gaming encounters make anonymous strangers
present in the trader’s situation, as a basis for assessing the relevance of displays on the trading screen and for reacting to
these displays. At the same time, traders engage in repeated self-displays as a means for defining their own situation and for
projecting subsequent action sequences.

Brief encounters: Calculation and the interaction order
of anonymous electronic markets
Alex Preda
*
University of Edinburgh, Sociology, SSPS, Chrystal Macmillan Building, 15A George Square, EH8 9LD Edinburgh, United Kingdom
Abstract
Calculation has been recently discussed in relationship to market transactions as: (1) a set of operations, including clas-
si?cations and computations, which support decision-making processes by economic actors; (2) action plans or strategies
which can be evaluated against e?cacy criteria; (3) broader social processes which induce behavioral modi?cations and
transformations along (1) and (2). Calculations would appear as situated plans or strategies, bounded by institutional con-
straints, and anchored in classi?cations, computations, and evaluations, strategies which are implemented within trading
interactions. Such plans make use of available resources and adapt to constraints, but are prior with respect to live trading
interactions. Using a conceptual apparatus anchored in the work of Erving Go?man, I argue that calculation is situational
action. Its features are shaped by the interaction order of trading, and it can be conceptualized as emerging from gaming
encounters—i.e., competitive displays of the participants’ socially relevant attributes. These arguments are supported with
empirical data from online, anonymous ?nancial trading. In these markets, gaming encounters make anonymous strangers
present in the trader’s situation, as a basis for assessing the relevance of displays on the trading screen and for reacting to
these displays. At the same time, traders engage in repeated self-displays as a means for de?ning their own situation and for
projecting subsequent action sequences.
Ó 2008 Elsevier Ltd. All rights reserved.
Introduction
Calculation has emerged in the recent debates
about (?nancial) markets as re-focusing the ana-
lytical attention away from social–structural
aspects of market exchanges (such as ties within
networks and groups) to the processes through
which trading strategies are worked out in trad-
ing rooms (e.g., Beunza & Stark, 2005, p. 92).
Calculation has also been tied to the concept of
agency, designating the incorporation of scripts
for practical actions in (formal) economic repre-
sentations and technologies (e.g., Callon, 2004,
p. 123; Callon, 2007, pp. 337–338). Calculative
agency highlights the role of commonly held dis-
tinctions and classi?cations in making economic
entities suitable for formalization (and for the
operations implied by this latter, including com-
paring, ranking, and computing) (Callon &
Muniesa, 2005, p. 1231; Muniesa & Callon,
2007). Markets appear thus as ‘‘calculative collec-
tive devices,” emphasizing the common e?ort put
by groups of market actors into reaching various
0361-3682/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aos.2008.06.005
*
Tel.: +44 0 131 650 4052; fax: +44 0 131 650 3989.
E-mail address: [email protected]
Available online at www.sciencedirect.com
Accounting, Organizations and Society 34 (2009) 675–693
www.elsevier.com/locate/aos
degrees of consensus on value. At the same time,
calculative practices (di?erent from calculating
devices) emphasize the social processes through
which speci?c rationalization procedures becomes
institutionalized within organizational structures
(e.g., Miller, 2008, p. 53). These procedures, in
their turn, can change organizational dynamics,
the relationships between organizations and polit-
ical institutions, and can lead to the emergence of
expert bodies which set up implementation
standards.
At least three notions of calculation are implicit
in these arguments: the ?rst is that of cognitive
operations, including classi?cations and computa-
tions. Such operations have a sequential character,
are iterable, and their outcome can be estimated
and anticipated. The second notion, which includes,
but is not limited to the above, is that of selecting,
projecting and evaluating the outcomes of market
transactions. Calculations are seen as strategies
involving expectations about other actors’ beliefs,
the evaluation of alternative courses of action, as
well as criteria for selecting and implementing
courses of action. Trading strategies rely on classi-
?cations, the processing of numerical data, as well
as criteria and procedures for the optimization of
means with respect to transaction goals (Biggart
& Castanias, 2001, p. 473). For instance, the
manipulation of numbers contributes to the reliabil-
ity and accountability of strategies. Finally, calcula-
tion designates the social processes through which
entities are selected and transformed in such a
way that they become the object of market transac-
tions submitted to e?ciency criteria. These pro-
cesses create institutions—sets of rules
incorporated in artifacts and organizational struc-
tures—which provide the resources for, as well as
the constraints of (?nancial) transactions (Miller,
2001, p. 380).
From the perspective of ?nancial transactions as
live and lived interactions, then, calculations appear
as strategies (or action plans) implemented in the
interaction order of trading. Is the latter then some-
thing external with respect to such plans—a setting
in which plans are realized—or is it intrinsic to the
features of transactions? More generally: what is
the relationship between calculation and trading
interactions?
An appropriate instance for examining these
questions is that of anonymous, online ?nancial
markets. In some situations at least, the interaction
order seems to be reduced here to a bare minimum.
Lay online traders
1
, for instance, anonymously
trade ?nancial securities on electronic platforms
on their own account, earning all or a considerable
part of their income from online trading activities.
They are not part of any organization and are not
employed in any capacity by any ?nancial institu-
tion. Organizational habitats are part of the traders’
system, but not of their lifeworld, understood as the
traders’ primary reality (Habermas, 1987; Schutz &
Luckmann, 1974, p. 35). Moreover, anonymous
electronic transactions are click and trade, not talk
and trade. In a certain sense, online anonymous
trading seems to come close to the normative model
of isolated, calculating individuals making choices
based on their strategies. By studying the apparently
minimal interaction order of lay online trading, we
can investigate whether calculation consists of
implementing strategies or not.
In the following, I examine the above questions
based on ethnographic observation (including audio
and video recordings) and interviews with lay online
traders. In the ?rst step, I elaborate the theoretical
frame of the analysis to follow. I argue that calcula-
tion remains an interaction-based achievement and
can be best conceptualized as a relational and situa-
tional activity. It should not be understood as the
mechanistic application of a plan or formula. Build-
ing upon Erving Go?man’s concept of encounter, I
suggest that in anonymous online markets calcula-
tion emerges from gaming encounters geared
toward the display of the participants’ socially rele-
vant features. Secondly, I present the methods and
the data on which this paper is based. Afterwards,
I shortly present the technological and institutional
developments which support lay online trading, as
well as some basic elements of their trading activi-
ties. Fourthly, I show how lay online traders calcu-
late trades based on at least two interaction
achievements: projection of the self and actualiza-
tion of other presences in the given situation. In
anonymous, online ?nancial markets calculation
does not appear as the application of a plan or for-
mula, or a strategy whose outcomes are checked
against projected results. Calculation emerges from
gaming encounters with anonymous strangers. It
1
Lay traders are usually referred to in the popular media as
day traders. This term is perceived as pejorative; moreover, day
trading designates a set of speci?c techniques not always used by
lay traders. Therefore, I will refer throughout this paper to lay, in
the sense of non-professional, albeit full-time or near full-time
traders.
676 A. Preda / Accounting, Organizations and Society 34 (2009) 675–693
requires bodily work di?erent from that of face-to-
face encounters. Its main prerequisites are the sta-
bilization of the screen and the de?nition of speci?c
situations. The screen appears as an observational
instrument, as well as a portal through which anon-
ymous traders ‘‘cross over” into each other’s situa-
tion. Since the display of social attributes is
achieved, among others, through price variations,
this contributes to explaining the price volatility
observed in electronic markets.
Calculation, plans, and the interaction order
Calculation as planned action implies that actors
implement rule-based sequences of activity which
can be evaluated against criteria of e?ciency. With
respect to ?nancial trading, for instance, such
sequences include establishing the objective, consid-
ering alternative courses of action, computing the
likely outcomes and enacting the outcome ‘‘that
optimizes risk and return according to pre-deter-
mined decision criteria” (Fenton-O’Creevy, Nichol-
son, Soane, & Willman, 2005, p. 79). This division
between devising a trading plan and implementing
it seems to resonate with an organizational division
of labour between fund managers, for instance, who
draft plans, and traders implementing such plans
under speci?c constraints. Planned actions can
require the use of various technologies of evaluation
and execution (e.g., charts, software programs), the
application of formulas, informational inputs and
data processing activities (e.g., analyses, access to
price data), as well as webs of relationships,
exchanges, and collaborative processes (as provided,
for instance, by the division of labour found on
trading ?oors). Organizational contexts can provide
the resources for planned actions, but they also set
constraints upon them: for instance, organization-
based networks of relationships can function as
the pipes (Podolny, 2001) through which informa-
tion ?ows, while also setting boundaries to the cir-
culation of information.
While institutional constraints and resources
(Abola?a, 1996), adhesiontoanalytical tools or styles
(Smith, 1999), judgment biases (Fenton-O’Creevy,
Nicholson, Soane, & Willman, 2005, pp. 83–86), as
well as intuitive elements (Zaloom, 2006, p. 136)
can a?ect the concrete courses of trading, the planned
element is nevertheless crucial. This element concerns
not only the sequentiality of trading activities (from
planning to implementation andevaluation), but also
the fact that, at any given time in the process of
trading, rules and criteria can be separated from the
lived activities which embody them. Traders’ post
hoc accounts and rationalizations could be seen as
empirical evidence for this separability, assuming
that such accounts would be identical with what trad-
ers actually do. The analytically-minded observer
could then detach and analyse calculative elements
from actual actions, and evaluate the latter against
the former. Moreover, calculative elements would
have to be prior to the concrete, live trading: while
this latter may introduce modi?cations and adapta-
tions, the plan would need to be recognizable as such
(Suchman, 1987, p. 36) and implemented by the
trader.
From the perspective of the interaction order,
then, trading would be situated calculation. Situa-
tional resources and constraints may make traders
adapt or revise calculations; they may impede upon
traders’ choices, as well as upon their evaluations.
Nevertheless, such resources and constraints would
appear as external with respect to trading plans or
strategies. An observer should then be able to iden-
tify and describe at least three distinct phases of cal-
culation: plan preparation, plan implementation/
adaptation, and evaluation. Whether such a distinc-
tion can analytically and empirically hold depends,
among others, on a closer examination of the inter-
action order of trading.
Calculation as a situational and relational activity
Financial trading is a relational activity, implying
an orientation toward, as well as exchanges with
other market actors, and relying on observational
technologies: computer screens, display boards,
and the like. The possibilities for (direct or medi-
ated) action imply zones of operation in which
actors e?ect changes (Schutz & Luckmann, 1974,
pp. 36, 44), the boundaries of which are shaped by
technology. The shared expectations which underlie
transactions draw on local resources, physical as
well as human, verbal as well as non-verbal. From
the perspective of the individual actor, the assump-
tion that participants will draw on a pool of shared
(and largely not explicit) expectations is itself crucial
with respect to the possibility of action. Accord-
ingly, trading as a relational activity is not only sit-
uated, but also situational (Go?man, 1983, p. 9):
that is, the resources of the situation are not simple
vehicles for a strategic activity or a plan whose def-
initional features remain independent of these
resources. It is not the same whether traders receive
A. Preda / Accounting, Organizations and Society 34 (2009) 675–693 677
price data on their screens, through shouting
accompanied by hand signs, via a formal letter at
home, or on a paper slip left under the doormat.
In anonymous, online ‘‘click and trade,” direct
verbal or bodily interactions with other traders are
not part of situational resources. The screen dis-
plays alphanumerical data, of which price and vol-
ume are seen as the most important. As an
observation and manipulation zone, the screen con-
stitutes the trader’s ‘‘kernel of reality” (Schutz &
Luckmann, 1974, p. 42)
2
. While in some cases the
screen enables conversations-qua-transactions
(Knorr Cetina & Bruegger, 2002, p. 913), in anony-
mous trading the screen ful?lls a dual role: as a
board on which anonymously posted displays can
be observed, and for posting the trader’s own data.
Observation of price data is not identical with
gazing at the screen. It implies awareness, or
selected availability of what is being seen (Heath,
Sanchez Svensson, Hindmarsh, Lu?, & vom Lehn,
2002, p. 321). Observing a particular price data or
accepting a transaction (i.e., clicking a button on
the screen for a speci?c display) can be seen as an
acknowledgment and validation of the presence of
another trader which, for a moment, becomes rele-
vant. Other traders, however, are anonymous
strangers, whose position is characterized both by
remoteness and closeness (Simmel, 1971 [1908], p.
145). They are remote not only in the spatial sense
of the word, but also because of a lack of any infor-
mation about their identity, interests, or intentions,
as well as of the absence of any personal relation-
ships, of the kind institutional traders can develop
(Knorr Cetina & Bruegger, 2002, p. 941). They are
close in the sense that traders must work with the
assumption of basic similarities (of knowledge and
interests) between them and their unseen counter-
parts. But they are close also because traders, in
order to transact, must face unseen strangers as
being there. The data ?ickering on screens are taken
as appresentations (Husserl, 1995, p. 112)—repre-
sentation and perception fused together—of the
traders displaying them. Thus, anonymous strang-
ers have to be co-present in the trader’s situation.
The acknowledgment and validation of anony-
mous strangers, however, cannot be based on an
immediate and direct orientation to their presence.
First, strangers are remote and do not have a face.
Second, they display in a ?eeting manner. Third,
they compete for the trader’s attention. The very
limited range of resources displayed on the screen
(price and volume data)
3
requires a trader to draw
upon additional means in the process of acknowl-
edging and validating other presences.
Such means cannot be seen as a set of universal,
?xed rules, of the kind implied by the equivalence
between calculation and strategy. First, ?xed rules
would not do justice to permanently changing dis-
plays. Second, they would have to come from some-
where; neither the screen nor the (organizational)
context provides them as a resource.
4
Third, they
would have to stabilize the presence of other traders
as signi?cant, even if for a moment. A plan or a for-
mula (as part of a plan) can hardly provide for this,
since it cannot include ex ante criteria for evaluating
signi?cance.
5
In order to do this, a plan would have
not only to foresee the order in which data will
?icker on the screen, but also their value. In a sim-
ilar manner, if we understand calculation as the
application of a formula (which generates data),
then that formula would have to contain the criteria
according to which the signi?cance of the generated
data could be evaluated. For instance, if we see cal-
culation as the application of a formula for options
prices, then this formula would have to establish the
signi?cance of theoretical prices in the same manner
for every trader. A plan would have to provide
omniscient anticipations of other traders’ actions,
anticipations which would be then embedded and
adapted within the interaction order. If a plan or
calculation cannot be taken as distinct from the
interaction order in which it is enacted, then it
2
Relying on Mead, Schutz and Luckmann distinguish between
the manipulative zone and the zone of distant things. In the
manipulative zone, objects can be seen and touched, whereas in
the zone of distant things objects can be seen, but not experienced
via live corporeal contact.
3
Traders have the possibility of inferring categorical identities
from this data (e.g., whether displays are made by individual or
by institutional traders); they also have the possibility of asking
the brokerage house to reveal post hoc the identity of their
counterparts for speci?c trades. This identi?cation is a longer
bureaucratic process which can unfold only after the trading
moment, when it had lost relevance. The observed traders do not
use it, preferring, for all practical purposes, anonymity.
4
Science and technology studies point to the ways in which
plans, such as engineering blueprints, schemes, drawings, etc. can
act as props for action. These are not simply representations or
sets of instructions for actors, but rather technologies of social
interaction (for a recent example, see Vertesi, 2008).
5
This resonates with Ludwig Wittgenstein’s (Wittgenstein,
1984, p. 416) critique of calculation as the application of a
formula.
678 A. Preda / Accounting, Organizations and Society 34 (2009) 675–693
would have to be intrinsic to the trader’s
orientation.
This orientation, however, is not an a priori with
respect to action, but established within the interac-
tion order.
6
If such rules or plans exist at all, they
would have to emerge in this order—which contra-
dicts the presumption of generality and universality.
Within this context, calculation can hardly be seen
as following a given set of instructions: such a set
is not part of the resources at hand. Even assuming
its location in the actors’ consciousness, it could not
support action without a set of instructions for cor-
relating rules to the resources of the situation. From
here on, however, in?nite regression sets in, since
correlation would require in its turn a set of specify-
ing rules, etc.
Calculation and trading encounters
I will contrast now plans with the notion of
encounter, as developed in the work of Erving Go?-
man. While plans and strategies are intrinsic to
games (which, indeed, have often been equated with
calculations), during the play of a game a variety of
relevant interactions occur, which cannot be de?ned
in terms of the game’s rules (Go?man, 1972, p. 33).
This variety can be seen as a gaming encounter: a
play of a game of chess, for instance, is a special
abstraction from the gaming encounter between
speci?c players. As usually presented in chess man-
uals, these plays abstract, select, and re-work con-
crete interactions into schematic visual
representations of moves. It is this re-working which
allows the reference to plans and strategies.
Glances, bodily movements, worded exchanges, or
tantrums are left out, enabling the codi?cation of
chess encounters as plays of the game of chess, with
the latter being analyzable in terms of moves, coun-
ter-moves, and strategies.
If, however, we regard the gaming encounter as
relevant for what is going on in the play of a game
(what Go?man calls an occasion for gaming), ele-
ments such as protests, interruptions, gestures, or
glances form an organic system of interactions
highly relevant for the outcome of the encounter
(and hence of the play). A participant’s actions,
then, cannot be seen as the implementation of a
strategy, but as a situational reaction to the previ-
ous action turn. A gaming encounter can be charac-
terized by a problematic outcome and by sanctioned
displays of socially relevant attributes (Go?man,
1972, p. 61), such as dexterity, endurance, self-con-
trol, resilience to humiliation, and the like. Games,
then, can be seen as arrangements or conventions
for ‘‘integrating into gaming encounters . . . socially
signi?cant externally based matters” (Go?man,
1972, p. 64), centered around speci?c sets of rou-
tines. Gaming encounters of chess, for instance,
include sets of routines executed during competitive
displays of social attributes. In such displays, partic-
ipants can switch across various keys (Go?man,
1974, p. 49) in which they perform their routines.
In a game of basketball, for instance, a player can
switch from a dribble to a mock pass (a make
believe) in order to confuse his opponent; such re-
keyings can be part and parcel of gaming
encounters.
The competitive display of social attributes res-
onates with Cli?ord Geertz’s notion of deep play
(1973, p. 433), in which cock?ght routines provide
the occasion for engaging in status competitions
by way of betting. In other words, gaming
encounters bring forth value-relevant issues, in
relationship to speci?c conventions and routine-
like procedures (such as those of football, or
chess). In a chess encounter, for instance, such
value-relevant issues can be ‘‘un?ippability,”
endurance, or quick response. Apparently unre-
lated gaming types, then, can provide occasions
for similar value-relevant issues (think of rugby,
car racing, and chess, for instance, with respect
to endurance, among others).
Face-to-face gaming encounters can be charac-
terized by symbolic distance from the environment
in which they take place (Go?man, 1972, p. 65)—
card games are a case in point here. In card games,
participants sitting around the table distance them-
selves from their audience, which may have the right
to look, but has to remain silent, is not allowed to
intervene, give advice, etc. The audience is then spa-
tially near the players and symbolically distanced
from them. In other kinds of encounters, not only
that the opponents/partners are not known, but
their relevance as opponents/partners is not known
(blind speed dating can be such an encounter). In
instances of online trading, the relevance of some-
body being a partner is not known previous to them
displaying on the screen. This prevents the gaming
encounter taking symbolic distance from the envi-
ronment. The traders have to use environmental
6
Similarly, in auctions prices emerge within the participants’
interactions (Heath & Lu?, 2007, p. 81).
A. Preda / Accounting, Organizations and Society 34 (2009) 675–693 679
resources—the screen, among others—in order to
establish this relevance. Symbolic closeness replaces
distance. The same can be illustrated with ?oor-
based trading v. online trading. In the former, the
status and identity of competitors is known before
they make the ?rst hand sign (Zaloom, 2006, pp.
98–103). The pit is symbolically distanced from
the clerical desks and phone booths situated only
a couple of steps away. In online trading, this rele-
vance must be inferred from the screen displays,
without the latter providing enough resources for
establishing this relevance. Hence, the encounter
must be kept close to its environment. Online trad-
ing can then be re-keyed in ways in which ?oor
based trading can not: it can be re-keyed as cafe´
encounters (such as observed by this author, among
others), or as household encounters.
The resources used by a trader in acknowledg-
ing and validating the signi?cant presence of anon-
ymous others cannot come but from the situation
of which the trader is part. A major resource is
the trader’s own presence. A basic form of address-
ing the presence of the other is when the actor
‘‘replies to himself as truly as the other person
replies to him” (Mead, 1964 [1934], p. 203). In car-
rying on a ‘‘conversation of gestures” (Mead, 1964
[1934], p. 205) with herself, as well as in talking to
herself, the actor can use her own presence as a
resource for establishing a meaningful orientation
to the presence of (remote) others. Self-talk does
not emerge as indulgence or taboo breaking (Go?-
man, 1978, p. 788), but as a way of embedding the
assumed presence of other actors into the trader’s
situation and context of action. Similarly, the pres-
ence of familiar persons (or of familiar strangers)
in the situation can be taken both as an occasion
and as a resource for brining remote, anonymous,
strange traders into the situation. By addressing
others present in the situation, or simply by taking
the presence of others as an occasion for address-
ing herself, the trader can momentarily stabilize
apparitions on screen, and validate them as
signi?cant.
Against this background, online trading encoun-
ters are anchored in the orientation toward other
potentially relevant presences, and toward oneself
(as a major situational resource). This action
extends into the future: its aim is to process given
elements in order to obtain a result which can be sig-
ni?cantly related to other results, obtained through
past operations. Displayed data must be processed
in such a way that the results of these activities
can be connected to each other. Processing dis-
played data, in its turn, depends on actively adding,
deleting, or combining existing data in new con?gu-
rations and evaluating their relevance. It also
depends on selecting and actualizing some displays
as relevant in the given situation, while others are
overlooked. This actualization process implies
bringing the displayed data into the trader’s situa-
tion, relating it to her own actions as relevant. It
is a process of actualizing the actions of absent,
anonymous traders as relevant for one’s own future
actions. Finally, these actualized presences have to
be endowed with minimal stability, in order for
one’s own actions to unfold. Even if only for a brief
moment, they must be maintained as signi?cant. In
order to achieve this, trader would use the most ele-
mentary resources at hand: their own bodies and
voices.
The trader’s display to others cannot exclude
from the start socially relevant attributes; numerical
screen displays have existential qualities (Vollmer,
2007, p. 593) and intervene therefore in establishing
social relationships. Endowing other, unknown
presences with signi?cance as the basis for actions
is related to the display of such attributes through
numerical data. Traders must show that they are
‘‘there”; they must show that they are attractive to
others. Endurance, even obstinacy, coolness, or
attractiveness have their place among these
attributes.
In online ?nancial markets, then, calculation
emerges from encounters rather than the applica-
tion of a plan. It emerges as the execution, adapta-
tion, ongoing combination, and modi?cation of
speci?c routines (e.g., buying a put, selling a call
on an index) according to the requirements of the
encounter. Encounters can be seen as social rela-
tionships of short duration, characterized by com-
petitive displays of socially relevant attributes, and
having uncertain outcomes. In online trading, where
participants engage with anonymous strangers
through the computer screen, social attributes can-
not be displayed directly to anonymous others.
They can, however, be displayed in strips of actions,
like posting numbers on screen, or reacting (or not)
to the numbers posted by strangers. Trading
encounters would have therefore at least two dimen-
sions: the actualization of other presences in the tra-
der’s situation, and the trader’s self-displays to
others on the trading screen. Trading encounters
are symbolically close to the environment in which
they unfold. They involve bodily engagement with
680 A. Preda / Accounting, Organizations and Society 34 (2009) 675–693
strangers, an engagement mediated by the screen.
7
Since the execution of trading routines requires
ongoing adaptation to the characteristic of speci?c
encounters (i.e., to competitive displays of attri-
butes), it follows that transaction sequences cannot
be foretold. These sequences will vary not only from
market participant to market participant; they will
also have an intrinsic variability, due to the neces-
sity of adapting them on an ongoing basis.
Methods
An appropriate way of spelling out the empirical
underpinnings of the above propositions would be
to look close up at how calculations unfold in the
live process of trading. The aim is not ?nding repre-
sentative frequencies of types of strategies. (What is
taken as a strategy often reveals itself as a post hoc
rationalization.) Best suited for such an examina-
tion is naturally occurring data (e.g., Scheglo?,
1996, p. 167)—that is, observational data obtained
from online anonymous traders in their natural hab-
itat. Instead of asking lay traders, for instance, if
and how they calculate while trading, visual and
audio recordings of the trading process itself,
together with protocols of participant observations
8
can be used in order to identify and analyze calcula-
tive processes. This can be complemented with inter-
view data providing more background information
about the traders’ past experiences, about the con-
straints under which they operate, and the like.
Gaming encounters do not necessarily include
verbalizations and explanations of actions per-
formed by traders. Nevertheless, the presence of
the ethnographer can be taken as an occasion for
thinking aloud, or for providing accounts—that is,
evaluative inquiries (Scott & Lyman, 1968, p. 46)
of what is going on, while it is going on. In such
cases, the ethnographer’s presence can act less as a
disturbing or distorting factor and more like an
occasion—and resource—used by traders in formu-
lating what they are doing. This means that traders
will seize upon the presence of the ethnographer as
an occasion for glossing upon what is going on,
and for making dialogues with unseen trading part-
ners audible, dialogues which otherwise may take
place in petto. Especially since such formulations
are uttered in the process of trading—where so much
is at stake—they should not be seen as performing
for the bene?t of the ethnographer, but as concreti-
zations of encounter moments (i.e., as making them
audible). Any performance for the bene?t of the eth-
nographer would mean here interrupting an ongoing
on-screen gaming encounter and engaging in a face-
to-face one. Such shifts would be disruptive; they
would be observable, and would require repair. It
is the interaction itself (i.e., the presence or absence
of such repairs) which indicates whether the pres-
ence of the ethnographer is an occasion for audibil-
ity or a disruption of on-screen encounters.
An additional, appropriate way for checking
upon such moments is to examine interaction strips
where the ethnographer is absent. By recording
trading days without the presence of the ethnogra-
pher, we can see whether verbalizations occur as
intrinsic to the trading process or not. I have done
recordings of full trading days (a total of 10) with-
out my presence and checked them against record-
ings of trading days when I was present. The
comparison shows that traders verbalized interac-
tions even when they were completely alone. The
presence (or intervention) of family members did
not appear as a disruption, but as an occasion for
commentary, evaluations, and for making audible
strips of inner dialogue with unseen transaction
partners. Comparing the presence/interference of
family members with the presence of the ethnogra-
pher, disruptions or modi?cations of the trading
process were not recognizable in either case.
The following analysis is primarily anchored in
data obtained through observation of lay US trad-
ers in the period October 2005–March 2007, and
consisting in video and audio recordings of trading,
together with protocols of participant observation.
The traders used US-based electronic platforms;
nevertheless, as they traveled, observations and
interviews with the same traders were made both
in the US and in Europe. The data discussed here
consist of the following: extensive recordings of
trading days (over 82 h of recordings); video record-
ing of trading with observation protocols (over 3 h);
observation protocols without recording (over 9 h);
observation protocols with audio recording (over
7 h); interviews (over 40 h).
7
Conversely, in boxing, which is apparently exclusively body-
centric, talk plays a signi?cant role (Wacquant, 2004, p. 66).
8
While protocols of participant observation cannot be consid-
ered naturally occurring data in the strict sense of the word, they
can capture situation-relevant elements which cannot be retained
by an audio recording, such as gestures. Video recordings
certainly allow the analysis of such elements, but they o?er
mostly a restricted angle, not identical with the vision ?eld of the
observer. In many cases, a combination of video recordings and
participant observation has the potential for better data yields.
A. Preda / Accounting, Organizations and Society 34 (2009) 675–693 681
The use of this combination is motivated by the
necessity to: (a) check interaction consistencies; (b)
complement audio data with visual ones, thus get-
ting access to behavioral detail (Pera¨kyla¨, 2006, p.
95); complement audio and video data with inter-
views, gaining thus insight into the traders’ operat-
ing context, background, and self-perception. In
addition to participant observation, recordings,
and interviews, this author has attended training
and presentation sessions organized by the elec-
tronic brokerage ?rm used by the traders. The ses-
sions were aimed at introducing lay traders to the
trading software and brokerage services of the ?rm;
they were organized in cooperation with and on the
premises of a global stock exchange.
Before a step-by-step discussion of various
aspects of calculation within the interaction order
of anonymous online trading, a presentation of
the modus operandi, the constraints, and the sys-
temic possibilities available to lay traders is neces-
sary. Understanding the modus operandi—that is,
the way in which traders assemble, modify, and pre-
serve the trading screen—as well as the ?nancial and
material constraints under which they operate is
essential not only for grasping their various activi-
ties, but also for how calculation is grounded in
the interaction order of trading.
The framework of lay online trading
Non-institutional (or lay) traders are individuals
who do not belong to a ?nancial organization and
trade full- or part-time on their own account. Obser-
vations suggest that an o?ine non-institutional tra-
der will trade at least 4–5 times per week. The
Securities and Exchange Commission (SEC) has
now a formal, frequency-based de?nition of lay
online trading, but there is no such de?nition for o?-
line trading.
Traders use the services of brokerage houses
(electronic or not) in order to transact stocks, com-
modities, indices, currencies, or derivative products
thereof, among others. While the historical dynam-
ics of lay trading still awaits sociological attention,
in the late 1990s some traders took advantage of
the widespread availability of electronic communi-
cation technologies (e.g., the Internet, trading soft-
ware, electronic trading platforms) and switched
to transacting ?nancial securities online (see also
Schatzki, 2002, pp. 157–161). Many lay traders use
the Nasdaq (National Association of Securities
Dealers Automated Quotation) system, which is
over-the-counter (i.e., not tied to a centralized auc-
tion system) and was already entirely automated
in the late 1990s.
The Securities and Exchange Commission (SEC)
estimated the number of full time lay traders in the
US (called day traders)
9
at about 7000 and their
share at 15% of the total Nasdaq trading volume,
while the number of brokerages was 133 (SEC,
2000, sections III.A, III.B). In 1999, the then SEC
chairman Arthur Levitt had de?ned the day trader
as ‘‘an individual, not registered as a broker-dealer
or as a registered representative, who trades stock
at a ?rm that allow the individual real time access
to the major stock exchanges and the Nasdaq mar-
ket” (SEC, 2000, section III.C). Characteristic for
day traders is their non-institutional, organization-
independent status (the legally relevant distinction
from professional traders also makes di?cult to
estimate their total number).
Using data compiled from brokerage houses, the
SEC estimated that most day traders (57%) were
between 30 and 45 years; more than half (53%)
earned over $100,000/year and 78% had a net worth
of $200,000 or greater (SEC, 2000, section IV.C).
Discussing more stringent margin requirements,
the SEC de?ned the ‘‘pattern day trader” as ‘‘any
customer that executes four or more day trades
10
within ?ve business days, provided the number of
trades is more than six percent in the account for
the ?ve day period,” and proposed a new rule to
the e?ect that pattern day traders will be required
to maintain a minimum equity of $25,000 at all
times (SEC, 2000, section IV.H.1). Judging after
the continued revenue growth of electronic broker-
ages in the early 2000s (accompanied by continu-
ously falling fees), the cautionary tone of the SEC
does not seem to have deterred lay traders.
We can distinguish at least among three mean-
ings of ‘‘day trader”: (1) a non-institutional trader
working on her own account and using her own
9
The term day trader is present in the academic literature since
at least the early 1980s (e.g., Van Landingham, 1980), but with a
di?erent connotation, determined by the opposition with buy-to-
hold investors. The distinction lay/professional is missing here.
10
‘‘Day trades” implies here that the traders buys and sells the
same security within the trading day. Observations suggest,
however, that this is not always the case; lay trading is not
restricted to, and does not necessarily imply buying and selling
the same security within one day. This technique is sometimes
used by lay traders and designated as such—i.e., day trading. In
interviews with lay traders, these considered day trading as risky
and (even when practicing it) not to be recommended.
682 A. Preda / Accounting, Organizations and Society 34 (2009) 675–693
?nancial means; (2) the repeated buying and selling
of a security within one trading day; (3) a regulatory
classi?cation of the SEC, imposing a set of speci?c
restrictions with regard to capital and margins on
(1).
Lay online trading has been rather favored by
several recent technological developments. Among
the most important are: (1) the adoption of a new
encryption standard (the AES, or Advanced
Encryption Standard) in 2001 (e.g., Landau,
2000)
11
; (2) the spread of ADSL, modem and
high-capacity cable networks, which has facilitated
real time (or near real time) transmission of ?nan-
cial data outside networks of professional institu-
tions, contributing to lower trading fees (from
$15–25 per trade in 2000 to $1 and less today); (3)
the expansion of electronic communication net-
works (ECN)—i.e., of computer-mediated ?nancial
transaction networks; (4) the expansion of US-
based electronic brokerage houses.
The screen as an observational and as an encounter
device
In the cases discussed here, lay traders logged in
on a NASDAQ level II screen, which looks like a
grid of di?erently colored cells, revealing the bids
and o?ers for every anonymous market participant
posting quotes. The ?rst three columns contain the
security’s trading code, the platform/exchange on
which it is tradable, and the expiry date (for deriva-
tives). The next columns provide information (in
this order) about position, average cost, quantity
unrealized, bid size, bid price, ask size, ask price,
last size, change, last price. The amount of data
which is visible at once depends, however, on screen
size, and this is an important limitation.
12
In addi-
tion to information about bids and asks, traders
can choose to display information about their asset
balance and liquidity, which is automatically
updated. Due to the physical limitations of the
screen, multiple windows can display various kinds
of information: price charts, news, expert recom-
mendations from specialized services, or a chat
room. This latter capability was not used by the
traders observed by the author.
The possibilities o?ered by the trading screen
indicate that not all activities of traders are calcula-
tion: searching for news, perusing of expert recom-
mendations, accounting of past relevant events or
interpretation of charts are some of the activities
which take place during a trading day. Yet, among
the most important moments are those of engage-
ment with other traders. They are important not
only because there is considerable more excitement,
and sometimes even tension, in such encounters, but
also because the trader’s ?nancial situation changes
as a direct consequence of them.
On a trading screen interface, 325 active cells (25
rows and 13 columns, arranged in a grid) are visible
at once. The interface can contain many more cells,
but they will not be visible without scrolling down.
It also has several (usually ?ve) rows of buttons (sit-
uated at the top of the screen). When a button is hit,
a new window pops up, showing new sets of data, or
triggering a di?erent activity: for instance, the trader
has ‘‘news,” ‘‘ticker,” and ‘‘chat” buttons at the top
of the screen. The active cells contain numbers and
letters, but some of them can also be used as but-
tons: for instance, an order is transmitted by trans-
forming a data cell into a button and hitting it. In
this case, the letter ‘‘T” will appear in the cell on a
colored background. Additionally, a sent order
can be cancelled by transforming an adjacent data
cell into a button and hitting it. (The letter ‘‘C” will
appear in this case in the cell against a colored back-
ground, di?erent from the ‘‘T” one.) The usual
background color of the trading interface is white
(on the left side) and black (on the right side). Col-
ors of letters, numbers, and cell buttons can be cho-
sen by the trader, who can also program di?erent
acoustic signals to be triggered by speci?c events
(e.g., when a sale occurs).
The trading interface contains thus three areas:
the button area at the top, shaded in grey; the white
area on the left hand; the black area on the right
hand. The grey and white areas occupy each about
one-?fth of the computer screen. The black area
takes the other three ?fths. This is also the zone
which necessitates the most attention and e?ort
from the trader. The grey area contains buttons cor-
responding to various activities (read news, or get
more data, or get other kinds of data) which, in
their turn, are contained on di?erent screen inter-
11
While investment banks and exchanges operate secure intra-
nets, lay traders depend on secure broadband connections and on
commercial encryption software.
12
Some of the lay traders observed were conscious about this
physical limitation and were considering working with a second
screen, in a manner similar with professional traders. Other
traders worked indeed with two screens. The fact that notebook
computers, with which many lay traders work, have smaller
screens, increases these physical limits to the amount of available
information.
A. Preda / Accounting, Organizations and Society 34 (2009) 675–693 683
faces. The cells in the white zone contain acronyms
of ?nancial products, the name of the trading plat-
form, and expiry dates of derivative products.
Observation and trading do not necessarily overlap:
for instance, if traders transact derivatives, they will
also put the underlying asset in one of the top rows,
in order to monitor it. The cells in the black area
will contain speci?c data for each of the names in
the white area: price (average, limit, bid, ask, last),
quantity in hundreds (unrealized, limit, bid, ask,
last), pro?ts and losses, status, and destination,
among others. The cells can be transformed into
active buttons: when traders want to send orders,
the ask price cell will become a transmit, or T-cell,
while the ask size cell will become a cancel, or C-cell.
The data contained in columns is not necessarily of
the same kind: for instance, the ?rst ?ve or six cells
in a column will contain data about the bid size, the
next ?ve or six data about limit prices, after which
the cells will revert to bid size.
During market hours, the cells in the black zone
of the screen can ?icker at once, at a perceived fre-
quency of up to 4–5 times per second: this is because
the data display changes in color, length, and con-
tent. Not only that the same type of data varies
within a cell, but a cell might display a certain kind
of data, of a certain length, and in a certain color
(e.g., bid size), and, in a fraction of a second, shift
to displaying a di?erent kind of data, of a di?erent
length, and in a di?erent color (e.g., limit price).
Thus, for the same cell or location, the trader is con-
fronted with sensory variations (color and length of
data sequences) which have to be related to intra-
data changes and/or to shifts across data types,
according to situation. Additionally, these changes
occur simultaneously in di?erent cells in the black
zone of the screen. Monitoring them is crucial and
takes the most time of online traders. Not only that
the trader’s ?nancial situation and status depend on
the accurate monitoring of the screen (and espe-
cially of the black zone); any transaction traders
may want to conduct is built upon and depends lar-
gely on uninterrupted monitoring. In fact, about
90% of the trader’s active time, if not more, is spent
on such monitoring.
From the viewpoint of the trader, each ?icker is a
display of action competing for attention with other
displays. It indicates that somebody has posted
something. The general notion of ‘‘stu? is happen-
ing” is still far from the awareness of a presence
which can be acknowledged and validated by the
trader as signi?cant. A ?icker can become a pres-
ence only when brought into relationship to the tra-
der’s positions and orientation. In order for this
transformation to take place, the remote stranger
has to be brought near. While the trading screen
provides some elements to trigger and support this
transformation, they are not enough. Moreover,
the screen is a fundamentally unstable object, ?ick-
ering all the time, without any apparent rhythm or
overall, precise, underlying rationale. This instabil-
ity is ampli?ed by traders who, in the search for
meaningful presences (but also in response to con-
straints) continuously modify the composition of
the screen.
The screen appears thus as a laminated object
(Go?man, 1974, pp. 82, 156–157), framing together
various layers of activity, some of which are ori-
ented toward anonymous strangers, and some
toward searching for information which should
enable this orientation. Some of these layers enable
contemplation, some enable experimentation, and
some enable engagement. Some zones of the screen
contain durable data (e.g, the codenames of securi-
ties), while others contain highly variable data (e.g.,
prices and volumes). Traders can set up di?erent
combinations of securities to be observed, can
increase the number of combinations, and can seek
out encounters with various combinations of trades.
Thus, at any given time, each trading screen is
unique, in the sense that it incorporates unique com-
binations of anonymous presences and unique
responses to them.
Calculation as bodily work and material coordination
Orientation toward, and interaction with the
screen is crucial for engaging in encounters. The
traders’ attitude is far from a passive, contemplative
one. Observing the screen is more than ‘‘looking
at”; it implies physical closeness and active bodily
engagement:
1 Trader: [16
00
] [points ?nger to screen cell with
underlying asset on which he trades puts]
2 still there quite a bit though I mean [moves cur-
sor up] .hh but I have stu? next month so.
3 hh [13
00
] [brings cursor back down] [moves cursor
up opens portfolio window] not a
4 pretty month for me .hh I’ll say that much [4
00
]
that might not be as bad as it looks, I
5 mean [points with ?nger to index cell] if that
?nishes even above seventyseven [2
00
] it
684 A. Preda / Accounting, Organizations and Society 34 (2009) 675–693
6 will wipe out all these [points with ?nger down to
row of puts he has put up for sale] you
7 know I mean I won’t have to buy’em back"
[opens again portfolio window] but that’s the
8 problem with getting the squeeze, you know what
I mean
In the above sequence, the trader must decide
whether to buy back a bunch of puts he has sold
earlier. He must decide upon (or calculate) the
implications of his actions, with respect to present,
as well as to future situations. Calculation appears
neither as a set of purely mental operations, nor
as the application of a plan. It is not recognizable
as the implementation of a (standard) move, nor
taken as such by the trader. In order to establish
what they are going to do next, traders have to
de?ne their situation at a given moment. De?ning
the situation, however, is not simply a matter of
words: good, bad, acceptable, etc. Nor is it a matter
of comparing data with a template, or of evaluating
them according to a set of criteria. It is a matter of
identifying relevant spots on the screen and correlat-
ing across spots, so that actions can be adapted to
an ever changing display with multiple, coordinated
sites, where relevant encounters might happen.
Manipulating numbers is contingent on downkey-
ing them (Vollmer, 2007, p. 587)—that is, on identi-
fying them as relevant spots on the screen.
Adjectives alone cannot do justice to such a situa-
tion; traders have to use their body in order to coor-
dinate with the experimental device. Hitches,
response cries (Go?man, 1978), and hand move-
ments are used to mark places and moments of rel-
evance for future actions and to support
calculations. The entire body, although strapped
to the chair, is geared toward such operations: lean-
ing forward, pointing at the screen with the ?nger,
clicking the mouse, sighing are as many bodily
actions marking relevant moments in calculative
activities.
This marking is important in an ever changing
environment, and allows the trader to stabilize, even
if only for a moment, the spots on the screen which
need to be correlated and accounted for. The very
activity of accounting for a screen ?ickering—
intrinsic to encounters—cannot dispense with bod-
ily work. In the lines 5–6 above, the trader does
exactly this—he projects the relevance of a spot on
the screen for his future situation by engaging his
body in interaction with the screen (through coordi-
nation of hand and eye movements with vocaliza-
tions) and by glossing upon his own movements in
connection to screen spots. In this perspective,
encounters are prepared and sustained by material
and discursive coordination with the device, based
on stabilizing observations, on de?ning situations,
and on projecting future consequences of present
situations.
On the trading ?oor, encounters imply a heavy
amount of body work, including elbowing and
shoving aside, when necessary (Zaloom, 2006, p.
136). In the trading rooms of investment banks,
trading also involves coordination with other trad-
ers and with the screen. Even in situations like the
above, stripped of any immediate organizational
environment and of physical co-presence of other
traders, bodily engagement is still required by
encounters.
Calculation as anonymous encounters
Within the experimental ?eld incorporated in the
trading screen, participants display their trades,
waiting for an anonymous stranger to appear and
engage with them. Before this happens, traders can-
not know whether their posted trades are relevant to
others or not. Knowing this, however, is essential
for deciding upon future trades: which products,
at what prices, and for what expiry dates (in the case
of derivatives) should traders post? The answer
depends on the success of the trades already posted,
as well as on the traders’ reactions to anonymous
others. Engaging with others is essential for evaluat-
ing the relevance of numbers seen on the screen, as
well as for the question ‘‘what to do next”? Engage-
ments imply making others present in the situation,
in spite of their remoteness and anonymity.
One tool for bringing strangers in is talk: by talk-
ing to absent strangers, traders create the conditions
for calculating their trades neither as abstract and
impersonal computations, nor as the application
of a pre-established plan, but as responses to rele-
vant presences. The encounter, however brief, stabi-
lizes the numbers ?ickering on the screen and
enables their manipulation and rationalization as
responses within an interaction frame. Transactions
come out of such brief, anonymous encounters,
which can be framed by soliloquies. Like somebody
waiting in the street for a blind date which may be
late or could fail to appear, and who eyes unknown
passersby asking in petto: ‘‘is this the one?”, ‘‘This
one looks like s/he might be, etc.”, traders project
relationships by means of internal conversations
A. Preda / Accounting, Organizations and Society 34 (2009) 675–693 685
(Mead, 1934) and act depending on such
projections.
Making strangers present in the situation can
require addressing them, talking to them, as indica-
tive of the acceptance or refusal to engage. Such
talk, which cannot of course be heard through the
screen, often includes response cries, marking rele-
vant moments in the encounter (Go?man, 1978, p.
814; see also Hutchins, 1995, p. 313):
Trader: [15
00
] hah. [10
00
] [music playing] Raking in
the cash!" haha" I’m taking cash for risk. Hrrh
hah. Dim digidim digidim dim dim [slaps hand]
[32
00
] Ohkay.; Ohkay.; We’re going to learn
how it works out. Haha haha hh. [45
00
] [music
playing, tapping, computer sounds] Ohkay."
I’m not having that kind of day, bud. hihi Give
me a call. [pushes chair]
The trader is here completely alone. He has just
sold a batch of derivatives when an anonymous
presence on screen attracts his attention: he refuses
to engage in the encounter (that is, to take the trade
posted by the stranger) by addressing this directly,
as the ‘‘bud” who should maybe call later. It is
not recognizable here that the refusal is part of an
overall strategy of rejecting particular kinds of
trades, or that it follows any criteria for evaluating
the implications or the attractiveness of the posted
trade. The rejection comes because, after having
sold, the trader does not have ‘‘that kind of day.”
This is linked to what had happened immediately
before, not to any general decision-making frame.
It should be noted here that the outcome of the pre-
vious transaction is still open, but it is the previous
successful encounter which makes the trader declin-
ing a new one. Display of disinterest as a socially
relevant attribute, similar to somebody not return-
ing a glance in the street, or ignoring the remark
of a stranger, is intrinsic to this strip of trading
activity.
In the following sequence, toward the end of the
trading day, a trader engages in a conversation
while talking to his wife in front of the active trad-
ing screen:
Trader: We’re playing a little mad over trades.
[1
00
] It’s like, whatever, dude. I might, I might nib-
ble a little more, but [2
00
] I mean [.5
00
] obviously he
sold these at a much cheaper price;, but, you
know, hahaha I might just say, well, give me
those back;, I’ll sell you a few [1
00
] I didn’t really
want you to sell it to me at eighty?ve, I count, I
think seventy?ve is more appropriate, further
out, but you’ve got to pay me more" than you
would for the eighty?ves, because, you know,
it’s ten dollars cheaper that you can sell it to
me for. So you’ve got to give me more money
for it. That’s my attitude". hihihi you with me?
[7
00
] This sure is a crazy market time". I guess
today’s everybody’s, everybody decided today’s
not a good day to buy. Hehe
Immediately before this sequence, the wife has
entered the room saying that the dogs are barking
outside, which might mean that a code enforcement
o?cer (a county inspector) is in the neighborhood.
The trader uses the question to change the topic
to his situation, and, having elicited interest, utters
the above sequence. He had bought options on a
company stock at a speci?c strike price and wants
to sell options on the same stock at a di?erent strike
price. What is the appropriate price to ask for each
strike price? Instead of using a general formula for
calculating options prices, the traders relates to the
anonymous ‘‘dude” he has bought earlier and tries
thus to make sense of his own real intentions as a
basis for identifying the right price for each strike
price. At least in this instance, calculation does
not appear as the mechanic application of a formula
(which traders might even ignore), but as grounded
in establishing a social relationship with a
counterpart.
The ?rst step is to create the encounter by bring
the anonymous ‘‘dude” in, within a conversation
embedded in the conversation with the trader’s part-
ner. Then, the ‘‘right” asking prices for two di?erent
strike prices (75 and 85) are made dependent on a
change of mind (‘‘I didn’t really want”) and on fair-
ness imperatives (‘‘you’ve got to give me more
money for it”). Finally, this calculation is rational-
ized by the trader’s de?ning his attitude with respect
to outside, uncontrollable constraints (‘‘crazy mar-
ket time”) and with the unwillingness of passersby
to give him a glance (‘‘nobody’s willing to buy”).
All this happens while the algorithm for calculat-
ing options prices, embedded in the trading soft-
ware, is at hand; at the click of a screen button,
prices can be obtained for any strike price and any
underlying asset. Yet, there is a fundamental tension
between using a trading screen assembled as a
device for engaging in anonymous, brief encounters,
on the one hand, and using a formula, on the other
hand. Since the traders’ activity is geared towards
successful encounters, price calculation cannot but
686 A. Preda / Accounting, Organizations and Society 34 (2009) 675–693
unfold within this frame too, bracketing out algo-
rithms (see also Lave, 1988, p. 122). The right price
becomes then dependent on fairness issues, on the
traders’ changing their mind, on them not really
meaning it—in short, on well known interaction
and relational issues which can surface when
encountering strangers.
The price emerges in a direct negotiation with the
unseen trading partner, in a way similar to that in
which participants in a gaming encounter of cards,
sitting around the table, can ask each other to put
more money on the table before they reveal their
hands. This is not the price generated by a formula
which is only a couple of clicks away, but by per-
sonal interaction. Trading means engaging with
‘‘guys,” ‘‘dudes,” and ‘‘buds,” not following on-
screen instructions. As engagement with ‘‘guys,”
trading relies on displaying socially relevant fea-
tures. ‘‘Coolness” can be one of them, in the same
way in which a game of poker can be at least as
much about social rankings as it is about money.
Being in margin: calculation, encounters, and control
When asked about their strategy, many traders
can give elaborate accounts of the underlying phi-
losophy, of the gurus who have in?uenced them
(or not), or of long-term plans appearing as the pin-
nacle of rational behavior. Such accounts appear as
socially approved vocabularies by means of which
traders accept or mitigate responsibility for past
actions (Scott & Lyman, 1968, p. 48). When
observed in the actual process of trading, however,
what traders do appears less as the result of well-
thought, accurately applied plans (or formulas)
and more as a struggle to cope with the constraints
of the moment, to prevent the available ?nancial
resources from drifting away, to recoup past losses,
or to try and prevent future possible losses (see also
Suchman & Trigg, 1993, p. 173). Of course, lay trad-
ers can have sets of routines, which they apply
according to the speci?c constraints of their trading
accounts (e.g., standard v. retirement). Traders can
also have speci?c notions at the start of each day,
notions which they justify by reference to events
they held to be signi?cant. For instance, traders
may want to do a straddle on a particular day,
because they know that earnings are due on a spe-
ci?c stock. When active in their current account,
traders may want to ‘‘play day trading” (buying
and selling the same asset within one day), or they
may decide against it; they may want to trade deriv-
atives on stocks, on indexes, or on currencies. Such
notions, however, while justi?ed post hoc by provid-
ing rational accounts about their bene?ts, about the
‘‘mind of the market”, and the like, can provide the
starting point for, but do not determine what will
actually happen in the trading encounter. As one
trader put it, ‘‘ultimately you have to judge poten-
tial opportunities as they open up before you on
the screen.”
Some constraints can occasionally come from
personal circumstances: a large due payment may
constrain the trader to risk more in his trades, in
the hope of gaining more. The ?nancial demands
and constraints of household life, for instance, echo
in trading goals. Professional trading can be a?ected
by institutional benchmarks, the colleagues’ perfor-
mance, or that of ‘‘star traders.” Lay trading, in its
turn, can be a?ected by personal and family plans
and commitments.
A constraint looms every day in the background:
keeping out of the margin. Getting into margin can
trigger automatic liquidation of positions without
any warning. When this happens, the composition
and balance of a large chunk of the portfolio, if
not of the whole will change. Sudden changes intro-
duce an additional element of randomness, requir-
ing sometimes substantial repairs. In order to
avoid such situations, traders must, when margins
are at risk, primarily act ad hoc in order to keep
them safe: their actions are subsumed to this goal,
and they will even accept temporary, present or
future losses, in order to achieve it.
Getting into the margin is something which hap-
pens suddenly; it cannot be foreseen days or weeks
in advance. A trade which looks advantageous,
which perfectly hedges or complements previous
trades can throw traders into margin. Their focal
point is the manipulative zone of the trading
screen. The portfolio, however, together with spe-
ci?c margins for each position, is displayed in a dif-
ferent window. Keeping it open all the time would
obscure the trading screen. The required concentra-
tion on the manipulative zone of the trading
screen, together with the obliteration of the actual
portfolio situation from the traders’ visual area will
lead them to bracket out margin issues while trad-
ing. This, in turn, contributes to the suddenness
with which a crisis can occur right at the start of
the market:
Trader: .hh still in the margin though [gesture
with the right hand to the face] .hh
A. Preda / Accounting, Organizations and Society 34 (2009) 675–693 687
Ethnographer: what’s going up?
Trader: doubleju [.5
00
] it’s [points with the ?nger
to the cell on screen]
Ethnographer: oh, they’ve started. . .
Trader: [cursor on price cell on the screen; keeps
cursor there; moves cursor one cell up] [15
00
]
Christ, that doesn’t mean it’s not gonna turn
around in .hh [opens new window, checks portfo-
lio] [7
00
] "oh, I’m outta margin! Pshiii;
Ethnographer: You’re out of margin now?
Trader: yeah [opens portfolio window again,
points with the ?nger] see, no yellow, it lights
up when it’s yellow [1
0
4
00
] [puts cursor on price
cell] [6
00
] still some big losses
Ethnographer: [2
00
] lots of red there
Trader: yeah; [cursor is kept on same price cell]
[28
00
] ;damn! .Heh we’ll see how long it lasts
At the opening of the trading day the trader
banks his hopes on an open situation—prices are
falling (and he has bet on a rise). Is a turnaround
going to happen? The tension arises here between
the trader’s position (which he has de?ned as ‘‘not
good” at the start of the day, and the situation
(which will indeed change during the day). Bene?t-
ing from a change in the situation, however,
depends on tending to the position and not slipping
into margin. At the same time, nursing a weak posi-
tion prevents the trader from taking advantage of
changes on the screen. Keeping stable the de?nition
of the situation (uncertain and hopeful at the same
time) requires repairing the margin. A simple indica-
tor is the yellow band which highlights the margins
the trader has eaten into, so that the task will
become that of keeping the yellow band out of sight,
for instance by buying back some of the products
the trader has earlier posted for sale. In this case,
traders will deal with themselves. Such actions are
coordinated with, and depend on repeated checks
on the yellow band.
In more stressful situations, keeping margins
intact can take one or more days; traders will work
towards this goal until the storm has been weath-
ered. Keeping the margin is not a simple operation
which takes only a couple of seconds, but a process
of continuous portfolio readjustment. For instance,
if traders, in order to keep the margin, have bought
back some of the options they have posted earlier
for sale, other positions will be a?ected.
Shielding the trading apparatus from external,
uncontrollable disruptions is a key condition for
engaging in encounters with other traders. At the
same time, repairs cannot be conducted but by
engaging in such encounters. Engaging in trading
encounters with oneself can be done, if necessary,
but it means lack of social relationships: it is like a
temporary game of solitaire, and playing it for too
long can be damaging. The task, therefore, is to
search for encounters which will make the crisis sig-
nal—the yellow band—disappear from the screen.
Looking for relevant encounters, however, has to
take into account not only the trader’s actual situa-
tion, but also future ones.
Seeking out encounters
In order to safeguard margin, traders need to cal-
culate how the price dynamics of the underlying
asset is a?ecting the options they have sold. They
will sometimes need to buy some options back, in
order to stay within margin. If they have sold puts
with an expiration date in the following month, buy-
ing them back now will a?ect the situation in a
month from now. The task is to calculate the price
of the underlying asset at which the puts will be
wiped out—i.e., they will have to take a loss next
month. The trades (put sales) are displayed in cells
on the screen. The cells need to be correlated among
them, but this correlation cannot be done without
establishing ?rst the link between a cell, or position,
and what is in that position, a re?exive work (Sche-
glo?, 2006, p. 154) which cannot follow any pre-
established plan. In a situation of fundamental
instability, when cells ?icker all the time without
any apparent rhythm, this becomes even more
important.
Any margin-maintenance intervention (in order
to avoid forced liquidation of positions) will entail
a corresponding adjustment of other related posi-
tions; this adjustment will a?ect the margin of other
positions, which will require correction; this correc-
tion, in its turn, will trigger a readjustment of the
portfolio, which may a?ect other margins, to the
e?ect of further readjustments, etc. The trader’s sit-
uation is one of ?ow (Knorr Cetina & Preda, 2007,
p. 137), where the next sequence of action modi?es
projections of future situations, leading to other
actions and other modi?cations, etc. All these read-
justments will take place within an ever changing
688 A. Preda / Accounting, Organizations and Society 34 (2009) 675–693
setting: the screen ?ickers permanently. Calculating
trades means correlating present situations to pro-
jections of future ones, and modifying these correla-
tions with every step of action. Thus, the trader’s
task is that of tuning positions to a continuously
changing display board (well knowing that these
adjustments will introduce additional changes into
this state), while trying to avoid random, uncon-
trolled shifts (such as given by a forced liquidation).
Under these circumstances, calculating according to
a pre-set plan is futile: no such plan can forecast
how the next changes on the screen, triggered by
the interventions of anonymous strangers, will a?ect
the trader’s positions and, with them, his margin. In
such situations, open engagement with others is
required. Like a person waiting for an encounter
in the street, and who decides to become active in
approaching strangers, traders may need to actively
display their presence to anonymous others:
Trader: [after having checked portfolio page
again] ;That’s easy, I’m still in margin [3
00
] I’m
tired of this shit. [1
0
40
00
] [cursor moves across
price cells for puts on the underlying asset, lingers
on them: $0.70, $0.79, then moves over to $1.54
and transforms it into $1.55, then moves over
to $0.5 and changes the order limit from 1 to 4,
then moves on ‘‘T,” clicks, selling at $0.5. Then
opens portfolio page, checks. A yellow strip is
there.] .ah shit, back in the margin suit .hh "ha
hahaha ha [continues moving the cursor across
price cells. Changes a price, clicks ‘‘T”. Then
opens the portfolio page.] [35
00
] fuhckhhh!
Ethnographer: the yellow.
Trader: yeah. All right, well, then I guess we gon-
na get some, get some bad ass" on this.// Ya
know, it’s called gittin the baadass".
Ethnographer: //.hh [19
00
]
Trader: [10’’] "dadadadadaaadahdahaa tootoo
tatatatatataataa; [48
00
] Allright, so oh I’m getting
down there. But we might just have to roll all
these all down; [11’’] Can’t wait forever, noone
waits forever [3
00
] [moves the cursor, changes
the limit order to 20 for $1.67, there is a buy cell
on the right of the limit cell] twentee freaking sell
twenty [then moves the cursor down to the sell
order for the same price and positions the cursor
on the limit cell, waits there. Does not click, but
moves the cursor up, lingers with it in the $1.65
cell, then moves it to the right and opens the
portfolio page. The yellow strip is still there.]
Son of a ;bitchhh! Ohkey, .hhhhh [16
00
]:hhoh-
hohhohoho .hh tatata tatata tatatattaataataa
tatatatatatataaa [slaps hand on knee repeatedly]
allright, we are not gonna wait too long [moves
cursor upwards next to a buy cell, changes limit
to 50 for a $77 put, then prepares to move to
the price cell]
Ethnographer: You’re buying now?//
Trader: //I have to it’s strrr tth [4
00
] all right well
[5
00
] actually, seventy or seventynine puts [moves
cursor, scrolls page down, inserts new put at
$79. Clicks ‘‘Buy,” sets limit at 10, sets price at
$2.59, moves cursor to ‘‘T”, moves cursor back
to price, changes it to $2.60, moves cursor back
to ‘‘T”, clicks, checks portfolio page, moves
upwards to ‘‘sell” at $1.00, clicks ‘‘T”, moves
upwards to ‘‘sell” at $0.56, clicks ‘‘T”, checks
portfolio, moves cursor down to ‘‘sell” at $1.70,
clicks ‘‘T”, moves cursor upwards to sell at
$0.56, clicks ‘‘T” again, checks portfolio. Opera-
tions continue for 3
0
39
00
.]
In the above situation, the trader decides to
make his presence visible to others, to display him-
self without waiting for anonymous others to
appear ?rst. Such a display means posting trades
to be seen by others: every trade displayed can
a?ect other positions, and the trader’s task is to
balance various positions so as to make the yellow
strip disappear while not (entirely) disrupting his
future trades.
Making the yellow band disappear is not some-
thing which can be simply done by clicking a
series of buttons on the screen, according to a
set of pre-existing rules. It involves glosses which
make bodily actions (such as placing the cursor
on a speci?c cell) accountable. It is structured by
hitches, rhythmic vocalizations, and response cries
(like ‘‘son of a bitch”). It involves the trader
repeatedly announcing his presence to others
(‘‘we are going to get bad ass;” ‘‘we are not gonna
wait too long”) as a means of de?ning the situa-
tion and of making potential, remote and anony-
mous partners present in his situation.
Announcements of the own presence and accep-
tance of other presences as relevant (‘‘all right
well”) are intrinsic to accomplishing the task at
hand. The display to others is repeatedly
announced by vocalizations, situational de?nitions,
and self-summons (lines 14–16, 22–23).
A. Preda / Accounting, Organizations and Society 34 (2009) 675–693 689
Calculation is oriented here toward two interre-
lated tasks: making the yellow band disappear and
?nding relevant encounters quickly. The latter con-
tributes to the former, but should not be exagger-
ated: once enough quick encounters have been
found, this should stop. Therefore the step-by-step
and recursive character of the calculative process:
making a transaction, checking on the yellow strip,
making another couple of trades, checking again.
There is no overarching goal to this, only immedi-
ate, step- and sequence-relevant goals, analogous
to what Edwin Hutchins calls evolutionary search
(Hutchins, 1995, p. 349). If necessary, the trader
transacts with himself—i.e., buys back the puts he
has posted earlier for sale—becoming thus his
immediately available other. Choices—buying at a
strike price of seventy or seventy-nine, for $2.59 or
$2.60—seem to be determined by available pres-
ences, as well as by the necessity of quick displays
to potential anonymous partners.
The process-like, unplanned (and unplannable)
character of this ?ow of reciprocal adjustments
makes trading into something which is developed
ad hoc, depending on and drawing upon the
resources of the situation. The trader cannot
directly and immediately recognize screen changes
as expressions of an underlying plan (Hutchby,
2001, p. 138; Suchman, 1987, p. 33), to which he
should adhere or react. The trader has an array of
put options on the underlying asset (an industrial
index) which he had previously sold and which he
is forced to buy back now. The prices of puts, as
well as the strike prices of the underlying asset vary.
For instance, the trader can decide to buy puts
priced at $0.50, $0.65, etc. on strike prices of the
underlying asset of $77, $75, or $79. He has plas-
tered various combinations on the trading screen
and needs to decide among them. Every trade, how-
ever, changes the situation, which needs to be con-
tinuously assessed and then modi?ed again. Put
simply, the task is to make the yellow strip—which
indicates combinations of transactions gone awry—
disappear from the portfolio window.
Adjustments are relational: they are conducted
through brief, anonymous encounters, which
require displaying oneself to others and reacting to
the displays of others. Their unfolding does not fol-
low pre-set steps; adjustments will have to be
de?ned and justi?ed in this very process. De?nition
implies here not only an answer to the question
‘‘what is going on?” but also anticipatory elements
of the responses to the trader’s action (Smith,
1999, p. 135). A situational de?nition can require
forms of talk drawing either upon the mundane
presence of other actors, upon the trader’s own
presence, or upon the potentially relevant presence
of anonymous traders.
In this step by step process emerge interaction
forms which, while ?rmly anchored in the situation,
are not reduced to the physical presence of the tra-
der, but combine it with absent presences—that is,
with anonymous others who are brought into the
situation. At the same time, traders can cross over
to the other side of the screen, so to speak, when
their presence has to be displayed more forcefully:
Trader: [16
00
] Okay", I guess that’s enough selling
for one day [5
00
] .hh hahaha ha [16
00
] I guess I’m
gonna go back and try again [13
00
] [drinks] so
what’s my average [points ?nger to screen] it’s
seventynine", not very good, [?nger moves up
and to the left on ‘‘buy”] ?ftytwo, not very good
[continues buying at di?erent prices and limits]
[30
00
] I said I’m in margin it’s not selling lately
[22
00
] .hh [6
00
] ah donno pretty ugly, pretty ;down
ugly [checks portfolio again. Yellow strip has
reappeared.] yeah, here we go" here we go"
[31
00
] [scrolls down the page, inserts buying puts
at $79, sets limit to 3, cursor on ‘‘T”, goes back,
changes price to $2.01] ohkey", let’s see if it’s
found a ?oor now [14
00
] [checks portfolio, yellow
band still there] ;ugly, mister, bitch ugly, ;all I
can say
The announcement about withdrawing from
encounters (enough selling) is immediately followed
by a second, contradictory one (try again), condi-
tional upon locating averages, which in its turn
leads to a de?nition of the situation (‘‘pretty down
ugly”) and then to an announcement of engaging
in further encounters (‘‘here we go”). This latter
fails, and the trader addresses the presence of
remote others as a means of justifying the failure
(‘‘ugly, all I can say”).
Finding the average means here the average price
at which the trader has bought back the puts he had
previously sold. While a simple arithmetical opera-
tion, it has to be done in real time—that is, while
adjusting to the changing situation through contin-
uously buying and selling. Here, the identi?cation of
the average price on screen (through pointing at the
respective price slots) requires concomitant judg-
ment: further actions depend on whether the aver-
age is seen as good or bad. This identi?cation is
achieved by bringing others into the situation: not
690 A. Preda / Accounting, Organizations and Society 34 (2009) 675–693
only that the ethnographer’s display of availability
is taken advantage of, but anonymous others are
brought into the situation for making the failed
action accountable. The trader, who has announced
several times his intention to stop, continues trading
(and fails in making the yellow strip disappear)
because the situation is still ‘‘bitch ugly.” Such
insults and imprecations, while leaving the impres-
sion that the market is treated in an anthropomor-
phic fashion and endowed with a gender, serve in
the ?rst place situation-speci?c accountability pur-
poses, and help orient the trader’s actions. The use
of talk during trading interactions (intrinsic to the
tasks of de?ning what is going on, of calculating,
and judging) includes mundane, vernacular expres-
sions (‘‘bitch,” ‘‘baby,” ‘‘bad ass”) as consequential
situational markers.
Conclusion
The interaction order of anonymous online mar-
kets has at least the following features: (1) bodily
work oriented towards the screen as a means of sta-
bilizing and correlating data displays; (2) bringing
anonymous strangers in the trader’s situation; (3)
self-displays to others. While previous ethnographic
work has identi?ed face-to-face (Abola?a, 1996;
Zaloom, 2006) and face-to-screen (Knorr Cetina &
Bruegger, 2002) as forms of the interaction order
in ?oor-based and trading room-based transactions,
respectively, a third form can be suggested here: that
of presence-in-absence, characterized by brief,
anonymous encounters within the trader’s situation.
The screen presents itself as an observation and
experimental instrument, but also as a tool for
crossing over: an interface through which strangers
can be ‘‘pulled” into the traders’ situation, while
these latter display themselves to strangers. The fun-
damental instability of this instrument, subjected to
external constraints and continuous interventions,
uncontrollable by a single trader, makes stabiliza-
tion a crucial task. Stabilization requires a variety
of situational resources, including the trader’s own
body, as well as vocal interventions in the trading
process.
In anonymous online trading, calculations are
not identical with pre-set plans. Neither are they
identical with the application of a formula, the
results of which determine trading decisions. While
the institutional adoption of economic models
(including formulas) has been recently debated
under the banner of performativity (see, for
instance, MacKenzie, Muniesa, & Siu, 2007), the
concrete uses of models and formulas do not follow
automatically from their institutionalization.
Calculations are situational actions, geared
toward brief anonymous encounters. Such encoun-
ters are realized by making absent strangers present
into the traders’ situations, as well as by the latter
‘‘crossing over” to display themselves to unknown
others. While online traders employ routines, such
as doing a straddle or trading index options, these
routines do not represent strategic applications
based on anticipations of the opponents’ moves
and on the evaluation of the latter according to cri-
teria of e?cacy. An analogy can make this point
clear: basketball players can employ routines, such
as passes or dribbles, but have ultimately to engage
in encounters with other players on the ?eld,
encounters which will determine not only the char-
acter of a dribble and its outcome, but also the sub-
sequent sequences of action. Doing a speci?c
routine (a dribble or a straddle) appears less as a
decision enacted in speci?c situations and more like
the outcome of encounters in which participants
engage with each other in socially relevant ways.
On the trading screen, where strangers come close
to each other, socially relevant attribute relate to
acceptance, rejection, or resilience, among others.
Games of acceptance and rejection appear to be
signi?cant in online trading, but they are not
unknown in institutional trading either. To give
but one recent example: the huge losses incurred
by the French bank Socie´te´ Ge´nerale appear to have
been caused by a trader’s drive to become accepted
in the circle of the ‘‘big guys” who earn big bonuses.
Yet, returning to Geertz’s notion of deep play, the
kind of play taking place in online anonymous mar-
kets appear to be di?erent from those of institu-
tional trading: the latter seem to be about status
competitions within relatively small groups, the
members of which know each other. Here, indeed,
the analogy with Geertz’s cock?ght arena could be
pushed further, pointing at the need to examine in
detail the interaction order of status competitions,
together with the associated rituals (and conse-
quences), within ?nancial organizations.
Non-institutional online trading seems to be
more about repeatedly occurring short bursts of
social competitiveness among strangers. If, in
Geertz’s (1973, p. 449) interpretation, the cock?ght
was to be seen as a Balinese reading of the Balinese
experience, online markets can be regarded as the
A. Preda / Accounting, Organizations and Society 34 (2009) 675–693 691
story a highly individualized, fragmented, competi-
tive society tells itself.
Another relevant aspect is how price variations
emerge in the interaction order of online trading.
Such variations have been tied, among others, to
ambiguous classi?cations done by analysts (Zucker-
man, 2004) or, in the case of lay traders, to shared
judgment errors due to a lack of sophistication
(e.g., Shleifer, 2000, p. 12). This latter would imply
unavoidably bad calculations on the part of non-
institutional traders. The examination of the inter-
action order of trading, however, reveals that price
variations are an intrinsic feature of trading encoun-
ters. Their anonymity and individual character
include, if not outright require individualized price
displays, a feature supported by the capabilities of
the trading software. Since rejection and acceptance
play a signi?cant role, traders are encouraged to
vary their prices as a means of encountering other
traders. The sources of price volatility, then, should
be sought less in imperfect calculations than in the
very characteristics of this interaction order. This
also points at competitive rituals as a possible
source of price volatility within institutional
trading. It is perhaps ironic that online ?nancial
markets, with an unprecedented degree of techno-
logical penetration and the explicit aim of attracting
more and more laypeople into ?nancial activities
emerges as a platform for brief encounters, laying
thus bare calculation as social competition.
Acknowledgements
Research for this paper has been supported by a
grant from the British Academy. I am very grateful
to Karin Knorr Cetina, Donald MacKenzie, Donna
Messner, Barbara Grimpe, Stefan Laube, Vanessa
Dirksen, Cornelius Schubert, Ingo Schulz Schae?er,
and Bernt Schnettler for their comments. I am also
indebted to the anonymous reviewers for their com-
ments and suggestions, which have provided me
with very valuable insights. My greatest debt goes
towards the traders who have granted me access
to the world of online ?nancial trading.
References
Abola?a, Mitchel (1996). Making markets: Opportunism and
restraint on Wall Street. Cambridge, MA: Harvard University
Press.
Beunza, Daniel, & Stark, David (2005). How to recognize
opportunities. Heterarchical search in a trading room. In
Karin Knorr Cetina & Alex Preda (Eds.), The sociology of
?nancial markets (pp. 84–101). Oxford: Oxford University
Press.
Biggart, Nicole Woolsey, & Castanias, Richard P. (2001).
Collateralized social relations. The social in economic calcu-
lations. American Journal of Economics and Sociology (60/2),
471–500.
Callon, Michel (2004). Europe wrestling with technology. Econ-
omy & Society (33/1), 121–134.
Callon, Michel. (2007). What does it mean to say that economics
is performative?. In Fabian Muniesa, Donald MacKenzie, &
Lucia Siu (Eds.), Do economists make markets? On the
performativity of economics (pp. 311–357). Princeton, NJ
and Oxford: Princeton University Press.
Callon, Michel & Fabian Muniesa (2005). Economic markets as
calculative collective devices. Organization Studies (26/8),
1229–1250.
Fenton-O’Creevy, Mark, Nicholson, Nigel, Soane, Emma, &
Willman, Paul (2005). Traders. Risks, decisions, and manage-
ment in ?nancial markets. Oxford: Oxford University Press.
Geertz, Cli?ord (1973). The interpretation of cultures. New York:
Basic Books.
Go?man, Erving. (1972). Encounters. Two studies in the sociology
of interaction. Harmondsworth: Penguin.
Go?man, Erving (1974). Frame analysis. New York: Harper.
Go?man, Erving (1978). Response cries. Language (54/4),
787–815.
Go?man, Erving (1983). The interaction order. American Socio-
logical Review(48/1), 1–17.
Habermas, Ju¨ rgen (1987). The theory of communicative action.
Vol. 2, Lifeworld and system: A critique of functionalist reason.
Cambridge: Polity.
Heath, Christian, Sanchez Svensson, Marcus, Hindmarsh, Jon,
Lu?, Paul, & vom Lehn, Dirk (2002). Con?guring awareness.
Computer Supported Cooperative Work, 11, 317–347.
Heath, Christian, & Lu?, Paul (2007). Ordering competition. The
interactional accomplishment of the sale of art and antiques
at auction. British Journal of Sociology (58/1), 63–85.
Husserl, Edmund (1995). Cartesianische Meditationen. Hamburg:
Felix Meiner.
Hutchby, Ian (2001). Conversation and technology. From the
telephone to the Internet. Cambridge: Polity.
Hutchins, Edwin (1995). Cognition in the wild. Cambridge, MA:
MIT Press.
Knorr Cetina, Karin, & Bruegger, Urs (2002). Global micro-
structures: The virtual societies of ?nancial markets. American
Journal of Sociology (107/4), 905–950.
Knorr Cetina, Karin, & Preda, Alex (2007). The temporalization
of ?nancial markets. From network to ?ow. Theory, Culture
& Society (24/7–8), 123–145.
Landau, Susan. (2000). Communication security for the twenty-
?rst century. The advanced encryption standard. Notices of
the AMS (47/4), 450–459.
Lave, Jean (1988). Cognition in practice. Mind, mathematics and
culture in everyday life. Cambridge: Cambridge University
Press.
MacKenzie, Donald., Muniesa, Fabian, &Siu, Lucia(Eds.). (2007).
Do economists make markets? On the performativity of econom-
ics. Princeton and Oxford: Princeton University Press.
Mead, George Herbert (1934). Mind, self and society. Chicago:
University of Chicago Press.
Mead, George Herbert (1964 [1934]). On social psychology.
Chicago: University of Chicago Press.
692 A. Preda / Accounting, Organizations and Society 34 (2009) 675–693
Miller, Peter (2001). Governing by numbers. Why calculative
practices matter. Social Research (68/2), 379–396.
Miller, Peter (2008). Calculating economic life. Journal of
Cultural Economy (1/1), 51–64.
Muniesa, Fabian, & Callon, Michel (2007). Economic experi-
ments and the construction of markets. In Donald MacKen-
zie, Fabian Muniesa, & Lucia Siu (Eds.), Do economists make
markets? On the performativity of economics (pp. 163–189).
Princeton, NJ and Oxford: Princeton University Press.
Pera¨kyla¨, Anssi (2006). Observation, video, and ethnography.
Case studies in AIDS counseling and greetings. In Paul Drew,
Geo?rey Raymond, & Darin Weinberg (Eds.), Talk and
interaction in social research methods (pp. 81–96). London:
Sage.
Podolny, Joel (2001). Networks as the pipes and prisms of the
market. American Journal of Sociology (107/1), 33–60.
Schatzki, Theodore (2002). The site of the social. A philosophical
account of the constitution of social life and change. University
Park, PA: The Pennsylvania State University Press.
Scheglo?, Emanuel (1996). Con?rming allusions. Towards an
empirical account of action. American Journal of Sociology
(102/1), 161–216.
Scheglo?, Emanuel (2006). On Possibilities. Discourse Studies
(8/1), 141–157.
Schutz, Alfred, & Luckmann, Thomas (1974). The structures of
the life-world. London: Heinemann.
Scott, Marvin B., & Lyman, Stanford (1968). Accounts. Amer-
ican Sociological Review (33/1), 46–62.
SEC, 2000. Special study: Report of examinations of day-trading
broker-dealers. February 25, 2000. Available from http://
www.sec.gov/news/studies/daytrading.htm.
Shleifer, Andrei (2000). Ine?cient markets. An introduction to
behavioral ?nance. Oxford: Oxford University Press.
Simmel, Georg (1971 [1908]). The stranger. In Donald N. Levine
(Ed.), On individuality and social forms (pp. 143–149).
Chicago: University of Chicago Press.
Smith, Charles W. (1999). Success and survival on Wall Street.
Understanding the mind of the market. Lanham, MD: Row-
man & Little?eld.
Suchman, Lucy (1987). Plans and situated actions. Cambridge:
Cambridge University Press.
Suchman, Lucy, & Trigg, Randall H. (1993). Arti?cial intelli-
gence as craftwork. In Jean. Lave (Ed.), Understanding
practice. Perspectives on activity and content (pp. 144–178).
Cambridge: Cambridge University Press.
Vertesi, Janet (2008). Mind the gap. The London Underground
map and users’ representations of urban space. Social Studies
of Science (38/1), 7–33.
Vollmer, Hendrik (2007). How to do more with numbers.
Elementary stakes, framing, keying, and the three-dimen-
sional character of numerical signs. Accounting, Organizations
and Society (32), 577–600.
Wacquant, Loic (2004) Body and soul. Notebooks of an appren-
tice boxer Oxford: Oxford University Press.
Wittgenstein, Ludwig (1984). Vorlesungen 1930–1935. Frankfurt:
Suhrkamp.
Zaloom, Caitlin (2006). Out of the pits. Traders and technology
from Chicago to London. Chicago: University of Chicago
Press.
Zuckerman, Ezra (2004). Structural incoherence and stock
market activity. American Sociological Review (69), 405–432.
A. Preda / Accounting, Organizations and Society 34 (2009) 675–693 693

doc_831566103.pdf