The emergence of trust in extremely hostile

1/32
The emergence of trust in extremely
hostile environments. The case of
covert support networks for persecuted
Jews in National Socialist Germany*
Abstract...........................................................................................................................2
Keywords ........................................................................................................................2
Introduction ....................................................................................................................3
Literature review ................................................................................................................................................................ 3
Definitions.............................................................................................................................................................................. 6
Historical context: Segregation and persecution 1933-1945 ................................................................................. 6
Method ............................................................................................................................7
Sources ..................................................................................................................................................................................... 8
From text to data ................................................................................................................................................................... 8
Case studies ..........................................................................................................................................................................10
The emergence of trust .................................................................................................. 11
Trust dyads and transitive trust triads between strangers......................................................................................13
Transitive trust chains .......................................................................................................................................................15
Trust in resistant communities .......................................................................................................................................17
Dynamic growth of support networks .........................................................................................................................20
Discussion and Conclusion ............................................................................................ 22
Bibliography .................................................................................................................. 26
Word count including notes and bibliography: 11 632
*
Author1 wishes to thank the Gerda Henkel Foundation for generous funding which made this research possible. In addition,
we thank the Gedenkstätte deutscher Widerstand Berlin for their support in pursuit of this project.
2/32
Abstract
Trust is essential to the formation and functioning of social networks. Little is known how
trust spreads in extremely hostile environments such as dictatorships where actors face
persecution and operate in secret. We examine trust in support networks for Jewish fugitives
in Berlin 1942-1945 based on manually extracted data on acts of help from historical sources.
We find that trust emerged in multifaceted dyads sometimes in a grey area between knowing
and not-knowing, grew into transitive triads by means of contact brokerage which in turn
formed short-lived chains. These chains provided fugitives with new helpers and support
networks with essential resources but also lead to their uncontrollable growth and eventual
demise.
Keywords
Trust, transitivity, social support, covert networks, National Socialism, mixed methods,
historical network research
3/32
Introduction
This paper examines the emergence of trust in an extremely hostile environment. We
examine how Jewish fugitives and their helpers created trusted ties and support networks in
National Socialist Germany. Jews had been systematically disempowered and isolated for
years and found themselves in a mortal fight against an overbearing dictatorship which used
its secret police, administration and control over citizens to create highly effective measures
of persecution. Between 1938 when emigration practically became impossible and 1940,
when deportations from the Reich began, conditions for effective help changed dramatically.
Resources became scarce due to the war effort. For fugitives, old trusted ties were broken and
the risk associated with trusting strangers rose. Survival depended on access to
accommodation but also special skills and knowledge including forgery, black market trading
for documents and food stamps. To gain access to these resources, fugitives and helpers had
to create trusted relationships under constant time-pressure and without the benefit of the
shadow of the past or future. Contact brokerage and fake identities had to compensate for
broken up communities and old trusted ties. While contact brokerage reduced the associated
risks, often fugitives and their helpers had no other choice but to take leaps of faith.
We find that dyadic trust ties grew into transitive trust triads in which one person took the
position of trust broker, a hitherto overlooked, highly effective form of help. These triads
linked into long chains which mobilised new helpers, facilitated the exchange of resources
and were overall relatively safe. These chains enabled support networks to help effectively
but also lead to their uncontrollable growth and detection. Embeddedness in social structures
softened the seeming dichotomies of “trustor” and “trustee”, “helper” and “recipient”.
Fugitives and their helpers regularly found themselves in states of mutual dependency.
Our research is based on eight networks which were active in Berlin, four of them are
focused on the survival of fugitives, the other four focus on the emergence of dedicated
support networks. Data were extracted manually from primary and secondary sources and we
have selected three of the eight case studies to illustrate our findings. Our detailed
reconstruction of selected examples is not statistically representative. Instead, we hope to
shed more light on the social dynamics which underlie the creation of trust. We argue that
trust in extremely hostile environments is typically multi-layered and grows within network
structures rather than in isolated dyads
Literature review
The findings of this study contribute to several fields: social support, trust, social network
analysis as well as history of National Socialist Germany. In our literature review we
therefore limit ourselves to studies with immediate relevance for our topic.
4/32
Trust has widely been acknowledged to be essential to the formation and functioning of
social networks. McKnight and Chervany (2001) point to the different conceptualisations of
trust across the social sciences. They observe that “that psychologists analyzed the
personality side, sociologists interviewed the social structural side, and economists calculated
the rational choice side of the trust elephant” (McKnight and Chervany 2001:29).
Trust is particularly important for covert networks in which actors face personal persecution
and need to operate in secret. Research on covert networks typically covers activities which
are commonly considered to be undesirable, such as drug and arms trafficking (Milward and
Raab 2002), illegal collusion (Baker and Faulkner 1993) or terrorism (Krebs 2002). Actors in
covert networks may face imprisonment (Erickson 1981; Lampe and Johansen 2004;
Morselli, Giguère, and Petit 2007; Simmel 1906) but the stakes are still significantly lower
than in our case. However, even though trust is often argued to be necessary and functional
for the development, sustainability and resilience of dark or covert networks (Bakker, Raab,
and Milward 2012; Morselli et al. 2007), this claim has rarely been investigated in more
detail to date.
Similarly, research on social support typically considers much less extreme environments
such as workplaces or neighbourhoods (Kupfer and Nestmann 2015; Podolny and Baron
1997; Wellman 1979; Wellman and Wortley 1990). Social support research is typically
concerned with wellbeing, stress management, performance or health risks and focuses on
long-lasting social relations. Our case studies differ from those considered by similar research
in at least three ways: First, they involved high risks for both parties of the relationship, i.e.
fugitives and their non-persecuted helpers. Betrayed trust had severe consequences for
fugitives who faced certain death and for their helpers who had reason to fear imprisonment
in a concentration camp and other reprimations (Kosmala 2007). Second, we study deviant
behaviour which is prosocial and highly desirable from a normative perspective. Third,
commonly observed mechanisms of trust creation and repercussions for defectors do not
apply.
Helping behaviour in the face of mass violence and the Holocaust in particular has interested
historians and social scientists alike. Historical research points to the diversity of motivations
among helpers and in the practices of help and is skeptical towards attempts to identify
common features (Benz n.d.; Grabowski 2008; Kosmala 2007). Past research has focused on
helping behaviour in the occupied countries (Moore 2010), money and other rewards as
motivation (Grabowski 2008), local histories of help and resistance in different contexts
(Benz n.d.; Moore 2010; Semelin 2011) but also refers to the importance of situational factors
(Gudehus 2015; Paldiel 2004). Social scientists in contrast have sought to identify
commonalities among helpers. Wolfson (1971, 1975) studied the demographics of 70 helpers
and found that age, education or gender were evenly distributed. London (1970) identified a
desire for adventure, strong ties with parents and a perceived social marginality as the three
common traits of helpers. Oliner and Oliner (1992) showed that roughly two-thirds of all
helpers whose cases were documented in the Israeli memorial site Yad Vashem responded to
requests for help. Yad Vashem however has strict guidelines which consider only selfless and
unpaid acts of help. Oliner and Oliner compared 400 helpers with 100 non-helpers and found
5/32
that helpers had a greater sense for responsibility and stronger attachments to people which
they attribute to their upbringing. Fogelman (1995) distinguishes between emotional and
moral motivations but also points to values and convictions deeply rooted in the personalities
of helpers. Monroe (2008, 2011) and Monroe, Barton, and Klingemann (1990) highlight the
importance of helpers’ empathy with fugitives based on a close understanding of their
suffering. Varese and Yaish 2000 did a secondary analysis of Oliner and Oliner’s data and
highlight the importance of requests for help as well as pre-existing relationships and the
expectation of contact brokerage.
Overall, historians link helping behaviour to national scopes of actions and social scientists
tend to focus on helper personalities. We argue alongside Varese and Yaish (2000) that the
decision to help was not exclusively a question of scopes of action or predisposition but also
one of social embeddedness. This implies that selfless and selfish motivations and interests
among those involved did not exist in separate spheres but that survival as well as effective
help depended on their integration.
To map such embeddedness, we make use of Social Network Analysis methods. A number of
articles engage with the potential and challenges of historical network analysis (Düring and
Eumann 2013; Erickson 1997; Lemercier 2015; Neurath 2008; Reitmayer and Marx 2010;
Wetherell 1998). In what follows we will therefore simply contrast different domains and
applications of network analysis methods and focus on studies related to our own research.
Erickson (1981) describes underground organizations in a structural perspective in Auschwitz
concentration camp on the basis of memoirs and documents as well as some survivor
interviews in a qualitative way as well as the rebellious White Lotus Sect with data from
confessions of arrested members, palace memorials, official communications, and captured
sect documents. In his attempt to explain the political development in Italy that led to the rise
of Fascism in the 1920s, Franzosi (1998, 1999) uses newspaper reports and police records to
extract relational data and to map the changing relations between major societal groups and
collective actors. Crossley et al. (2012) investigate the secrecy-efficiency trade off in covert
networks with data on the UK suffragette movement in the early 20th century. More recently,
Crossley et al. (2012) studied meetings of the inner circle of the IRA. In some studies,
membership or sponsor lists are used to map structures and explain developments of protest
movements over time as in Osa (2003) for Poland or in Bearman and Everett (1993) for
protest movements in the United States from 1961 to 1983 (co-sponsoring of protest rallies).
Rosenthal et al. (1985) used biographical dictionaries to identify women reformers in
nineteenth-century New York State (1840-1914) and the inter-organizational structure within
the women’s movement. This research shows that in order to gain relational information
about social structures from archival sources, researchers have to make creative use of the
available sources. Scholars however rarely extract relational data from texts.
Help, rescue and support
In the course of this paper we use the words “help” and “support” interchangeably. They refer
to any documented actions which directly or indirectly provided fugitives with essential
6/32
resources which facilitated their survival whether or not they were selfless in nature. Among
such resources are accommodation, food or food stamps, forged passports, new contacts but
also emotional support. Anyone who provides such resources is considered a “helper”,
anyone who receives them is considered a “recipient”. The two terms are used independently
of the racial categorizations which applied in Germany at the time and people can be helpers
in one situation and recipients in another. We prefer the terms “help” and “support” to
“rescue”. Where “rescue” is associated with a passive recipient, “help” and “support”
highlight the agency of fugitives as well as the temporary nature of any contributions by third
parties (Merriam-Webster n.d., n.d.). Where this distinction is relevant, we characterize
helpers as “non-persecuted” to describe those who may or may not have counted as “Aryan”
but either way were spared from active persecution. We define support networks as groups of
helpers who supported more than one helper at the same time. Persecution occurred on the
grounds of National Socialist race classifications which determined who was to be considered
“Jewish”, “Half-Jewish” and “living in a privileged mixed marriage”. We operate with these
classifications and use them as labels since they determined the scope of action of a given
person within National Socialist Germany. These terms do not necessarily represent a
person’s self-perception or religious beliefs.
Historical context: Segregation and persecution 1933-1945
The actions of helpers and survivors can only be understood against the background of highly
restrictive social control in National Socialist Germany. Immediately after coming to power,
Hitler had ordered the arrest of numerous political opponents and forced many others into
emigration. This meant a significant weakening of any political opposition and the
destruction of trusted networks which could have been transformed into support networks, as
was the case in the occupied countries (Moore 2010). In Germany most forms of public
protest had been quieted after 1939. Whoever wanted to resist National Socialism would
become part of the “loneliest of all European resistance groups” and belong to a small
minority without any influence before and during the war – observed and denounced by large
numbers of spies, Blockwarte, members of housing communities who ensured that residents
complied with the regime's demands as well as ordinary Germans who sought to do their duty
and report any “suspicious” activities (Lustiger 2011:17). Together, they facilitated an
effective control instrument and posed a constant danger for fugitives who needed shelter and
their hosts (Gellately 1996).
Soon after Adolf Hitler’s rise to power on 30 January 1933, anti-Jewish measures were put in
place: by March 1933, Jewish civil servants were forced to retire. This was the first of many
newly introduced laws and measures which served to systematically isolate Jews from the
rest of society. Policies for Jewish emigration were coupled with the systematic robbery of
material goods and financial assets and by 1938 emigration became nearly impossible. In
October 1940 the first deportations of Jews in the southwest of Germany to the Gurs
concentration camp began, one year later deportations began all over the Reich. Berlin had
become a place of refuge in the years before for many who had hoped to find a way to
7/32
emigrate. After years of systematic segregation and faced with their deportation to
concentration camps, 164 000 Jews remained in Germany. An estimated 12 000 Jews chose
to go into hiding within Germany, 7 000 tried to hide in Berlin. On 27 February 1943, during
the so-called Fabrikaktion all remaining Jews in forced labour were supposed to be deported.
An estimated 4 000 Jews managed to evade their deportation and made an attempt to survive
in Berlin (Kosmala 2004:140). Some 1 500 of them managed to survive until the end of the
war (Kosmala 2011:34). While deportations in the Reich had already begun in October 1941,
they were held back in Berlin until a year later; valuable time for Jews and their helpers to
learn about the real purpose of the “relocations” and their consequences. Still, most Jews in
Berlin wavered “between not-knowing, knowing and not-wanting to know” (Karl-Heinz
Reuband cited in Kosmala 2004:136).
Jews and their helpers had to adapt to ever more aggressive measures of persecution and
wartime restrictions. Jews whose looks matched the anti-Semitic clichés of “Jewishness”
faced particularly high risks. On top of this, young men were always suspicious of having
deserted from the Wehrmacht and therefore were subject to frequent identity checks by the
police. They in particular had to rely on forged documents. Another threat came from Jewish
spies, fugitives who had been caught and forced to help catch others. Threatened with their
own death or that of their families they used their inside knowledge of covert networks to
find and denounce helpers and other fugitives (Tausendfreund 2006). Throughout the war
fugitives had been without protection from the Allied bombardments, only few could take the
risk to hide in bunkers, because entry required identity checks (Kosmala 2011:35). Still many
fugitives recounted that they associated the bombings with the prospect of a soon to come
Allied victory. In the course of the deportations from 1940 onwards, Jews had become more
and more absent from German society, to the extent that potential helpers would not even
believe Jewish fugitives when they dared to reveal their true identity towards the end of the
war (Schieb 1989). Previous social networks of Jews and (potential) helpers had been
shattered, only small fragments remained. Jews and those who were willing to help were
isolated and faced a high risk of denunciation by whoever they chose to confide in. Relations
between both parties therefore had to be transformed from weak or non-existent to very
strong and reliable in an extremely short time period. From a fugitive’s perspective, a large
number of smaller acts of help facilitated survival, albeit at the price of additional risk of
being reported to and arrested by the German Secret Policy (Gestapo). Clusters of helpers
emerged in small, un-Nazified communities of typically no more than 12-30 individuals
which Martin Broszat (Broszat 1986; Broszat and Fröhlich 1987) named “resistant” with
reference to the medical connotation of the term and hinting at previously underestimated
scopes of action.
Method
In the following section we describe our research approach which combines traditional close
reading techniques with methods developed in Qualitative Data Analysis and Social Network
8/32
Analysis. Given the historical nature of the research with a challenging situation with regard
to data collection we especially focus on the availability and accuracy of information.
Sources
A key criterion for our selection of case studies was the availability of sources which would
allow us to reconstruct support activities in great detail. A vast amount of materials is made
available by institutions such as Yad Vashem in Jerusalem or the German Gedenkstätte
deutscher Widerstand in Berlin. Survival, public and scholarly interest determine whether and
to which extent we know of helpers and fugitives. This causes problems of over- and underrepresentations of actors and activities.
The only way to collect a substantive amount of data was to consider all available sources:
Gestapo interrogation reports, applications for remuneration, memoirs, interviews and earlier
historical research. The consequence is that to some extent, biases inherent in these types of
documents are introduced into our dataset: During an interrogation, we may expect that
people tried to diminish their own roles, merely confirm what the interrogator knows already
and to avoid mentions of those who are not yet known to the interrogator. In applications for
remunerations applicants had to give evidence of their resistance activities and different
forms of damages (financial, physical, careers, etc.) which they suffered as a result. Memoirs
and interviews focus on one person and their interpretation of events and their role in them.
We observe an emphasis on the individual and little interest, to provide a representative list of
involved third parties. Historical research publications were useful inasmuch as they were
based on efforts to collect everything that was known about specific acts of helping behaviour
and to some extent on accurate descriptions of how persons interacted with each other. Still,
any reconstruction is inevitably based on fragments of sources with strong inherent biases.
For example, descriptions of the creation of trusted dyads to new helpers tend to be very rich
in survivor’s memoirs whereas their authors often cannot make statements about the overall
embeddedness of their helpers. Sources on support networks tend to lack detailed information
on helpers who were more than two steps away from the centre. But they do provide highly
detailed descriptions on the roots of the relations among the core members of the network.
From text to data
In the overwhelming majority of sociological network studies data collection is conducted via
questionnaires. This method of data collection clearly has many advantages such as providing
a way to bound the network, define what constitutes a link and collect data on larger
networks (Calloway and Morrissey 1993:380). Moreover, before the recent advent of data
mining possibilities, collecting relational data through surveys has presumably been the most
cost effective method for the variety and amount of data that can be collected. For historical
networks we need to rely on archival sources which were created either during the events or
9/32
in their aftermath which poses a number of limitations and challenges. While these data are
relatively easy to collect, and therefore, inexpensive, especially when in electronic form, the
underlying logic of tie formation is often not clear Marsden 2005. In what follows, a method
will be presented to collect relational data from archival sources, especially when available
data is limited and fragmented. While the number and range of network studies collecting
data from archival sources has increased, according to (Marsden 2005:24), relatively few
explicitly methodological studies of archival data have been conducted, warranting additional
attention.
We applied methods developed in Qualitative Data Analysis to bridge the gap between
hermeneutics and formal data collection (Miles, Huberman, and Saldaña 2014). We
developed an abstract definition for each subtype of a relation (such as “provision with food”)
accompanied by examples from the primary sources which indicate whether or not a subtype
applies. Help with accommodation for example was defined as follows:
All descriptions of provisions of housing, shelter, hideouts. The code applies to shortterm as well as long-term accommodation. This includes non- persecuted persons as
well but excludes people who hide e.g. in a subway station without the help of
somebody else.
Example:
“Through him, we were referred to a landlord who had a vacant one- bedroom
furnished apartment. For twice or thrice the normal rent, he would let us have it.
Mom agreed, and we had a place to live and hide in.” (Neuman 2005:4)
Definitions for each code are published in German in Author 2015. The process of data
extraction is grounded in historical context knowledge and careful interpretation. The
meaning of concepts such as “acquaintance” or “friend” for example is just as manifold as are
the forms of help and how they have been experienced: An apple given to a Jew wearing the
yellow star could become a meaningful act of emotional support while demonstratively
offering a Jew a seat in a crowded bus could be interpreted as a likely cause for public
shaming or punishment and therefore a threat (despite the good intentions).
Each act of help (food and commodities, accommodation, forged documents, border crossing,
legal aid, emotional support, contact brokerage, other) was coded in one row and describes
the form of help, the quality of relations between helper and recipient of help, the time frame
of their first meeting and the time frame when the act of help occurred. Where two or more
acts of help co-occurred (e.g. food and accommodation), we added a separate row to
acknowledge this. Similarly, reciprocal help was also represented by two rows. Contact
brokerage, one of the most effective forms of help as we will show, was coded slightly
differently. The process involves three people: A beneficiary, a broker and a new helper. To
represent this triad, we first code relations between broker and beneficiary and broker and
new helper with the code “contact brokerage”. In a second step we code the act of help which
was later provided by the new helper to the beneficiary, for example the provision of
10/32
accommodation. This allows us to keep track of who acts as a broker and the path structures
which emerged as a consequence.
The periodization of acts of help posed problems. While some acts of help could be dated to a
specific day, most others could be either tied only to a season (“in the summer”) or a specific
year. To keep our approach simple and robust, we limited ourselves to biannual time steps.
Attributes such as the National Socialist race classification provided an approximation of
scopes of action within the German society.
This drastic abstraction and reduction of the complexity of social interactions followed the
concrete interest to map network structures which emerged as a consequence of dyadic
helping behaviour. All codes and code categories were developed, tested and revised in
several iterations using different documents to ensure the best possible match between
research questions and available information. Wherever sources were too vague or unreliable
we chose to not add a new relation. This practice and the fact that many acts of help were
never reported and are now forgotten means that our network models likely lack density.
Network boundaries were defined by the available archival sources. In some cases, links
existed to other networks. These relations were considered inasmuch as they were
immediately relevant for the operations of a network. Immediate relevance was
acknowledged when essential resources such as money or documents travelled from one
cluster into a network. Where no such relevance was observed, we chose to ignore them.
In total we manually extracted data on 1 500 actors connected by 5 000 acts of help across
eight case studies.
Case studies
From our eight case studies we have selected three which we will describe in greater detail to
exemplify our findings: The first case revolves around Erna Segal (born in 1898) who was a
married Jewish woman with two children aged 12 and 15 in 1939. She survived in Berlin
together with her family and died in August 1989 in Denver, Colorado. While most Jewish
families who went into hiding split up, the Segals remained in close contact and exchanged
information about helpers. Erna Segal’s detailed report is a very rich source from the
perspective of fugitives who are in constant need of new helpers and hide-outs. In the case of
Erna Segal not only for herself but also for her family. Their was as much a question of good
luck as it was testament to their resourcefulness and the help of 79 helpers, half of whom
were complete strangers. The second case study focuses on Luise Meier, a catholic widow
born in 1885 who experienced measures against the Jewish population from close proximity
and decided to help her Jewish neighbours escape to Switzerland. When she was approached
to help again, Meier agreed and thereby unwillingly started a chain reaction which lead to the
emergence of a support network. We can reconstruct the activities of this network from the
perspectives of fugitives and their helpers. Luise Meier survived the Second World War and
died in Soest, Germany in 1979.
The support network surrounding Franz Kaufmann (our third case study) provided fugitives
with forged documents and food stamps and quickly grew into the largest covert support
network in the German Reich with ca. 400 actors involved. Kaufmann, a former mid-level
civil servant in the state of Prussia was born in 1886 to Jewish parents but was raised
11/32
protestant in a German nationalist family. In the early 1930s Kaufmann is responsible for the
reorganisation of municipal administrations. He was ambitious and did not avoid conflict
when he consciously bypassed his superiors to make his new concept known to the Minister
of Finance. Kaufmann was arrested by the Gestapo in August 1943, incarcerated and killed in
the Sachsenhausen concentration camp in February 1944. Several of the actors have
described the actions of this network and a dedicated monograph (Rudolph 2005) was
available for analysis as well.
We find key initiators in all of the support networks we studied. Like Meier, Segal and
Kaufmann they stand out as strong-willed personalities who were ultimately willing to
overcome boundaries and break rules, identify scopes of actions where others failed to see
them and lead passive albeit sympathetic bystanders into action. All of them (like many
helpers) had started to help in the mid 1930s by means of emotional support or legal advice
and thereby witnessed the radicalisation of National Socialist persecution policies from close
proximity. The three case studies we selected for this paper all highlight the key dynamics,
strategies and social interactions which were the basis for the formation of trust within
National Socialist Germany. In the following we will investigate how the process of trust
creation between individuals and in these helper networks evolved under these extreme
circumstances. We hereby focus on four themes: The rapid building of trust between
complete strangers, the emergence of transitive trust chains, trust within resistant
communities, and the rapid and dynamic growth of support networks which in their success
carried the seed of their destruction.
The emergence of trust
Recruitment, resources and structure are considered to be the key elements for the emergence
and functioning of covert networks (Erickson 1981) as is the necessity to trust. In extreme
circumstances such as those in Berlin at the time with a high risk of detection and arrest, it
was hard to develop trust for several reasons: Strangers could turn out to be spies and trusted
ties could be corrupted either by arrests or a change of hearts by the helpers. Both helpers and
fugitives therefore had to develop ways to reduce risks before they could engage in
recruitment or resource exchange. In our data, we observe combinations of different forms of
trust: trust in the willingness to help, trust based on mutual benefits, trust based on fake
identities, transitive trust based on contact brokerage. Trust manifested itself in different
structures: dyadic relationships, alongside paths, within clusters and across clusters as
exemplified in Table 1. Trust grew by means of brokerage and the acts of help which followed
spread uncontrollably, driven by short term necessities. Both the support networks and the
ego networks of fugitives therefore differ from other covert networks or “secret societies”,
which are more frequently characterized by a higher proportion of planning and more formal
organizational elements.
12/32
Strengths
Weaknesses
Dyads
Can either be very strong and durable
or fleeting.
Once trust is established, easy to
reactivate.
Hard to create.
Limited access to new resources.
Triads
Robust means to broker trust between
strangers based on two trusted dyads.
Hard to find a trustworthy broker.
Broker’s selection may be risky.
Out of control of fugitives.
Paths
Span geographical and social
distances.
Relatively low risk for everybody
involved due to low density, relatively
easy to protect from investigations.
Facilitates movement of people and
resources.
Fleeting, high risk of fragmentation,
low degree of redundancy.
Limited performance.
Can be reconstructed
Clusters
High level of redundancy allows
division of labour and high
performance.
Embeddedness ensures trust and
continued collaboration.
Risk of isolation, lacking access to
resources.
Higher density means higher risk of
detection.
Table 1 Strengths and weaknesses of structural patterns in the context of covert support networks.
As we will show below, networks could grow against the will of at least some of the leading
actors. Fugitives were not able to plan ahead for more than a few weeks at most but usually
less. Most helpers were only involved for a limited period of time before for example
sheltering fugitives would raise suspicions. Those who were active in support networks often
helped a large number of fugitives one step along the way by means of documents or food
stamps.
Within support networks, we do not observe formal hierarchies but rather a very small
number of highly involved actors typically led by one initiator as the driving force. These
initiators were able to organise an infrastructure for the exchange of rare resources such as
passports or guides for illegal border-crossings between those who needed them,
intermediaries and those who could provide them. These infrastructures opened up new
scopes of action which allowed potential helpers to get involved, independently of their
personal motivations and interests. Involvements in such exchanges did not necessarily equal
a commitment to help Jews or an awareness to be part of a group. There are hardly any
attempts to effectively steer recruitment and the distribution of resources beyond the attempt
to meet the demands for help. Similar to Erickson we observe that “the more decentralized
13/32
the control over recruitment, the more recruitment simply spreads like a diffusion process and
generates a structure with very few hierarchical features” (Erickson 1981:197). This lead to a
paradoxical situation: Given the circumstances, planned and formally organized action was
very limited and dangerous. People were forced to commit to unplanned, spontaneous trust to
reduce risk. However, as we will see below in the case of the Kaufmann and Meier support
network, it was exactly that uncontrollable and dynamic process of relationship formation
and development of the network structure that ultimately again increased the risk of
detection.
Trust dyads and transitive trust triads between strangers
When Jews decided to go into hiding, they often found themselves socially segregated with
most of their friends and family gone or out of reach. Naturally, they would look for help
from those few contacts they could still rely on. Many of these early helpers were
acquaintances, people they had met only months before they went into hiding. Most helpers a
fugitive would encounter after this initial phase, however, were complete strangers. There are
no reliable statistics available, but an estimate by historian Beate Kosmala points towards
seven dedicated helpers per fugitive (Kosmala 2007). Our case studies suggest that this
number could also be as high as 80-100 helpers per person. This estimate takes into account
that fugitives had many fleeting or indirect relationships with helpers who nevertheless
provided essential support. Dyadic relationships between fugitives and their non-persecuted
helpers could have extremely uneven distributions of power: A helper who would choose to
defect had to fear few negative consequences; abuse of this power would simply go unnoticed
and unpunished. Fugitives were therefore often very much concerned to maintain good
relations and for example not to overstay their welcome.
Erna Segal, whom we mentioned in the beginning gave a detailed account of how she and her
family made contact to helpers and built relationships with them. Their experiences, tactics,
fears and problems are exemplary for those of other fugitives. Erna Segal describes her
search for suitable hideouts at the beginning of her life in the underground:
„Now we considered all Aryans which were absolutely trustworthy. I had a good tailor which
a friend of mine and I regularly used. (…) I would sit there for hours, helped out a little in
order to learn about sewing with the idea to make a living with this once we emigrate. I
decided to trust these people. I went to see them and told them about my plan. They were very
well informed and had even agreed to look after a number of suitcases with valuables of one
of our friends who had been forced to leave the country. These people hoped to come back
and had great trust in the tailor. Mr and Mrs Dovsky [the tailors] immediately agreed to my
plan. They would take my son and me in. We agreed on a substantial monthly payment and in
addition I would run their household and help them with the sewing.” (Segal 1956:94–95).
Segal selected these helpers with great care: She had known them as good craftsmen and
gotten to know them personally during a few days of joint work. Secondly she relied on
information by third parties which confirmed that the couple had in the past proven their
14/32
opposition to the Nazis and had helped some of her friends. Knowledge of this previous act
of help for a mutually trusted fugitive served as a direct recommendation and confirmed her
trust in the couple. This helped both Erna and the tailors to minimize the risk of being
betrayed by the other and allowed them to build a trusted relation. Fear of Gestapo spies who
played the part of fugitives was real among helpers and therefore they too needed to be able
to trust whoever they decided to help. This simple triangular constellation between two
strangers and a mutually trusted third party who would allow strangers to develop trusted ties
occurs in all cases of support for fugitives and was the basis for the brokerage chains which
will be discussed below.
The tie between Erna and the tailor couple was further strengthened by the fact that she
would pay them a “substantial” amount of money, help them in their household and
contribute to their business. This must have made this arrangement significantly more
attractive to the helpers. In summary, Erna had four reasons to hope that the Dovsky couple
would keep their promise:
1. Opposition: Her knowledge of their opposition to National Socialism
2. First impressions: Her own impression of them
3. Track record: Her friend’s recommendation
4. Material gains/ Exchanges: Her payments and other services she offered.
However, when she noticed that Dovsky’s daughter went out with a “Nazi”, the family left
their hideout immediately, fearing that it had become unsafe for them to stay. But wherever
possible, Erna would rely on these four factors when developing trusted ties to strangers.
There are variations to this procedure. When Erna started to chat to an older man in a park
she self-identified as Aryan. She constructed a narrative which consisted of half-truths. She
presented herself as a victim of a recent bomb raid and claimed one of her sons was in the
Wehrmacht (in fact he was held in a concentration camp in France). The man promises to
introduce her to acquaintances outside Berlin. When they both arrive there, he introduces
Erna as an old friend in need of help and thereby vouching for her trustworthiness. His act of
help may be based on the false premise that Erna was Aryan, he may also have doubted her
and kept these doubts to himself. Either way, his lie to the new hosts created a new trusted tie
between them and Erna. Both fugitives and their helpers could hide behind such half-truths:
Fugitives appeared more credible and thereby exploited “Volksgemeinschaft” solidarity in
Germany.
In these contexts, honesty and mutual benefit in a dyad are replaced by uncertainty on both
sides. What counts is the willingness to help and it is irrelevant to which extent one believes a
story. This leads to three scenarios in which help can occur:
1. The person believes the fugitive’s cover story and helps.
2. The person has a suspicion but helps anyway.
3. The person does not believe the fugitive but helps anyway.
Fugitives cannot know for sure whether or not their new helper believes their story but, as the
example from above shows, it does not matter. From the perspective of potential helpers,
half-truths made it easier to legitimize involvement and to avoid cognitive dissonance
15/32
(Festinger 1957). Solidarity with Wehrmacht soldiers and their families as well as with
victims of air-raids was socially expected behaviour. While Erna’s needs and worries were
real, the way she framed them let her tap several sources of solidarity at the same time (with
fugitives, with victims of air raids, with people in need in general, with “Aryans”) and receive
help without knowing what her helpers thought of her. But even if a fake cover story was
detected and the approached person was not willing to help but also not keen to denounce
after a suspicious request for help, the cover story permitted a non-confrontational way to
decline.
Most fugitives were under constant pressure to find new helpers and had to develop a
minimum of trust with strangers in a short period of time. Cover identities and the above
mentioned uncertainties on both sides provided an effective way to sidestep long trust
building processes. In cases where helpers were aware of the true identity of fugitives, the act
of help itself generated some degree of trust since the provision of help meant that a person
was willing to incriminate his- or herself and be exposed to the risk of detection and
punishment.
Transitive trust chains
Transitivity usually describes the observation that a friend of a friend is also a friend
(Wasserman and Faust 1994:150;165). Transitivity can be represented by closed triads in
which all possible relations are realised. We find that such transitive triads were essential for
the creation of trust across all case studies. In the previous section we discussed a situation in
which a stranger who trusted Erna Segal went on to introduce her as an old friend to his
acquaintances. Faced with this direct request for help, which is known to be a crucial factor in
triggering helping behaviour (Latané and Darley 1970; Latané and Rodin 1969; Varese and
Yaish 2000), the approached individuals were not so much facing the decision whether they
should help strangers they had never met before; they were confronted with a request for help
from someone they knew and who trusted them and felt an obligation to comply: Declining
the request would mean refusing to help whoever approached them. This created a significant
incentive for outsiders to get involved in helping behaviour.
Figure 1 Contact brokerages provided by non-persecuted helpers, "Half-Jewish" helpers and “Jewish” fugitives
16/32
Figure 1 shows that in the majority of cases non-persecuted helpers brokered contacts between
each other or for fugitives. Still, Jewish-Jewish and Jewish-non-persecuted contact
brokerages rank third and fourth. This defies the dichotomy of passive fugitives and their
active non-persecuted “rescuers”. Rather, the high number of contact brokerages by fugitives
suggests that despite their precarious situation they played an active role in their own survival
and that of others.
But brokerage did not stop there: Like Segal’s helper in the park, others also began to search
for new helpers in their own social networks and subsequently took on the role of a trusted
broker. Thereby, each new trusted tie opened up a new pool of potential helpers. We find
these chains of trust brokerages in all fugitive ego networks. Figure 2 shows brokerage chains
for the Segal family (the family was removed from the graph to increase legibility). The
graph connects brokers with each other; arrows indicate who approached whom. For
example, an unspecified couple (“bekanntes Ehepaar”) approached a couple named Scholz
who continued to approach a woman named “Frau Weil”. The Scholz couple therefore
switched from being considered as potential new helpers into the role of brokers. The long
chain-like structures reveal that this form of trust brokerage occurred up to seven times in a
row. A closer examination of the graph shows that these brokerage chains were also capable
of crossing diverse social classes, ranging from upper-middle class to lower working class,
from a catholic monastery to a countess. Again, it did not matter, whether or not brokers
knew that they were helping Jewish fugitives, but most did.
17/32
Figure 2 Brokerage chains in the network of the Segal family. Helpers found new helpers within their own social
networks, up to seven times in a row. Arrows indicate directionality of action. 34 nodes, 26 ties, manual layout.
Such chains were fleeting constructs and also provided support networks with (in times of
war) rare resources e.g. as red carton and boxwood for the creation of forged documents. Ties
between fugitives and helpers usually dissolved the moment they moved on to new helpers.
Together with the basic principle of any covert action, to only share an absolute minimum of
information, brokerage chains were relatively hard to detect.
Trust in resistant communities
Following the National Socialist seizure of power in January 1933 all known opposition
movements and with them, organisations which could have acted as roots for future
resistance were systematically disbanded. Leading activists were murdered or arrested and
18/32
members intimidated. In this climate, support networks based on pre-war organizational
structures were forced to organize themselves in small groups which could remain below the
radar of the Gestapo, spies and denunciators. As a consequence, these communities were
typically isolated, lacked resources and the means to obtain them.
All of the support networks we have studied were based on foci (Feld 1981) or social relays
(Ohlemacher 1996): they consisted of a core of individuals who opposed National Socialism,
did not face immediate persecution but witnessed it from close proximity, and shared longlasting trusted ties which typically preceded the regime. These clusters had survived mainly
because they did not previously committed any acts of resistance and remained passively in
niches of the National Socialist society. In our case studies, we find a circle of friends, a
church community or a group of neighbours. All of them began to support fugitives in the
1930s, by working around the flood of restrictions, help with emigration and by breaking the
social segregation. These early and still legal activities were an important prerequisite for the
subsequent illegal support of Jews in hiding.
The biggest support network within the German Reich emerged around Franz Kaufmann.
Despite his protests, Kaufmann was forced to retire in December 1935. Several attempts to
emigrate fail. With the outbreak of the war, Kaufmann makes unsuccessful attempts to join
the Wehrmacht and later the Red Cross. The fact that his wife was considered “Aryan” and
the birth of his daughter Angelika in 1940 meant that his legal status was “life in privileged
mixed marriage”; which, according to National Socialist race legislation, provided him with
relative safety from persecution. Around this time he became involved in a Berlin-based
initiative by Otto Grüber which provided legal aid to Jews who either wished to emigrate or
needed assistance in dealing with the anti-Jewish legislation. Kaufmann also became active in
the Bekennende Kirche (Confessing Church), a protestant church community which stood in
opposition to the Nazified Deutsche Christen. Within this group, Kaufmann met
“Nichtarische Christen” (Non-Aryan Christians), members of the protestant church who were
considered Jews or Half-Jews by National Socialist law. Kaufmann and his peers remained in
close contact with these members of the community throughout the 1930s. Members of the
Bekennende Kirche shared a critical view on the Deutsche Christen and National Socialism
but were not immune to Anti-semitism (Rudolph 2005:26). For the most part, their opposition
became apparent in theological arguments, frequently pondering the relationship between
Christian values and the duty to obey the state. The Dogmatische Arbeitsgemeinschaft
(Dogmatic circle) emerged as one bible study group in Berlin. Members of the Dogmatische
Arbeitsgemeinschaft had met once a week for two years and learned to know and trust each
other through previous forms of support and through their bible studies. Within this group,
over a period of two years, Kaufmann emerged as a leading figure who sought to translate the
theological insight that resistance against National Socialism was legitimate into action.
When one of the “non-Aryan” members of the community faced her deportation in August
1942, Kaufmann solicited help from individuals associated with this group and provided her
with a first forged passport. Within a year, Kaufmann became the centre of a rapidly growing
network, at least 400 individuals were involved in the provision of mainly forged documents
and food stamps but also other forms of help. Roughly half of the members of the
19/32
Dogmatische Arbeitsgemeinschaft supported Kaufmann in these efforts. This shows, first,
that even years of collaboration, shared norms, peer support and immediate exposure to the
consequences of persecution did not automatically lead to helping behaviour. Secondly, it
illustrates the unique environment in which the support network emerged: Many members of
the Bekennende Kirche would refuse to break the law themselves but still cover the
network’s actions. Others organized the collection of money and passports during church
services, brokered contacts to the Red Cross in Switzerland to obtain funds or recommended
Kaufmann to help-seeking fugitives. Within the Bekennende Kirche of Berlin, Kaufmann’s
network was safe from denunciation and opened up possibilities to contribute according to
one’s respective sense of justice.
Other networks have evolved in similar communities, always led by at least one charismatic
person who first initiated support and set an example. They too were held together by years
of trusted relationships and joint collaboration. In contrast to the dyadic and path-like
structures we have examined before, these clusters offered a far more stable source of support
with some degree of redundancy in case individual helpers were no longer available. Years of
trust and close cooperation in these tight-knit communities also meant that individuals felt an
obligation to the initiators and their peers to prolong their involvement.
The transition from legal means of support for Jews to violate contemporary laws followed
National Socialist legislation: Increasing measures of persecution demanded increasingly
antagonistic measures of support. By 1942 effective help had moved from breaking social
segregation and help in legal matters to forging passports, theft and black market trade.
Access to people with the required skills and contacts, however, was rare and in the case of a
bible study group also a moral dilemma. In the Kaufmann network, Christian helpers debated
at length whether their obligation to respect the law prevailed over their duty to help those in
need. They found relief in the works of the theologian Karl Barth whose works helped them
justify their actions. Still, disobeying the state order did not come easy for many, no matter
how inhumane its actions were. Gertrud Staewen, one of the helpers in Kaufmann’s network
wrote to her priest: „Hellmut, one year ago I was a relatively civil decent woman. Now I am
becoming a gangster. Relate this to the terrible work which unites me and Jacobs” (cited in
Flesch-Thebesius 2004:247, our translation). While Staewen continued to struggle to commit
to any unlawful actions, others, like Helene Jacobs whom Staewen mentions in the letter,
eventually committed to breaking the law but were always careful to draw a line between
them and “ordinary criminals” (Andreas-Friedrich 1996:227; Neiss et al. 1983). Contacts to
black marketeers were, however, essential for the creation of forged documents and provision
with food stamps. In order to become effective helpers, support networks needed to break
their isolation, overcome their moral inhibitions and gain access to these resources. Within
tight-knit clusters, personalities like Franz Kaufmann and Luise Meier, who organized escape
routes from Berlin across the German-Swiss border, acted as examples, provided concrete
opportunities to get involved and to some extent legitimized these actions.
20/32
Dynamic growth of support networks
Helpers like Kaufmann or Meier became known among fugitives and non-persecuted helpers
alike. Their continued success in providing extremely rare resources, such as forged
documents or illegal border crossings, made them and by association also their collaborators
trustworthy in the eyes of fugitives and other helpers. In turn they became highly attractive
focal points for those who sought help. The isolation of support networks was often
overcome with the help of fugitives themselves: In the case of the Kaufmann network, for
example, fugitives engaged in the necessary black market trading. In contrast to the rather
static and slowly evolving networks of the helpers, fugitives were forced to be mobile,
change their hideouts and helpers constantly. Despite the high risks and insecurities for
individuals, this meant a relatively large number of contacts to helpers.
Figure 3 Transitive trust chains which lead towards Luise Meier. Squares represent fugitives, diamonds nonpersecuted helpers, grey color indicates unknown National Socialist race classification. 44 nodes, 42 ties, manual
layout.
21/32
Fugitives exchanged information about capable helpers among themselves which again
traveled along chains through the small but highly mobile networks which they could
maintain during their lives in the underground. Figure 3 shows the paths alongside which
fugitives found Luise Meier. Fugitives could approach support networks with reasonable
confidence that the lead was not a trap by the Gestapo or a scam. Being known as capable
helpers among fugitives led to ever increasing numbers of requests for help. Helping more
fugitives therefore meant the network’s rapid expansion beyond control and to increased
activity. Therefore, the success of a support network also led to a significantly higher risk of
detection and carried the seed for its destruction. Within 18 months, the Kaufmann network
for example provided help for an estimated 200 fugitives and had around 400 actors who
were involved in these activities. Kaufmann and Luise Meier found that the steadily rising
number of requests for help made it impossible for them to limit the rapid expansion of the
networks they had initiated. More and more fugitives sought their help, more and more
outsiders needed to be brought in and as a consequence, the risk of detection rose steadily. In
the transcript of his Gestapo interrogation Kaufmann is quoted saying1:
„The more I gave in to the pity I felt, the greater were the requests I received which
explains the scale of my wrong-doings.“
(...)
„Thereby I became, without wanting to, a point of attraction and collection for Jewish
refugees. Their trust and hope made in that I could also help them emotionally made
it impossible to send them away.“ (our translation, cited in Rudolph 2005:68)
Luise Meier’s indictment documents state:
„In the meantime, word had come out among the illegal Jews in Berlin that Meier had
helped several Rassegenossen to escape. As a consequence a large number of Jews
unknown to her approached her for her aid in their escape.“ (our translation, cited in:
Anon 1944)
Figure 4 Arrests made by the Gestapo which led to the detection of the Kaufmann network within four days in late
July 1943
1
Transcripts like these are written in the language of the interrogators and represent their viewpoints. They can not be
trusted to represent literal quotes.
22/32
Even though brokerage chains were safer than long-lasting clusters given their extremely low
density, they were not fail proof: The same brokerage chains which facilitated the emergence
of far-reaching chains of trust within the Reich were reversed and thereby rolled up by the
Gestapo. During interrogations, helpers and fugitives were forced to reveal their respective
next contact and step by step the Gestapo moved towards the center of the support networks
as illustrated in Figure 4.
Discussion and Conclusion
The aim of this study was to shed light on the trust formation processes among fugitives and
helpers in one of the most extreme situations of social isolation and persecution in human
history, the persecution of Jewish people in Nazi Germany from 1939 to 1945. In a situation
of almost complete isolation and in an institutionally hostile environment, general
assumptions about trust formation between fugitive and helper do not apply. Trust did not
emerge in simple dyads between a trustor and a trustee. It was based on multi-layered
relationships which grew into triads, chains and clusters. Like fugitives, helpers too took (an
albeit smaller) risk when they established a support relationship. So trust had to
simultaneously be established from both sides and the roles of trustor and trustee became
blurred or even merged as helper - recipient relationships evolved into mutual dependency.
Selection and Influence
Steglich, Snijders, and Pearson (2010) name actor-initiated selection and social influence as
the two main factors on network formation and actor behaviour. For fugitives and helpers,
selection was essential. In a brokerage situation however, actors alternate between selection
and influence: We find selection (“whom can I trust with this?”) by those who initiate an act
of brokerage followed by influence (“it’s me who is asking” / “I helped already, would you
help as well?”) experienced by those who were approached as potential new helpers. This
experience of influence lead people to actively seek out potential new helpers. Within
resistant clusters selection brought like-minded people together who built relationships based
on their opposition to National Socialism and their bond based on faith or friendship.
Exposure to initiators and practices of peers convinced individuals to also commit themselves
to illegal acts of help and the associated risks
Perhaps surprisingly, fugitives and their helpers performed very similar tasks: Determine
whether somebody is trustworthy, obtain and exchange resources. While overall, fugitives
faced certain death if they were caught, their helpers had to fear punishment. Fugitives and
23/32
their helpers faced the same challenges to create trusted ties and used very similar strategies
to overcome them.
Relationships between fugitives and other fugitives, fugitives and their helpers and helpers
with other helpers vary considerably regarding their duration, balance between selfish- and
selflessness and the foundations of trust. One or combinations of several of the following
components constituted multi-layered relationships: A (reported) track record of previous acts
of help, old strong ties, material gains or other forms of exchange, known opposition to the
regime. Helpers who knew the true identity of fugitives could count on their desire to survive.
Table x lists four distinct scenarios and identifies reasons to trust and threats which applied.
We distinguish between threats within the control of helpers and fugitives such as betrayal
and non-conspiratorial behaviour and external threats such as Gestapo-traps, denunciation or
raids. Over time, the risk of intentional betrayal by one of the parties went down while the
risk of detection and traps (when encountering strangers) went up due to increased frequency.
A helper’s or a support network’s track records were crucial: A demonstration of previous
effective help and trustworthiness reduced the risks considerably. Track records were created
through acts of help.
Situation
Reasons to trust
Threats
Fugitives/helpers encounter a
(potential new) helper
First impressions/
Track record/
Material gains and
Exchanges/
Opposition
First impressions/
Fugitive’s desire to survive/
Betrayal/
Trap/
Non-conspiratorial
behaviour
Helper engages with fugitive
Fugitives/helpers engage in black
market trading
Helpers in support networks
First impressions/
Material gains/
Complicity
Track record/
Complicity/
Opposition
Trap/
Non-conspiratorial
behaviour
Betrayal/
Trap
Get caught with the
whole cluster/
Non-conspiratorial
behaviour
Table x
Finally, we are interested in what this extreme case can teach us for our general
understanding of trust in social networks in general and in covert networks in particular.
24/32
Fugitives first used a strategy that we would expect and the trust literature suggests. They
turned to people they knew and who would at least not betray them. Trust grew over time and
with mutual dependency. Erna Segal offers a good example of that strategy. As her
knowledge of the opposition of a helper to National Socialism grew, she formed her opinion,
had a friend’s recommendation and entered into an economic exchange relationship.
But soon these contacts were exhausted and fugitives had to turn to complete strangers in the
best of cases brokered by an actor with whom they had an old and established trust tie.
Therefore affection-based goodwill trust was more important than the more cognition-based
competence trust (Newell and Swan 2000). Both fugitives and their helpers had to first of all
find somebody who would not turn them in or exploit their vulnerability (goodwill). The need
to immediately get competent help (competence) was secondary. Only later, when it came to
highly specific means and resources, competence trust became relevant.
Shadow of the past and the future were not existent or very short. Fugitives and their helpers
often had no history together and were not likely to meet again in the future. Swift trust
(Meyerson, Weick, and Kramer 1996; Meyerson et al. 1996) which is often discussed as an
alternative to trust building processes based on frequent past interaction (shadow of the past)
is based on assumed shared institutional norms and similar experiences and therefore cannot
explain the trust formation in the situations under investigation here. It is therefore something
that we could call rapid trust both parties have to develop based on initial clues and first
impressions which, however, still involves a high risk.
Since risk was involved for both parties, one could also describe it as a relationship initiation
“dance” during which both sides gather crucial information about the other based on
reputation, subtle critical comments about the Nazi regime, hints at exchange relationships as
well as a subjective impression of sympathy towards the other party. Work in the area of
impression formation has shown the strong and enduring effect of the first impressions during
a first encounter on the future quality of a relationship (Jones 1990; Rosenthal and Ambady
1992). First impression theory explains how it was possible for fugitives to successfully
connect to helpers.
The situation in helper-helper relationships is somewhat different, since helpers could fall
back on relationships they have forged in other contexts where it was easier and less risky to
learn about somebody’s political orientation and character traits. In triads and in brokerage
chains with two non-persecuted helpers abuse of power was easier to be noticed and could
have negative repercussions for the helper. Within resistent clusters, helpers typically knew
each other for a long time and defectors would have to expect repercussions or at least
exclusion. In general we can observe that the following factors contributed to helpers not to
defect: peer expectations, loss of reputation, own political motivations, low personal risk,
solidarity with fugitives, personal bond beyond “the cause”, money and other remunerations
and solidarity with the group if being part of a resistant cluster.
Also for more fleeting helper-helper as well as fugitive-helper relationships, trusting
dispositions such as faith in humanity or trusting stance (Yu, Saleem, and Gonzalez 2014) are
very likely to not have played a role, since distrust had become the norm. It was rather the
ability to read and accurately assess the subtle clues that people gave about their position
25/32
towards the National Socialist system and the trusting beliefs (competence, benevolence,
integrity) (Yu et al. 2014) that people placed in another person that lead to a trust and support
relationship. After a support relationship was established trust formation followed again the
usual assumptions about familiarity between people and reinforcing mechanisms with
positive experiences based on concrete actions. The specific circumstances, however,
provided an additional and extremely valuable clue. If a new contact performed an act of
great risk to him or herself this person can be trusted – especially if that person’s gains are
small or even non-existent.
The key to understanding the evolution of these support networks and their structural features
is the development of transitive trust chains. Trust dyads morphed into transitive trust triads
through the brokering role most frequently by helpers but also by fugitives. Through that
mechanism it became possible to mobilize resources and build trust relationships with
additional helpers that were often complete strangers. These sequential triads formed long
and sparse chains that were relatively hard to detect by the Gestapo because contacts did
usually not last that long and people only knew three persons, the fugitive, the helper before
and the helper after oneself in the chain. Due to the relatively short contact time, they also
grew fast and uncontrollably into serendipitous networks in form of a chain without a center.
However, they therefore also had their limitations in terms of resource mobilization and
planned action capabilities and mainly served to hide people and literally keep them alive.
We have to contrast these very fast growing brokerage chains with clusters of helpers who
are based on many years of slowly built trust. Bonnie Erickson observed that “[t]he need for
trust then leads to the use of prior networks, which predate the secret society and set
constraints on it” (Erickson 1981:200). Almost all networks that emerged in the cases that we
investigated show a combination of old trusted ties such as those in the protestant church
community and newly created ties such as those between Erna Segal and the people she
approached. The stability of the former allowed division of labour but lacked the fresh
resources from the latter. Therefore, when brokerage chains connected to clusters, they
created a much larger availability of resources but also greatly increased the requests for help.
This combination causes the clusters themselves to become significantly more active and to
grow in size and become more centralized but in turn also be at a significantly higher risk for
detection.
This shows that it is possible to still build trust and social support relationships in situations
of extreme hostility and isolation with complete strangers which are neither based on prior
contact nor expectations of future contact even across social classes and ethnic groups. Even
though forging these kind of relationships was extremely risky in their beginning phases,
“successful” fugitives became skillful in using the first impressions to accurately predict
whether a potential helper would at least not turn them in. However, there were also many
instances, in which potential helpers turned out not to be trustworthy. Therefore, the less
risky and more promising strategy was to create transitive trust triads and have helpers act as
contact brokers in that process, i.e. the helper builds a trust tie to another helper and once that
is established introduces the fugitive who then builds a mutual trust relationship with the
second helper. That way, in an ideal case, long trust chains developed where any of the
helpers did not have to take great risks since a stay for example was only for a relatively short
26/32
time but that risks were actually distributed along these chains. Covert networks can use this
mechanism especially when it comes to hiding or transporting goods, money or people. There
is therefore evidence that building transitive trust triads into chain like structure is an
effective mechanism. This study is to our knowledge the first which can empirically show,
how this process works and what network structures evolve as a consequence.
Combining more dense networks in resistant communities with long chains of trust
relationships can mobilize a lot of resources and can partly overcome the isolation in a hostile
social environment. Our results show that in the end covert networks cannot overcome the
tension between efficiency and security (Morselli et al. 2007). While long brokerage chains
are relatively secure, they only present a very limited capacity for effective collective action.
As soon as this capacity is increased by combining these chains with denser structures of
networks in resistant communities that provide more opportunity for coordination and
planned action, covert networks run an increased risk of being discovered and destroyed
through the increase of redundant ties, more knowledge about other members of the network
and uncontrollable growth which gives state security forces an increased opportunity to learn
about and roll up a covert network.
As almost all historical studies, this study also faces limitations with regard to the
generalizability, reliability and validity of the data which we have discussed in the
introduction especially with regard to the relational nature of our data. In addition, we
investigated eight cases in one country in a specific historical situation which naturally limits
the external validity of our findings. However, we believe that especially the extreme
situation fugitives and helpers had to face in Nazi Germany between 1939 and 1945 presents
an excellent opportunity to gain important insights in the process of trust formation by means
of brokerage and the emergence of covert or dark networks across isolated clusters and
diverse social classes.
Bibliography
Andreas-Friedrich, Ruth. 1996. Der Schattenmann: Schauplatz Berlin.
Tagebuchaufzeichnungen 1938-1948. Berlin: Suhrkamp.
Anon. 1944. Anklageschrift Oberstaatsanwalt Beim Sondergericht Freiburg.
Strafsache Gegen Luise Meier. Gedenkstätte Deutscher Widerstand, Dossier Luise
Meier.
Baker, Wayne E. and Robert R. Faulkner. 1993. “The Social Organization of
Conspiracy: Illegal Networks in the Heavy Electrical Equipment Industry.” American
Sociological Review 58(6):837–60. Retrieved November 24, 2015
(http://www.jstor.org/stable/2095954).
Bakker, René M., Jörg Raab, and H. Brinton Milward. 2012. “A Preliminary Theory of
Dark Network Resilience.” Journal of policy analysis and management 31(1):33–62.
Retrieved October 26, 2015
27/32
(http://onlinelibrary.wiley.com/doi/10.1002/pam.20619/full).
Bearman, Peter S. and Kevin D. Everett. 1993. “The Structure of Social Protest,
1961–1983.” Social Networks 15(2):171–200. Retrieved November 25, 2015
(http://www.sciencedirect.com/science/article/pii/0378873393900045).
Benz, Wolfgang, ed. n.d. Solidarität Und Hilfe Für Juden Während Der NS-Zeit, Vol.
1-7, 1996-2004. Berlin: Metropol Verlag.
Broszat, Martin. 1986. “Zur Sozialgeschichte Des Deutschen Widerstands.”
Vierteljahrshefte für Zeitgeschichte 34(3):293–309.
Broszat, Martin and Elke Fröhlich. 1987. Alltag und Widerstand: Bayern im
Nationalsozialismus. Piper.
Calloway, Michael and Joseph P. Morrissey. 1993. “Accuracy and Reliability in SelfReported Data in Inter-Organizational Networks.” Social Networks 15(4):377–98.
Crossley, Nick, Gemma Edwards, Ellen Harries, and Rachel Stevenson. 2012.
“Covert Social Movement Networks and the Secrecy-Efficiency Trade off: The Case
of the UK Suffragettes (1906–1914).” Social Networks 34(4):634–44. Retrieved
October 22, 2015
(http://www.sciencedirect.com/science/article/pii/S0378873312000469).
Düring, Marten and Ulrich Eumann. 2013. “Historische Netzwerkforschung. Ein
Neuer Ansatz in Den Geschichtswissenschaften.” Geschichte und Gesellschaft
39:369–90.
Erickson, Bonnie H. 1981. “Secret Societies and Social Structure.” Social Forces
60(1):188–210. Retrieved October 15, 2015
(http://sf.oxfordjournals.org/content/60/1/188).
Erickson, Bonnie H. 1997. “Social Networks and History: A Review Essay.” Historical
Methods 30(3):149–57.
Feld, Scott L. 1981. “The Focused Organization of Social Ties.” American Journal of
Sociology 86(5):1015–35. Retrieved February 11, 2016
(http://www.jstor.org/stable/2778746).
Festinger, Leon. 1957. Theory of Cognitive Dissonance. New impression. Stanford
University Press.
Flesch-Thebesius, Marlies. 2004. Zu Den Außenseitern Gestellt: Die Geschichte Der
Gertrud Staewen 1894 - 1987. Berlin: Wichern-Verl.
Fogelman, Eva. 1995. Conscience and Courage: Rescuers of Jews During the
Holocaust. Anchor.
Franzosi, Roberto. 1998. “Narrative as Data: Linguistic and Statistical Tools for the
Quantitative Study of Historical Events.” International Review of Social History
43(Supplement S6):81–104. Retrieved November 25, 2015
(http://journals.cambridge.org/article_S002085900011510YES).
Franzosi, Roberto. 1999. “The Return of The Actor. Interaction Networks Among
Social Actors During Periods of High Mobilization (Italy, 1919-1922).” Mobilization:
An International Quarterly 4(2):131–49. Retrieved November 25, 2015
(http://mobilizationjournal.org/doi/abs/10.17813/maiq.4.2.480g12700535663l).
Gellately, Robert. 1996. “Denunciations in Twentieth-Century Germany: Aspects of
Self-Policing in the Third Reich and the German Democratic Republic.” The Journal
28/32
of Modern History 68(4):931–67. Retrieved November 24, 2015
(http://www.jstor.org/stable/2946725).
Grabowski, Jan. 2008. Rescue for Money: Paid Helpers in Poland, 1939 - 1945.
Jerusalem: Yad Vashem.
Gudehus, Christian. 2015. “Verfolgten Helfen. Heuristiken und Perspektiven (am
Beispiel des Holocaust).” Online Encyclopedia of Mass Violence. Retrieved
December 7, 2015 (http://massviolence.org/Verfolgten-Helfen-Heuristikenund?decoupe_recherche=gudehus).
Jones, Edward E. 1990. Interpersonal Perception. New York, NY, US: W H
Freeman/Times Books/ Henry Holt & Co.
Kosmala, Beate. 2004. “Zwischen Ahnen Und Wissen.” Pp. 135–59 in Die
Deportation der Juden aus Deutschland, edited by C. Dieckmann. Göttingen:
Wallstein-Verl.
Kosmala, Beate. 2007. “Stille Helden.” Aus Politik und Zeitgeschichte. Beilage zur
Wochenzeitung Das Parlament 14-15. Retrieved
(http://www.bpb.de/themen/OOFHRO.html).
Kosmala, Beate. 2011. “Stille Helden.” Pp. 34–48 in Rettungswiderstand, edited by
A. Lustiger. Berlin: Suhrkamp.
Krebs, Valdis. 2002. “Mapping Networks of Terrorist Cells.” CONNECTIONS
24(3):43–52.
Kupfer, Annett and Frank Nestmann. 2015. “Soziale Unterstützung – Social Support
Eine Zentrale Funktion Sozialer Netzwerke.” Pp. 151–79 in Knoten und Kanten III
Soziale Netzwerkanalyse in Geschichts- und Politikforschung, edited by M. Gamper,
L. Reschke, and M. Düring. Berlin, Boston: De Gruyter.
Lampe, Klaus von and Per Ole Johansen. 2004. “Organized Crime and Trust: On the
Conceptualization and Empirical Relevance of Trust in the Context of Criminal
Networks.” Global Crime 6(2):159–84. Retrieved November 25, 2015
(http://dx.doi.org/10.1080/17440570500096734).
Latané, Bibb and John M. Darley. 1970. The Unresponsive Bystander: Why Doesn’t
He Help? New York: Appleton-Century Crofts.
Latané, Bibb and Judith Rodin. 1969. “A Lady in Distress: Inhibiting Effects of
Friends and Strangers on Bystander Intervention.” Journal of Experimental Social
Psychology 5(2):189–202. Retrieved
(http://www.sciencedirect.com/science/article/pii/0022103169900468).
Lemercier, Claire. 2015. “Formal Network Methods in History: Why and How?” Pp.
281–304 in Social networks, political institutions, and rural societies, Rural history in
Europe, edited by G. Fertig. Turnhout: Brepols.
London, Perry. 1970. “The Rescuers: Motivational Hypotheses about Christians Who
Saved Jews from the Nazis.” Pp. 241–50 in Altruism and Helping Behavior, edited by
J. R. Macaulay and L. Berkowitz. New York [etc.]: Academic Press.
Lustiger, Arno, ed. 2011. Rettungswiderstand: Das Buch von Den Judenrettern Im
Nationalsozialismus. 1. ed. Berlin: Suhrkamp.
Marsden, Peter V. 2005. “Recent Developments in Network Measurement.” Pp. 8–30
in Models and Methods in Social Network Analysis, edited by P. J. Carrington, J.
29/32
Scott, and S. Wasserman. New York: Cambridge University Press.
McKnight, D. Harrison and Norman L. Chervany. 2001. “Trust and Distrust
Definitions: One Bite at a Time.” Pp. 27–54 in Trust in Cyber-societies, Lecture
Notes in Computer Science, edited by R. Falcone, M. Singh, and Y.-H. Tan. Springer
Berlin Heidelberg. Retrieved November 25, 2015
(http://link.springer.com/chapter/10.1007/3-540-45547-7_3).
Merriam-Webster. n.d. “Definition of HELP.” Retrieved February 10, 2016a
(http://www.merriam-webster.com/dictionary/help).
Merriam-Webster. n.d. “Definition of RESCUE.” Merriam-Webster. Retrieved
February 10, 2016b (http://www.merriam-webster.com/dictionary/rescue).
Meyerson, Debra, Karl E. Weick, and Roderick M. Kramer. 1996. “Trust in
Organizations: Frontiers of Theory and Research.” Pp. 166–96 in Swift Trust and
Temporary Groups. Thousand Oaks, CA: SAGE Publications, Inc.
Miles, Matthew B., A. M. Huberman, and Johnny Saldaña. 2014. Qualitative Data
Analysis: A Methods Sourcebook. 3rd ed. Thousand Oaks, Califorinia: SAGE
Publications, Inc.
Monroe, Kristen, Michael Barton, and Ute Klingemann. 1990. “Altruism and the
Theory of Rational Action: Rescuers of Jews in Nazi Europe.” Ethics 101(1):103–22.
Monroe, Kristen R. 2011. Ethics in an Age of Terror and Genocide: Identity and
Moral Choice. Princeton and N.J: Princeton University Press.
Monroe, Kristen Renwick. 2008. “Cracking the Code of Genocide: The Moral
Psychology of Rescuers, Bystanders, and Nazis during the Holocaust.” Political
Psychology 29(5):699–736.
Moore, Bob. 2010. Survivors: Jewish Self-Help and Rescue in Nazi-Occupied
Western Europe. Oxford, New York: Oxford University Press.
Morselli, Carlo, Cynthia Giguère, and Katia Petit. 2007. “The Efficiency/security
Trade-off in Criminal Networks.” Social Networks 29(1):143–53. Retrieved August
15, 2014 (http://linkinghub.elsevier.com/retrieve/pii/S0378873306000268).
Neiss, Marion, Birgit Wagner, Irina Wandrey, and Martina Voigt. 1983. Interview with
Helene Jacobs. Berlin: Gedenkstätte Deutscher Widerstand.
Neurath, Wolfgang. 2008. “Neue Perspektiven Für Die Geschichtswissenschaft
Durch Soziale Netzwerkanalyse (SNA).” Österreichische Zeitschrift für
Geschichtswissenschaft 19 (4):140–53.
Newell, Sue and Jacky Swan. 2000. “Trust and Inter-Organizational Networking.”
Human Relations - HUM RELAT 53(10):1287–1328.
Ohlemacher, Thomas. 1996. “Bridging People and Protest: Social Relays of Protest
Groups against Low-Flying Military Jets in West Germany.” Social Problems
43(2):197–218. Retrieved February 11, 2016 (http://www.jstor.org/stable/3096998).
Oliner, Samuel P. and Pearl M. Oliner. 1992. The Altruistic Personality: Rescuers of
Jews in Nazi Europe.
Osa, Maryjane. 2003. Solidarity and Contention: Networks of Polish Opposition.
Minneapolis: University of Minnesota Press.
Paldiel, Mordecai. 2004. “The Deeds and Behavior of Those We Term Righteous
Among the Nations.” Yad Vashem. Retrieved November 25, 2015
30/32
(https://www.yadvashem.org/yv/en/righteous/pdf/resources/paldiel.pdf).
Podolny, Joel M. and James N. Baron. 1997. “Resources and Relationships: Social
Networks and Mobility in the Workplace.” American Sociological Review 62(5):673–
93. Retrieved November 24, 2015 (http://www.jstor.org/stable/2657354).
Reitmayer, Morten and Christian Marx. 2010. “Netzwerkansätze in Der
Geschichtswissenschaft.” Pp. 869–80 in Handbuch Netzwerkforschung, edited by C.
Stegbauer and R. Häußling. Wiesbaden: VS.
Rosenthal, Naomi, Meryl Fingrutd, Michele Ethier, Roberta Karant, and David
McDonald. 1985. “Social Movements and Network Analysis: A Case Study of
Nineteenth-Century Women’s Reform in New York State.” The American Journal of
Sociology 90(5):1022–54. Retrieved (http://www.jstor.org/stable/2780088).
Rosenthal, Robert and Nalini Ambady. 1992. “Thin Slices of Expressive Behavior as
Predictors of Interpersonal Consequences: A Meta-Analysis.” Psychological Bulletin
111(2):256–74.
Rudolph, Katrin. 2005. Hilfe Beim Sprung Ins Nichts: Franz Kaufmann Und Die
Rettung von Juden Und “Nichtarischen” Christen. Berlin: Metropol-Verlag.
Schieb, Barbara. 1989. Interview with Konrad Latte. Berlin: Gedenkstätte Deutscher
Widerstand.
Segal, Erna. 1956. Autobiographical Report Erna Segal. Gedenkstätte Deutscher
Widerstand.
Semelin, Jacques. 2011. Resisting Genocide: The Multiple Forms of Rescue. edited
by C. Andrieu and S. Gensburger. C Hurst & Co Publishers Ltd.
Simmel, Georg. 1906. “The Sociology of Secrecy and of Secret Societies.” American
Journal of Sociology 11(4):441–98. Retrieved October 23, 2015
(http://www.jstor.org/stable/2762562).
Steglich, Christian, Tom AB Snijders, and Michael Pearson. 2010. “Dynamic
Networks and Behavior: Separating Selection from Influence.” Sociological
methodology 40(1):329–93. Retrieved February 24, 2016
(http://onlinelibrary.wiley.com/doi/10.1111/j.1467-9531.2010.01225.x/full).
Tausendfreund, Doris. 2006. Erzwungener Verrat: “Jüdische Greifer” Im Dienst Der
Gestapo 1943 - 1945: Techn. Univ., Diss.–Berlin, 2005. Berlin: Metropol Verl.
Varese, Federico and Meir Yaish. 2000. “The Importance of Being Asked the Rescue
of Jews in Nazi Europe.” Rationality and Society 12(3):307–34. Retrieved August 16,
2014 (http://rss.sagepub.com/content/12/3/307).
Wasserman, Stanley and Katherine Faust. 1994. Social Network Analysis: Methods
and Applications. New York: Cambridge University Press.
Wellman, Barry. 1979. “The Community Question: The Intimate Networks of East
Yorkers.” American Journal of Sociology 84(5):1201–31. Retrieved November 24,
2015 (http://www.jstor.org/stable/2778222).
Wellman, Barry and Scot Wortley. 1990. “Different Strokes from Different Folks:
Community Ties and Social Support.” American Journal of Sociology 96(3):558–88.
Wetherell, Charles. 1998. “Historical Social Network Analysis.” International Review
of Social History 43(Supplement 6):125–44.
Wolfson, Manfred. 1971. “Der Widerstand Gegen Hitler: Soziologische Skizze über
31/32
Retter von Juden in Deutschland.” Aus Politik und Zeitgeschichte. Beilage zur
Wochenzeitung “Das Parlament” (15):32–39.
Wolfson, Manfred. 1975. “Zum Widerstand Gegen Hitler: Umriß Eines
Gruppenportraits Deutscher Retter von Juden.” Pp. 391–407 in Tradition und
Neubeginn, edited by J. Hütter. Köln: Heymanns.
Yu, Michael, Muniba Saleem, and Cleotilde Gonzalez. 2014. “Developing Trust: First
Impressions and Experience.” Journal of Economic Psychology 43:16–29. Retrieved
November 27, 2015
(http://www.sciencedirect.com/science/article/pii/S0167487014000233).
Ansatz die von EL vorgeschlagenen tasks zu systematisieren. Von mir aus zu löschen, es
sei denn Du siehst noch etwas Interssantes drin:
Needs
At times required
skills and resources
Associated form
of trust
Associated risks
Accommodation during
the night and
Occupation during the
day
Space
Grown
Rapid
Blind
Reputation
False premises
Remuneration
In private homes, denunciation by
neighbours.
Detection by authorities when in
public places.
Hotels are required to report names of
guests.
Frequent raids by police.
Food/food stamps
Black market
trading
Forgery
False identity
Money
Limited availability due to wartime
rationing, need to reach out to others.
Acquire forged food stamps on black
market risky due to police raids.
Helpers: purchase more food than
normal.
Forged documents
Forgery
Black market
trading
Access to passports
etc.
Money
German bureaucracy set to identify
Jews.
Hard to find a forger and required
resources, need to reach out to others.
High costs motivate false promises
by would-be forgers.
Documents are identified as fake by
authorities.
Travelling
False identity
Cover story
Frequent police raids and regular
checks.
Young men are considered potential
deserters and require additional
documents.
Emotional support
Empathy
Social networks destroyed, hard to
32/32
establish contact.
Finding new helpers
People judgment
Cover story
Money
High risk to encounter a spy or
denunciator.