Agent-based models (ABMs) are computational models that simulatebehavior of individual agents in order to study emergent phenomena atthe level of the community. Depending on the application, agents mayrepresent humans, institutions, microorganisms, and so forth. Theagents’ actions are based on autonomous decision-making andother behavioral traits, implemented through formal rules. Bysimulating decentralized local interactions among agents, as well asinteractions between agents and their environment, ABMs enable us toobserve complex population-level phenomena in a controlled and gradualmanner.

This entry focuses on the applications of agent-based modeling in thephilosophy of science, specifically within the realm of formal socialepistemology of science. The questions examined through these modelsare typically of direct relevance to philosophical discussionsconcerning social aspects of scientific inquiry. Yet, many of thesequestions are not easily addressed using other methods since theyconcern complex dynamics of social phenomena. After providing a briefbackground on the origins of agent-based modeling in philosophy ofscience (Section 1), the entry introduces the method and its applications as follows. Webegin by surveying the central research questions that have beenexplored using ABMs, aiming to showwhy this method has beenof interest in philosophy of science (Section 2). Since each research question can be approached through variousmodeling frameworks, we next delve into some of the common frameworksutilized in philosophy of science to showhow ABMs tacklephilosophical problems (Section 3). Subsequently, we revisit the previously surveyed questions andexamine the insights gained through ABMs, addressingwhat hasbeen found to answer each question (Section 4). Finally, we turn to the epistemology of agent-based modeling and theunderlying epistemic function of ABMs (Section 5). Given the often highly idealized nature of ABMs, we examinewhichepistemic roles support the models’ capacity to engage withphilosophical issues, whether for exploratory or explanatory goals.The entry concludes by offering an outlook on future researchdirections in the field (Section 6).

Since the literature on agent-based modeling of science is vast andgrowing, it is impossible to give an exhaustive survey of modelsdeveloped on this topic. Instead, this entry aims to provide asystematic overview by focusing on paradigmatic examples of ABMsdeveloped in philosophy of science, with an eye to their relevancebeyond the confines of formal social epistemology.

• 1. Origins
• 2. Central Research Questions
  • 2.1 Theoretical diversity and the incentive structure of science
  • 2.2 Theoretical diversity and the communication structure of science
  • 2.3 Cognitive diversity
  • 2.4 Social diversity
  • 2.5 Peer-disagreement in science
  • 2.6 Scientific polarization
  • 2.7 Scientific collaboration
  • 2.8 Summing up
• 3. Common Modeling Frameworks
  • 3.1 Epistemic landscape models
    • 3.1.1 Application to the problem of cognitive diversity
    • 3.1.2 Applications to network epistemology
    • 3.1.3 Other applications
  • 3.2 Bandit models
  • 3.3 Bounded confidence models of opinion dynamics
  • 3.4 Evolutionary game-theoretic models of bargaining
  • 3.5 Summing up
• 4. Central Results
  • 4.1 Theoretical diversity and the incentive structure of science
  • 4.2 Theoretical diversity and the communication structure of science
    • 4.2.1 The “Zollman effect”
    • 4.2.2 The spread of disinformation
  • 4.3 Cognitive diversity
  • 4.4 Social diversity
  • 4.5 Peer-disagreement in science
  • 4.6 Scientific polarization
  • 4.7 Scientific collaboration
  • 4.8 Additional themes
• 5. Epistemology of Agent-Based Modeling
  • 5.1 The challenge of abstract modeling
  • 5.2 Verification and validation: from exploratory to explanatory functions of ABMs
  • 5.3 The “family of models perspective”
• 6. Conclusion and Outlook
• Bibliography
• Academic Tools
• Other Internet Resources
• Related Entries

1. Origins

The method of agent-based modeling was originally developed in the1970s and ’80s with models of social segregation (Schelling1971; Sakoda 1971; see also Hegselmann 2017) and cooperation (Axelrod& Hamilton 1981) in social sciences, and under the name of“individual-based modeling” in ecology (for an overviewsee Grimm & Railsback 2005). Following this tradition, ABMs drewthe interest of scholars studying social aspects of scientificinquiry. By representing scientists as agents equipped with rules forreasoning and decision-making, agent-based modeling could be used tostudy the social dynamics of scientific research. As a result, ABMs ofscience have been developed across various disciplines that includescience in their subject domain: from sociology of science,organizational sciences, cultural evolution theory, theinterdisciplinary field of meta-science (or “science ofscience”), to social epistemology and philosophy of science.While ABMs developed in philosophy of science often tackle themes thatare similar or related to those examined by ABMs of science in otherdomains, they are motivated by philosophical questions—issuesembedded in the broader literature in philosophy of science. Theirintroduction was influenced by several parallel research lines:analytical modeling in philosophy of science, computational modelingin related philosophical domains, and agent-based modeling in socialsciences. In the following, we take a brief look at each of theseprecursors.

One of the central ideas behind the development of formal socialepistemology of science is succinctly expressed by Philip Kitcher inhisThe Advancement of Science:

The general problem of social epistemology, as I conceive it, is toidentify the properties of epistemically well-designed social systems,that is, to specify the conditions under which a group of individuals,operating according to various rules for modifying their individualpractices, succeed, through their interactions, in generating aprogressive sequence of consensus practices. (Kitcher 1993: 303)

Such a perspective on social epistemology of science highlighted theneed for a better understanding of the relationship between individualand group inquiry. Following the tradition of formal modeling ineconomics, philosophers introduced analytical models to study tensionsinherent to this relationship, such as the tension between individualand group rationality. In this way, they sought to answer thequestion: how can individual scientists, who may be driven bynon-epistemic incentives, jointly form a community that achievesepistemic goals? Most prominently, Goldman and Shaked (1991) developeda model that examines the relationship between the goal of promotingone’s professional success and the promotion oftruth-acquisition, whereas Kitcher (1990, 1993) proposed a model ofthe division of cognitive labor, showing that a community consistingof scientists driven by non-epistemic interests may achieve an optimaldistribution of research efforts. This work was followed by a numberof other contributions (e.g., Zamora Bonilla 1999; Strevens 2003).Analytic models developed in this tradition endorsed economicapproaches to the study of science, rooted in the idea of a“generous invisible hand”, according to which individualsinteracting in a given community can bring about consequences that arebeneficial for the goals of the community without necessarily aimingat those consequences (Mäki 2005).

Around the same time, computational methods entered the philosophicalstudy of rational deliberation and cooperation in the context of gametheory (Skyrms 1990, 1996; Grim, Mar, & St. Denis 1998), theoryevaluation in philosophy of science (Thagard 1988) and the study ofopinion dynamics in social epistemology (Hegselmann & Krause 2002,2005; Deffuant, Amblard, Weisbuch, & Faure 2002). Computationalmodels introduced in this literature included ABMs: for instance, acellular automata model of the Prisoner’s Dilemma, or modelsexamining how opinions change within a group of agents.

Agent-based modeling was first applied to the study of science insociology of science, with the model developed by Nigel Gilbert (1997)(cf. Payette 2011). Gilbert’s ABM aimed at reproducingregularities that had previously been identified in quantitativesociological research on indicators of scientific growth (such as thegrowth rate of publications and the distribution of citations perpaper). The model followed an already established tradition ofsimulations of artificial societies in social sciences (cf. Epstein& Axtell 1996). In contrast to abstract and highly-idealizedmodels developed in other social sciences (such as economics andarchaeology), ABMs in sociology of science tended towards anintegration of simulations and empirical studies used for theirvalidation (cf. Gilbert & Troitzsch 2005).

Soon after, ABMs were introduced to the philosophy of science throughpioneering works by Zollman (2007), Weisberg and Muldoon (2009), Grim(2009), Douven (2010)—to mention some of the most prominentexamples. In contrast to ABMs developed in sociology of science, theseABMs followed the tradition of abstract and highly-idealized modeling.Similar to analytical models, they were introduced to study howvarious properties of individual scientists—such as theirreasoning, decision-making, actions and relations—bring aboutphenomena characterizing the scientific community—such as asuccess or a failure to acquire knowledge. By representing inquiry inan abstract and idealized way, they facilitated insights into therelationship between some aspects of individual inquiry and its impacton the community while abstracting away from numerous factors thatoccur in actual scientific practice. But in contrast to analyticalmodels, ABMs proved to be suitable for scenarios often too complex foranalytical approaches. These scenarios include heterogeneousscientific communities, with individual scientists differing in theirbeliefs, research heuristics, social networks, goals of inquiry, andso forth. Each of these properties can change over time, depending onthe agents’ local interactions. In this way ABMs can show howcertain features characterizing individual inquiry suffice to generatepopulation-level phenomena under a variety of initial conditions.Indeed, the introduction of ABMs to philosophy of science largelyfollowed the central idea of generative social science: to explain theemergence of a macroscopic regularity we need to show howdecentralized local interactions of heterogeneous autonomous agentscan generate it. As Joshua Epstein summed it up: “If youdidn’t grow it, you didn’t explain its emergence”(2006).

2. Central Research Questions

ABMs of science typically model the impact of certain aspects ofindividual inquiry on some measure of epistemic performance of thescientific community. This section surveys some of the centralresearch questions investigated in this way. Its aim is to explain whyABMs were introduced to study philosophical questions, and how theirintroduction relates to the broader literature in philosophy ofscience.

2.1 Theoretical diversity and the incentive structure of science

How does a community of scientists make sure to hedge its bets onfruitful lines of inquiry, instead of only pursuing suboptimal ones?Answering this question is a matter of coordination and organizationof cognitive labor which can generate an optimal diversity of pursuedtheories. The importance of a synchronous pursuit of a plurality oftheories in a given domain has long been recognized in thephilosophical literature (Mill 1859; Kuhn 1977; Feyerabend 1975;Kitcher 1993; Longino 2002; Chang 2012). But how does a scientificcommunity achieve an optimal distribution of research efforts? Whichfactors influence scientists to divide and coordinate their labor in away that stimulates theoretical diversity? In short, how istheoretical diversity achieved and maintained?

One way to address this question is by examining how differentincentives of scientists impact their division of labor. To see therelevance of this question, consider a community of scientists all ofwhom are driven by the same epistemic incentives. As Kitcher (1990)argued, in such a community everyone might end up pursuing the same,initially most promising line of inquiry, resulting in little to nodiversity. Traditionally, philosophers of science tried to addressthis worry by arguing that a diversity in theory choice may resultfrom diverse methodological approaches (Feyerabend 1975), diverseapplications of epistemic values (Kuhn 1977), or from differentcognitive attitudes towards theories, such as acceptance andpursuit-worthiness (Laudan 1977). Kitcher, however, wondered whethernon-epistemic incentives, such as fame and fortune—usuallythought of as interfering with the epistemic goals ofscience—might actually be beneficial for the community byencouraging scientists to deviate from dominant lines of inquiry.

The idea that scientists are rewarded for their achievements throughcredit, which impacts their research choices, had previously beenrecognized by Merton (1973) and Hull (1978, 1988). For example, ascientist may receive recognition for being the first to make adiscovery (known as the “priority rule”), which mayincentivize a specific approach to research. Yet, such non-epistemicincentives could also fail to promote an optimal kind of diversity.For instance, they may result in too much research being spent onfutile hypotheses and/or in too few scientists investigating the besttheories. Moreover, an incentive structure will have undesirableeffects if rewards are misallocated to scientists that arewell-networked rather than assigned to those who are actually first tomake discoveries, or if credit-driven science lowers the quality ofscientific output. This raises the question: which incentivestructures promote an optimal division of labor, without havingepistemically or morally harmful effects?

ABMs provide an apt ground for studying these issues: by modelingindividual scientists as driven by specific incentives, we can examinetheir division of labor and the resulting communal inquiry. We willlook at the models studying these issues inSection 4.1.

2.2 Theoretical diversity and the communication structure of science

Another way to study theoretical diversity is by focusing on thecommunication structure of scientific communities. In this case we areinterested in how the information flow among scientists impacts theirdistribution of research across different rival hypotheses. Theimportance of scientific interaction for the production of scientificknowledge has traditionally been emphasized in social epistemology(Goldman 1999). But how exactly does the structure of communicationimpact scientists’ generation of knowledge? Are scientistsbetter off communicating within strongly connected social networks, orrather within less connected ones, and under which conditions ofinquiry? These and related questions belong to the field ofnetwork epistemology, which studies the impact ofcommunication networks on the process of knowledge acquisition.Network epistemology has its origin in economics, sociology andorganizational sciences (e.g., Granovetter 1973; Burt 1992; Jackson& Wolinsky 1996; Bala & Goyal 1998) and it was first combinedwith agent-based modeling in the philosophical literature by Zollman(2007) (see also Zollman 2013).

Simulations of scientific interaction originated in the idea thatdifferent communication networks among scientists, characterized byvarying degrees of connectedness (seeFigure 1), may have a different impact on the balance between“exploration” and “exploitation” of scientificideas. Suppose a scientist is trying to find an optimal treatment fora certain disease, since the existing one is insufficiently effective.On the one hand, she could pursue the currently dominant hypothesisconcerning the causes of the disease, hoping that it will eventuallylead to better results. On the other hand, she could explore novelideas hoping to have a breakthrough leading to a more successful curefor the disease. The scientist thus faces a trade-off betweenexploitation as the use of existing ideas and exploration as thesearch of new possibilities, long studied in theories of formallearning and in organizational sciences (March 1991). The informationflow among scientists could impact this trade-off in the followingway: if an initially misleading idea is shared too quickly throughoutthe community, scientists may lock-in on researching it, prematurelyabandoning the search for better solutions. Alternatively, if theinformation flow is slow and sparse, important insights gained by somescientists, which could lead to an optimal solution, may remainundetected by the rest of the community for a lengthy period of time.ABMs were introduced to investigate whether and in which circumstancesthe information flow could have either of these effects. For instance,if scientists are assumed to be rational agents, could a tightlyconnected community end up exploring too little and miss out onsignificant lines of inquiry?

Besides studying communities consisting of “epistemicallyuncompromised” scientists—that is, agents whose inquiryand actions are directed at discovering and promoting thetruth—similar questions can be posed about communities in whichepistemic interests have been overridden. For instance, the impact ofindustrial interest groups on science may lead to biased or deceptivepractices, which may sway the scientific community away from itsepistemic goals (Holman & Elliott 2018). While recentphilosophical discussions on this problem have largely focused on therole of non-epistemic values in science (Douglas 2009; Holman &Wilholt 2022; Bueter 2022; Peters 2021), ABMs were introduced toexamine how epistemically pernicious strategies can impact the processof inquiry, as well as to identify interventions that can be used tomitigate their harmful effects.

In addition to the problem of theoretical diversity, networkepistemology has been applied to a number of other themes, such asoptimal forms of collaboration, factors leading to scientificpolarization, effects of conformity on the collective inquiry, effectsof demographic diversity, the position of minorities, optimalregulations of dual-use research, argumentation dynamics, and soforth. We will look at the models studying theoretical diversity andthe communication structure of science inSection 4.2 andSection 4.6.

(a)

(b)

(c)

Figure 1: Three types of idealizedcommunication networks, representing an increasing degree ofconnectedness: (a) a cycle, (b) a wheel, and (c) a complete graph. Thenodes in each graph stand for scientists, while edges between thenodes stand for information channels between two scientists.

2.3 Cognitive diversity

A diversity of cognitive features of individuals can be beneficial invarious problem-solving situations, including business andpolicy-making (Page 2017). But how does the diversity of cognitivefeatures of scientists, including their background beliefs, reasoningstyles, research preferences, heuristics and strategies impact theinquiry of the scientific community? In philosophy of science, thisissue gained traction with Kuhn’s distinction between normal andrevolutionary science (Kuhn 1962), suggesting that differentpropensities may push scientists towards one type of research ratherthan another (see also Hull 1988). This raises the question: how doesthe distribution of risk-seeking, maverick scientists and risk-averseones impact the inquiry of the community? Put more generally: do someways of dividing labor across different research heuristics result ina more successful collective inquiry than others?

By equipping agents with different cognitive features we can use ABMsto represent different cognitively diverse (or uniform) populations,and to study their impact on some measure of success of the communalinquiry. We will look at the models studying these issues inSection 4.3.

2.4 Social diversity

A scientific community is socially diverse when its members endorsedifferent non-epistemic values, such as moral and political ones, orwhen they have different social locations, such as gender, race andother aspects of demography (Rolin 2019). The importance of socialdiversity has long been emphasized in feminist epistemology, both forethical and epistemic reasons. For instance, many have pointed outthat social diversity is an important catalyst for cognitivediversity, which in turn is vital for the diversity of perspectives,and therefore for scientific objectivity (Longino 1990, 2002; Haraway1989; Wylie 1992, 2002; Grasswick 2011; Anderson 1995; for adiscussion on different notions of diversity in the context ofscientific inquiry see Steel et al. 2018).

Moreover, in the field of social psychology and organizationalsciences, it has been argued that social diversity is epistemicallybeneficial even if it doesn’t promote cognitive diversity.Instead, it may counteract epistemically pernicious tendencies ofhomogeneous groups, such as unwarranted trust in each other’stestimony or unwillingness to share dissenting opinions (for anoverview of the literature see Page 2017; Fazelpour & Steel 2022).While these hypotheses have received support in virtue of empiricalstudies, ABMs have proved a complementary testing ground, allowing foran investigation of minimal sets of conditions which need to hold forsocial diversity to be epistemically beneficial.

Another problem tackled by means of ABMs concerns factors that canundermine social diversity or disadvantage members of minorities inscience. For instance, how does one’s minority status impactone’s position in a collaborative environment, given the normsof collaboration that can emerge in scientific communities? Or howdoes one’s social identity impact the uptake of their ideas? Wewill look at the models studying these issues inSection 4.4 andSection 4.7.

2.5 Peer-disagreement in science

Scientific disagreements are commonly considered vital for scientificprogress (Kuhn 1977; Longino 2002; Solomon 2006). They typically gohand in hand with theoretical diversity (seeSection 2.1 and2.2) and stimulate critical interaction among scientists, important forthe achievement of scientific objectivity. Nevertheless, an inadequateresponse to disagreements may lead to premature rejections of fruitfulinquiries, to fragmentation of scientific domains and hence toconsequences that are counterproductive for the progress of science.This raises the question: how should scientists respond todisagreements with their peers, to lower the chance of a hinderedinquiry? Which epistemic and methodological norms should they followin such situations?

This issue has been discussed in the context of the more generaldebate on peer-disagreement in social epistemology. The problem ofpeer disagreement concerns the question: what is an adequate doxasticattitude towardsp, upon recognizing that one’s peerdisagrees onp? Should one follow a “ConciliatoryNorm”, demanding, for instance, to lower the confidence inp, split the difference by taking the middle ground between theopponent’s belief and one’s own on the issue, or tosuspend one’s judgment onp? Or should one rather followa “Steadfast Norm” demanding to stick to one’s gunsand keep the same belief with the same confidence as beforeencountering a disagreeing peer? (For initial arguments in favor ofthe Conciliatory Norm see, e.g., Elga 2007, Christensen 2010, Feldman2006; for arguments in favor of the Steadfast Norm see, e.g., De Cruz& De Smedt 2013, Kelp & Douven 2012; for reasons why norms arecontext-dependent see, e.g., Kelly 2010 [2011], Konigsberg 2013,Christensen 2010, Douven 2010; for a recent critical review of thedebate as applied to scientific practice see Longino 2022.)

Similarly, in case of scientific disagreements and controversies wecan ask: should a scientist who is involved in a peer disagreementstrive towards weakening her stance by means of a conciliatory norm,or should she remain steadfast? What makes this issue particularlychallenging in the context of scientific inquiry is that we are notonly interested in the epistemic question of an adequate doxasticresponse to a disagreement, but also in the methodological (orinquisitive) question of how the norms impact the success ofcollective inquiry as a process. In particular, if scientistsencounter multiple disagreements throughout their research on acertain topic, will their collective inquiry benefit more fromindividuals adopting conciliatory attitudes or steadfast ones? ABMsnaturally lend themselves as a method for investigating these issues:by modeling scientists as guided by different normative responses to adisagreement, we can study the impact of the norms on the communalinquiry. We will look at the models studying these issues inSection 4.5.

2.6 Scientific polarization

Closely related to the issue of scientific disagreements is theproblem of scientific polarization. While scientific controversiestypically resolve over time, they may include periods of polarization,with different parts of the community maintaining mutually conflictingattitudes even after an extensive debate. But how and why doespolarization emerge? Do scientific communities polarize only ifscientists are too dogmatic or biased towards their viewpoints, or canpolarization emerge even among rational agents?

Following a range of formal models in social and political sciencesaddressing a similar issue in society at large (for a review seeBramson et al. 2017), ABMs have been used to examine the emergence ofpolarization in the context of science. What makes agent-basedmodeling particularly apt for this task is not only that we can modeldifferent aspects of individual inquiry that may contribute to theemergence of polarization (such as different background beliefs,different communication networks, different trust relationships, andso on), but we can also observe the formation of polarized states,their duration (as stable or temporary states throughout the inquiry)and their features (such as the distribution of scientists across theopposing views). We will look at the models studying these issues inSection 4.6.

2.7 Scientific collaboration

As acquiring and analyzing scientific evidence can be highlyresource-demanding for any individual scientist, scientificcollaboration is a widespread form of group inquiry. Of course, thisis not the only reason why scientists collaborate: incentives leadingto collaborations range from epistemic ones (such as increasing thequality of research) to non-epistemic ones (such as striving forrecognition). This raises the question: when is collaboratingbeneficial, and which challenges may occur in collaborative research?Inspired by these questions, philosophers of science have investigatedwhy collaborations are beneficial (Wray 2002), which challenges theypose on epistemic trust and accountability (Kukla 2012; Wagenknecht2015), what kind of knowledge emerges through collaborative research(such as collective beliefs or acceptances, (M. Gilbert 2000; Wray2007; Andersen 2010), which values are at stake in collaborations(Rolin 2015), what an optimal structure of collaborations is(Perović, Radovanović, Sikimić, & Berber 2016), andso on.

ABMs of collaboration were introduced to study the above and relatedquestions, focusing on how collaborating can impact inquiry. Whilecollaborations can indeed be beneficial, determining conditions underwhich they are such is not straightforward. For instance, depending onhow scientists engage in collaborations, minorities in the communitymay end up disadvantaged. We will look at the models studying theseissues inSection 4.7.

2.8 Summing up

Beside the above themes, ABMs have been applied to the study ofnumerous other topics in philosophy of science: from the allocation ofresearch funding, testimonial norms, strategic behavior of scientists,all the way to different procedures for theory-choice (seeSection 4.8 where we list models studying additional themes). Moreover, one ABMcan simultaneously address multiple questions (for example, socialdiversity and scientific collaboration are often inquiredtogether).

To study the questions presented in this section, philosophers haveutilized different representational frameworks. Even if models areaimed at the same research question, they are often based on differentmodeling approaches. For instance, individual scientists may berepresented as Bayesian reasoners, as agents with limited memory, asagents searching for peaks on an epistemic landscape, as agents thatform their opinions by averaging information they receive from otherscientists and from their own inquiry, or as agents that are equippedwith argumentative reasoning. Similarly, the process of evidencegathering may be represented in terms of pulls from probabilitydistributions, as foraging on an epistemic or an argumentationlandscape, or as receiving signals from others and from the world. Wenow look into some of the common modeling frameworks employed in thestudy of the above questions.

3. Common Modeling Frameworks

When developing a model examining certain aspects of scientificinquiry, one first has to decide on a number of relevantrepresentational assumptions, such as:

1. How to represent the process ofinquiry andevidencegathering?
2. What doagents in the simulation stand for (e.g.,individual scientists, research groups, scientific labs, etc.)?
3. What are theunits of appraisal in scientific inquiry(e.g., hypotheses, theories, research programs, methods, etc.)?
4. How do scientistsreason and evaluate their units ofappraisal?
5. How do scientistsexchange information?

Similarly, if we wish to model a scenario in which scientists bargainabout the division of tasks or resources, we will have to decide howto represent theirinteractions andrewards they getout of them. These modeling choices are guided by the researchquestion the model aims to tackle, as well as the epistemic aim of themodel.

The majority of ABMs developed in philosophy of science are built assimple and highly idealized models. The simpler a model is, the easierit is to understand and analyze mechanisms behind the results ofsimulations (we return to this issue inSection 5). In this section, we delve into several common modeling frameworksthat have been used in this way. Each framework offers a differenttake on the above representational choices and has served as the basisfor a variety of ABMs. Particular models will not be discussed justyet—we leave this forSection 4.

3.1 Epistemic landscape models

Modeling the process of inquiry as an exploration of an epistemiclandscape draws its roots from models of fitness landscapes inevolutionary biology, first introduced by Sewall Wright (1932). Byrepresenting a genotype as a point on a multidimensional landscape,where the “height” of the landscape corresponds to itsfitness, the model has been used to study evolutionary paths ofpopulations.

The idea ofepistemic landscapes entered philosophy ofscience with the work of Weisberg and Muldoon (2009) and Grim (2009).In this reinterpretation of the model, the landscape represents aresearch topic, consisting of multiple projects or multiplehypotheses. A research topic can be understood either in a narrowsense (e.g., the study of treatments for a certain disease) or in abroader sense (e.g., the field of astrophysics). A point on thelandscape stands for a certain hypothesis or a specific approach toinvestigating the topic. Approaches can vary in terms of differentbackground assumptions, methods of inquiry, research questions, etc.Accordingly, the landscape can be modeled in terms of \(n\)dimensions, where \(n-1\) dimensions represent different aspects ofscientific approaches, while the \(n^{th}\) dimension (visualized asthe “height” of the landscape) stands for some measure ofepistemic value an agent gets by pursuing the corresponding approach.For instance, in case of a three-dimensional landscape, thexandy coordinates form a two-dimensional disciplinary matrix inwhich approaches are situated, whilez coordinate measurestheir “epistemic significance” (seeFigure 2). The latter can be understood in line with Kitcher’s idea thatsignificant approaches are those that enable the conceptual andexplanatory progress of science (Kitcher 1993: 95).

ABMs of science have utilized two-dimensional and three-dimensionallandscapes, as well as the generalized framework of NK-landscapes, inwhich the number of dimensions and the ruggedness of the landscape areparameters of the model.^[1] Scientists are modeled as agents who explore the landscape, trying tofind its peak(s), that is, the epistemically most significant points.The framework allows for different ways of measuring the success andefficiency of inquiry: in terms of the success of the community indiscovering the peak(s) of the landscape (rather than getting stuck inlocal maxima), the success in discovering any of the areas of non-zerovicinity, the time required for such discoveries, and so forth.

Figure 2: Two representations ofWeisberg and Muldoon’s epistemic landscape: a three-dimensionalrepresentation on the left and a two-dimensional representation of thesame landscape on the right, where the height of the landscape isrepresented by different shades of gray: the lighter the shade, themore significant the point on the landscape (adapted from Weisberg& Muldoon 2009).

3.1.1 Application to the problem of cognitive diversity

The framework of epistemic landscapes has been applied to a variety ofresearch questions in philosophy of science. Weisberg and Muldoon(2009) introduced it to examine the impact of cognitive diversity onthe performance of scientific communities. What makes the epistemiclandscape framework particularly attractive for the study of cognitivediversity and the division of labor is its capacity to representvarious research strategies (as different heuristics of exploring thelandscape), as well as a coordinated distribution of research efforts(cf. Pöyhönen 2017).

Weisberg and Muldoon’s ABM employs a three-dimensional landscape(seeFigure 2), built on a discrete toroidal grid, with two peaks representing thehighest points of epistemic significance. To study cognitivediversity, the model examines three research strategies, implementedas three types of agents:

• The “controls” who aim to find a higher point on thelandscape than their current location, while ignoring the explorationof other agents.
• The “followers” who aim to find already exploredapproaches in their direct neighborhood, which have a highersignificance than their current location.
• The “mavericks” who also aim to find points of highersignificance given previously explored points, but rather thanfollowing in the footsteps of other scientists, they prioritize thediscovery of new, so far unvisited, points.

The control strategy represents individual learning that disregardssocial information, while the follower and the maverick strategiesrepresent different ways of taking the latter into account. The modelexamines how homogeneous populations of each type of explorers, orheterogeneous populations consisting of diverse groups of explorers,impact the efficiency of the community in discovering the highestpoints on the landscape, and in covering all points of non-zerosignificance.

Following Weisberg and Muldoon’s contribution, the framework ofepistemic landscapes became highly influential, resulting in variousrefinements and extensions of the model. For instance, Alexander,Himmelreich, and Thompson (2015) introduced the “swarm”strategy, which describes scientists who can with certain probabilityidentify points of higher significance in their surrounding and adjusttheir approach so that it is similar, but different from approachespursued by others who are close to them. Thoma (2015) introduced“explorer” and “extractor” strategies, theformer of which describes a scientist seeking approaches verydifferent from those pursued by others, while the latter (similar toAlexander and colleagues’ swarm researcher) seeks approachesthat are similar to, yet distinct from those pursued by others.Moreover, Fernández Pinto and Fernández Pinto (2018)examined alternative rules for the follower strategy. Finally,Pöyhönen (2017) introduced a “dynamic”landscape, such that the exploration of patches “depletes”their significance.

3.1.2 Applications to network epistemology

Another application domain of the epistemic landscape framework is thecommunication structure of science and its impact on communal inquiry.To study these issues Grim and Singer (Grim 2009; Grim, Singer,Fisher, et al. 2013) developed ABMs employing a two-dimensionallandscape, where points on thex-axis represent alternativehypotheses for a given domain of phenomena, while they-axisindicates the “epistemic goodness” of each hypothesis orthe epistemic payoff an agents gets by pursuing it. Depending on howwell the best hypothesis is hidden, the shape of the epistemic terrainwill represent a more or less “fiendish” research problem(seeFigure 3). For instance, if the best hypothesis is a narrow peak in thelandscape (as inFigure 3c) finding it will resemble a search for a needle in a haystack andtherefore a fiendish research problem.^[2] An inquiry is considered successful if any scientist (eventuallyfollowed by the rest of the community) discovers the highest peak onthe landscape, while it is unsuccessful if the community converges ona local maximum. The model is used to study how different socialnetworks with varying degrees of connectedness (such as those inFigure 1), and different epistemic landscapes with different degrees offiendishness impact the success of inquiry.

(a)

(b)

(c)

Figure 3: Epistemic landscapes with anincreasing degree of fiendishness (adapted from Grim, Singer, Fisher,et al. 2013: 443)

3.1.3 Other applications

Beside the above themes, the framework of epistemic landscapes hasappeared in many other variants. For instance, De Langhe (2014a)proposed an ABM aimed at studying different notions of scientificprogress, which makes use of a “moving epistemiclandscape”: a landscape in which the significance of approachescan change as a result of exploration of other approaches. Balietti,Mäs, and Helbing (2015) developed an epistemic landscape model tostudy the relationship between fragmentation in scientific communitiesand scientific progress, in which agents explore a two-dimensionallandscape, representing a space of possible answers to a certainresearch question, with the correct answer located at the center ofthe landscape. Currie and Avin (2019) proposed a reinterpretation ofthe three-dimensional landscape aimed at studying the diversity ofscientific methods, in which thex andy axis stand fordifferent investigative techniques or methods of acquiring evidence,and thez axis for the “sharpness” of the obtainedevidence. Sharpness here refers to the relationship between researchresults and a hypothesis, where the more the evidence increasesone’s credence in the hypothesis, the sharper it is. The modelalso represents the “independence” of evidence as thedistance between two points on the landscape according to thexandy coordinates: the further apart two methods are, the lessoverlapping their background theories are, and the more independentthe evidence obtained by them is.

Finally, epistemic landscape modeling has inspired related landscapemodels. The argumentation-based model of scientific inquiry by Borg,Frey, Šešelja, and Straßer (2018, 2019) is onesuch example. The model, inspired by abstract argumentation frameworks(Dung 1995), employs an “argumentation landscape”. Thelandscape is composed of “argumentation trees” thatrepresent rival research programs (seeFigure 4). Each program is a rooted tree, with nodes as arguments and edges as a“discovery relation”, representing a path agents take tomove from one argument to another. Arguments in one theory may attackarguments in another theory. In contrast to epistemic landscapes,argumentation landscapes aim to capture the dialectic dimension ofinquiry where some points on the landscape, assumed to be acceptablearguments, may subsequently be rejected as undefendable. This allowsfor an explicit representation of false positives (accepting a falsehypothesis) as an argument on the landscape an agent accepts withoutknowing there exists an attack on it, and false negatives (rejecting atrue hypothesis) as an argument an agent rejects, without knowing thatit can be defended. The model has been used to study how argumentativedynamics among scientists, pursuing rival research programs, impactstheir efficiency in acquiring knowledge under various conditions ofinquiry (such as different social networks, different degrees ofcautiousness in decision making, and so on).

Figure 4: Argumentation-based ABM byBorg, Frey, Šešelja, and Straßer (2019) employingan argumentative landscape. The landscape represents two rivalresearch programs (RP), with darker shaded nodes standing forarguments that have been explored by agents and are thus visible tothem; brighter shaded nodes stand for arguments that are not visibleto agents. The biggest node in each RP is the root argument, fromwhich agents start their exploration via the discovery relation,connecting arguments within one RP. Arrows stand for attacks from anargument in one RP to an argument in another RP.

3.2 Bandit models

In the previous section we saw how epistemic landscape models havebeen used to represent inquiries involving mutually compatibleresearch projects (as in Weisberg and Muldoon’s model) as wellas inquiries involving rival hypotheses (as in Grim and Singer’smodel). Another framework used to study the latter scenario is basedon “bandit problems”.

The name of bandit models comes from the term “one-armedbandit”—a colloquial name for a slot machine. Multi-armedbandit problems, introduced in statistics (Robbins 1952; Berry &Fristedt 1985) and studied in economics (Bala & Goyal 1998; Bolton& Harris 1999), are decision-making problems that concern thefollowing kind of scenario: suppose a gambler is about to play severalslot machines. Each machine gives a random payoff according to aprobability distribution unknown to the gambler. To determine whichmachine will give a higher reward in the long run, the gambler has toexperiment by pulling arms of the machines. This will allow her tolearn from the obtained results. But which strategy should she use?For instance, should she alternate between the machines for the firstcouple of pulls, and then play only the machine that has given thehighest reward during the initial test run? Or should she rather havea lengthy test phase before deciding which machine is better? While inthe first case she might startexploiting her currentknowledge before she has sufficientlyexplored, in the secondcase she might explore for too long. In this way the gambler faces atrade-off between exploitation (playing the machine that has so fargiven the best payoff) and exploration (continued testing of differentmachines). The challenge is thus to come up with the strategy thatprovides an optimal balance between exploration and exploitation, soas to maximize one’s total winnings.

As we have seen above (seeSection 2.2), the exploration/exploitation trade-off may also occur in the contextof scientific research. Whether scientists are attempting to determineif a novel treatment for a disease is superior to an existing one, orselecting between two novel methods of evidence gathering with varyingsuccess rates, they may encounter the exploration/exploitationtrade-off. In other words, they may have to decide when to ceaseexploring alternatives and begin exploiting the one that appears mostsuitable for the task.

ABMs based on bandit problems were first introduced to philosophy ofscience by Zollman (2007, 2010). A bandit ABM usually looks asfollows. In analogy to slot machines, each scientific theory (orhypothesis, or method) is represented as having a designatedprobability of success. Scientists are typically modeled as“myopic” agents in the sense that they always pursue atheory they believe to give a higher payoff. They gather evidence fora theory by making a random draw from a probability distribution.Subsequently, they update their beliefs in a Bayesian way, on thebasis of results of their own research (the gathered evidence) as wellas results of neighboring scientists in their social network (such asthose inFigure 1). Scientists are thus not modeled as passive observers of evidence forall the available theories, but rather as agents who activelydetermine the type of evidence they gather by choosing which theory topursue. The inquiry of the community is considered successful if, forexample, scientists reach a consensus on the better of the twotheories.

Bandit models of this kind build on the analytical framework developedby economists (Bala & Goyal 1998) who examined the relationshipbetween different communication networks in infinite populations andthe process of social learning. Applied to the context of science,this variation of bandit problems concerns the puzzle mentioned aboveinSection 2.2: assuming a scientific community is trying to determine which of thetwo available hypotheses offers a better treatment for a givendisease, could the structure of the information flow among thescientists impact their chances of converging on the betterhypothesis? By applying Bala and Goyal’s framework to thecontext of science and scenarios involving finite populations, Zollmaninitiated the field of network epistemology (see Zollman 2013), whichstudies the impact of social networks on the process of knowledgeacquisition.

Beside the aim of examining network effects on the formation ofscientific consensus, bandit models of scientific interaction havebeen applied to a number of other topics, such as the impact ofpreferential attachment on social learning (Alexander 2013), bias anddeception in science (Holman & Bruner 2015, 2017; Weatherall,O’Connor, & Bruner 2020), optimal forms of collaboration(Zollman 2017), factors leading to scientific polarization(O’Connor & Weatherall 2018; Weatherall & O’Connor2021b), effects of conformity (Weatherall & O’Connor 2021a),effects of demographic diversity (Fazelpour & Steel 2022),regulations of dual-use research (Wagner & Herington 2021), sociallearning in which a dominant group ignores or devalues testimony froma marginalized group (Wu 2023), or disagreements on the diagnosticityof evidence for a certain hypothesis (Michelini & Osorio et al.forthcoming).

3.3 Bounded confidence models of opinion dynamics

As we have seen above, the performance of a scientific community issometimes assessed in terms of its success in achieving consensus onthe true hypothesis. The question how consensus is formed and whichfactors benefit or hinder its emergence is also studied within thetheme ofopinion dynamics in epistemic communities. Models ofopinion dynamics aim at investigating how opinions form and change ina group of agents who adjust their views (or beliefs) over a number ofrounds or time-steps, resulting in the formation ofconsensus,polarization, or aplurality ofviews. A modeling framework that has been particularly influential inthis context originates from the work of Hegselmann and Krause (2002,2005, 2006), drawing its roots from analytical models of consensusformation (French 1956; DeGroot 1974; Lehrer & Wagner 1981).^[3]

The basic model functions as follows. At the start of the simulation(that is, at time \(t=0\)) each agent \(x_i\) is assigned an opinionon a certain issue, expressed by a real number \(x_i(0) \in {(0,1]}\).Agents then exchange their opinions (or beliefs) with others andadjust them by taking the average of those beliefs that are “nottoo far away” from their own. These are opinions that fallwithin a “confidence interval” of size \(\epsilon\), whichis a parameter of the model. In this way, agents have boundedconfidence in opinions of others. By iterating this process, the modelsimulates the social dynamics of the opinion formation (seeFigure 5a).

Applied to the context of scientific inquiry, scientists are usuallyrepresented as truth seeking agents who are trying to determine thevalue of a certain parameter \(\tau\) for which they only know that itlies in the interval \((0,1]\) (Hegselmann & Krause 2006). At thestart of the simulation each agent is assigned a random initialbelief. As the model runs, agents adjust their beliefs by receiving a(noisy) signal about \(\tau\) and by learning opinions of others whofall within their confidence interval. For instance, an agent’sbelief can be updated in terms of the weighted average ofothers’ opinions and the signal from the world.^[4] Such dynamics represent scientists as agents who are able to generateevidence that points in the direction of \(\tau\), though theirupdates are also influenced by their prior beliefs and the beliefs oftheir peers (seeFigure 5b &Figure 5c).

(a)

(b)

(c)

Figure 5: Examples of runs of Hegselmannand Krause’s model, with thex-axis representing timesteps in the simulation and they-axis opinions of 100 agents.The change of each agent’s opinion is represented with a coloredline. The value of \(\tau\) (the assumed position of the truth) isrepresented as a black dotted line. Figure (a) shows opinion dynamicsin a community without any truth-seeking agents, resulting in aplurality of diverging views. Figure (b) shows the opinion dynamics ina community in which all agents are truth-seekers, which achieves aconsensus close to the truth. Figure (c) shows the opinion dynamics ina community in which only half of the population are truth-seekers,which also achieves a consensus close to the truth (adapted fromHegselmann & Krause 2006; for other parameters used in thesimulations see the original article).

An early application of the bounded confidence model was the problemof the division of cognitive labor (Hegselmann & Krause 2006).Since agents can be modeled as updating their beliefs only on thebasis of social information (that is, by averaging over opinions ofothers who fall in one’s confidence interval) or on the basis ofboth social information and the signal from the world, the model canbe used to study the opinion dynamics in a community in which noteveryone is a “truth-seeker”. The model then examines whatkind of division of labor between truth-seekers and agents who onlyreceive social information allows the community to converge to thetruth.

Subsequent applications of the framework in social epistemology andphilosophy of science focused on the study of scientific disagreementsand the question: what is the impact of different norms guidingdisagreeing scientists on the efficiency of collective inquiry (Douven2010; De Langhe 2013)? If agents update their beliefs both in view oftheir own research and in view of other agents’ opinions theyrepresent scientists that follow a Conciliatory Norm by“splitting the difference” between their own view and theviews of their peers (see aboveSection 2.5). In contrast, if they update their beliefs only in view of their ownresearch they represent scientists that follow a Steadfast Norm. Otherapplication themes of this framework include the impact of noisy dataon opinion dynamics (Douven & Riegler 2010), opinion dynamicsconcerning complex belief states, such as beliefs about scientifictheories (Riegler & Douven 2009), updating via an inference to thebest explanation in a social setting (Douven & Wenmackers 2017),deceit and spread of disinformation (Douven & Hegselmann 2021),network effects and theoretical diversity in scientific communities(Douven & Hegselmann 2022), and so forth.

3.4 Evolutionary game-theoretic models of bargaining

Game theory studies situations in which the outcome of one’saction depends not only on one’s choice of an action, but alsoon actions of others. A “game” in this sense is a model ofa strategic interaction between agents (“players”), eachof whom has a set of available actions or strategies. Each combinationof the players’ strategic responses has a designated outcome ora “payoff”. In contrast to traditional game theory, whichfocuses on agents’ rational decision-making aimed at maximizingtheir payoff in one-off interactions, the evolutionary approachfocuses on repeated interactions in a population. A game is assumed tobe played over and over again by players who are randomly drawn from alarge population. While agents start with a certain strategicbehavior, they learn and gradually adjust their responses according tospecific rules called “dynamics” (for example, byimitating other players or by considering their own pastinteractions). As a result, successful strategies will diffuse acrossthe community. In this way, evolutionary game-theoretic models can beused to explain how a distribution of strategies across the populationchanges over time as an outcome of long-term population-levelprocesses. While the standard approach to game theory has primarilyfocused on combinations of players’ strategies that lead to a“stable” state, such as the Nash equilibrium—a statein which no player can improve their payoff by unilaterally changingtheir strategy—the evolutionary approach has been used to studyhow equilibria emerge in a community (see the entries ongame theory and onevolutionary game theory).

Evolutionary game theory was originally introduced in biology(Lewontin 1961; Maynard Smith 1982). It subsequently gained interestof social scientists and philosophers as a tool for studying culturalevolution, that is, for investigations into how beliefs and normschange over time (Axelrod 1984; Skyrms 1996). The models can beimplemented using a mathematical treatment based on differentialequations, or using agent-based modeling. While the former approachemploys certain idealizations, such as an infinite population size orperfect mixing of populations, ABMs were introduced to study scenariosin which such assumptions are relaxed (see, e.g., Adami, Schossau,& Hintze 2016).

Applications of evolutionary game theory to social epistemology ofscience were especially inspired by models of bargaining, studying howdifferent bargaining norms emerge from local interactions ofindividuals (Skyrms 1996; Axtell, Epstein, & Young 2001). Theframework was introduced to the study of epistemic communities byO’Connor and Bruner (2019), building up on Bruner’s (2019)model of cultural interactions.^[5]

The basic idea of bargaining models is as follows: agents bargain overshares of available resources, where their demands and expectationsabout others’ demands evolve endogenously in view of theirprevious interactions. Applied to the context of science, bargainingconcerns not only explicit negotiations over financial resources, butalso situations in which scientists need to agree how to divide theirworkload in joint projects (O’Connor & Bruner 2019). Forexample, if two scientists are working on a joint paper or if they areorganizing a conference together they will have to agree on how muchtime and effort each of them will devote to it. The norms determiningsuch a division of labor may not be fair. For example, if scientistA puts much less effort into the project than scientistB, but they both get the same recognition for its success,B will be disadvantaged. Similarly, if they agree onAbeing the first author in a collaborative paper, whileB endsup working much more on it, the outcome will again be unfair. Suchnorms may become entrenched in the scientific community, especially inthe context of interactions between members of majority and minoritygroups in academia. But how do such norms emerge? Are biases favoringmembers of certain groups over others necessary for the emergence ofsuch discriminatory patterns, or can they become entrenched due toother, perhaps more surprising factors?

Evolutionary game-theoretic models have been used to study these andrelated questions. Bargaining is represented as a strategicinteraction between two agents, each of whom makes a demand concerningthe issue at hand (for instance, a certain amount of workload, theorder of authors’ names in a joint publication, and so on).Depending on each agent’s demand, each gets a certain payoff.For instance, supposeA andB wish to organize aconference and they start by negotiating who will cover which tasks.If they both make a high demand in the sense that each is willing toput only a minimal effort into the project while expecting the otherto cover the rest of the tasks, they will fail to organize the event.IfB, on the other hand, makes a low demand (by taking on alarger portion of the work) whileA makes a high demand, theywill be able to organize the conference, though the division of laborwill be unfair (assuming they both get equal credit for successfullyrealizing the project).

The game, originating in the work of John Nash (1950), is called the“Nash demand game” (or the “mini-Nash demandgame”). Each player in the game makes their demand (Low, Med orHigh). If the demands do not jointly exceed a given resource, eachplayer gets what they asked for. If they do exceed the availableresource, no-one gets anything. In the example above, High can beinterpreted as demanding to work less than the other on theorganization of the conference, or demanding the first authorship in ajoint paper while putting relatively lower effort into it. Similarly,Low corresponds to the willingness to take on a larger portion of thework (relative to the order of authors in case of the joint paper),while Med corresponds to demanding a fair distribution of labor.Table 1 displays the payoffs in such a game. Any combination of demands thatgives a joint payoff of 10 is a Nash equilibrium, which means thateither player’s strategy is the best response to the otherplayer’s one. While a Nash equilibrium may correspond to a fairdistribution of resources (if both players demand Med), it may alsocorrespond to an unfair one (if one player demands Low and the otherone High). This raises the question: which equilibrium will thecommunity achieve if agents learn from their previous interactions? Inparticular, if the individuals are divided into sub-groups (which maybe of different sizes), where their membership can be identified bymeans of markers visible to other agents, they can develop strategiesconditional on the group membership of their co-players. Whichequilibrium state will such a community evolve to? To study suchquestions, evolutionary models employ rules or dynamics that determinehow players update their strategies and how the distribution ofstrategies across the community changes over time.^[6]

The modeling framework based on bargaining has been used to studynorms in scientific collaborations and inequalities that may emergethrough them. For example, O’Connor and Bruner (2019)investigate the emergence of discriminatory norms in academicinteractions between minority and majority members. Rubin andO’Connor (2018) study the emergence of discriminatory patternsand their effects on diversity in collaboration networks, whileVentura (2023) examines the impact of the structure of collaborativenetworks on the emergence of discriminatory norms even if there are novisible markers of an agent’s membership to a sub-group.Moreover, Klein, Marx, and Scheller (2020) use a similar framework tostudy the relationship between rationality and inequality, that is,the success of different strategies for bargaining (such as maximizingexpected utility) and their impact on the emergence of inequality.

3.5 Summing up

Besides the above frameworks, numerous additional approaches have beenused to build simulations in social epistemology and philosophy ofscience. Some prominent frameworks not mentioned above include theBayesian framework “Laputa” developed by Angere(2010—in Other Internet Resources) and Olsson (2011), aimed atstudying social networks of information and trust, the model ofargumentation dynamics by Betz (2013), or the influential framework byHong and Page (2004) utilized in the study of cognitive diversity (formore on these frameworks see the entry oncomputational philosophy). Another evolutionary framework used in philosophy of science wasproposed by Smaldino and McElreath (2016). The model represents ascientific community as a population consisting of scientific labsthat employ culturally transmitted methodological practices, where thepractices undergo natural selection from one generation of scientiststo the next, and it has been employed, for example, to study theselection of conservative and risk-taking science (O’Connor2019). In the next section we take a look at some of the centralresults obtained by ABMs in philosophy of science, based on the aboveand some other frameworks.

4. Central Results

To provide an overview of the main findings obtained by means of ABMsin philosophy of science, we will revisit the research questionsdiscussed inSection 2 and look at how they have been answered through specific models.

4.1 Theoretical diversity and the incentive structure of science

Before we survey ABMs that study the incentive structure of science,we first look into the results of some analytical models whichinspired the development of simulations. To get a more precise grip onhow individual incentives shape epistemic coordination and theoreticaldiversity, philosophers introduced formal analytical models, inspiredby research in economics. One of the central results from this body ofliterature is that the optimal distribution of labor can be achievedwhen scientists act according to their self-interest rather thanfollowing epistemic ends (e.g., Kitcher 1990; Brock & Durlauf1999; Strevens 2003). More precisely, the models show that if weassume scientists aim at maximizing rewards from making discoveries,they will succeed to optimally distribute their research efforts ifthey take into account the probability of success of each researchline and how many other scientists currently pursue it. Assuming thatall scientists evaluate theories in the same way, their interest infame and fortune, rather than epistemic goals alone, will lead some ofthem to select avenues that initially appear less promising.^[7]

ABMs were introduced to address similar questions, but assuming morecomplex scenarios. For example, Muldoon and Weisberg (2011) developedan epistemic landscape model (seeSection 3.1) to examine the robustness of Kitcher’s and Strevens’results under the assumption that scientists have varying access toinformation about the pursued research projects in their community andthe future success of those projects. Their results indicate that oncescientists have limited information about what others are working on,or about the degree to which projects are likely to succeed, theirself-organized division of labor fails to be optimal. Another exampleis the model by De Langhe & Greiff (2010) who generalizeKitcher’s model to a situation with multiple epistemic standardsdetermining the background assumptions of scientists, acceptablemethods, acceptable puzzles, and so on. The simulations show that oncescientific practice is modeled as based on multiple standards, theincentive to compete fails to provide an optimal division of labor.

A closely related question concerns the “priorityrule”—a norm that allocates credit for a scientificdiscovery to the first one to make it (Strevens 2003, 2011)—andits impact on the division of labor. While Kitcher’s andStrevens’s models suggested that the priority rule incentivizesthe optimal distribution of scientists across rival research programs,a range of formal models, including ABMs, were developed to reexaminethese results and shed additional light on this norm. For instance,Rubin and Schneider (2021) examine what happens if credit is assignedby individuals, rather than by the scientific community as a whole, asin Strevens’ model. They further suppose that news aboutsimultaneous discoveries by two scientists spreads through a networkedcommunity. The simulations show that more connected scientists aremore likely to gain credit than the less connected ones, which may, onthe one hand, disadvantage minority members in the community, and onthe other hand, undermine the role of the priority rule as anincentive resulting in the optimal division of labor. Besides thequestion of how the priority rule impacts the division of labor, ABMshave also been used to study other effects of the priority rule. Forexample, Tiokhin, Yan, and Morgan (2021) develop an evolutionary ABMshowing that the priority rule leads the scientific community toevolve towards research based on smaller sample sizes, which in turnreduces the reliability of published findings.

The impact of incentives on the division of labor in science has alsobeen analyzed in terms of incentives to “exploit” existingprojects in contrast to incentives to “explore” novelideas. For instance, De Langhe (2014b) developed a generalized versionof Kitcher and Strevens’ model in which agents achieve theoptimal division of labor by weighing up relative costs and benefitsof exploiting available theories and exploring new ones. Within theframework of bandit models and network epistemology (seeSection 3.2), Kummerfeld and Zollman (2016) proposed an ABM that examines ascenario in which scientists face two rival hypotheses, one of whichis better though the agents don’t know which one. While agentsalways choose to pursue (or exploit) a hypothesis that seems morepromising, they may also occasionally research (and thereby explore)the alternative one. The simulations show that if the community isleft to be self-organized in the sense that each scientist explores tothe extent that they consider individually optimal, agents will beincentivized to leave exploration to others. As a result, scientistswill fail to develop a sufficiently high incentive for exploring novelideas, that is, an incentive which would be optimal from theperspective of their community at large.

4.2 Theoretical diversity and the communication structure of science

4.2.1 The “Zollman effect”

The study of theoretical diversity in terms of network epistemologyled to a novel hypothesis: that the communication structure of ascientific community can promote or hinder the emergence oftheoretical diversity and thereby impact the division of cognitivelabor. The idea was first demonstrated by bandit models developed byZollman (2007, 2010; seeSection 3.2) and came to be known as the “Zollman effect” (Rosenstock,Bruner, & O’Connor 2017). ABMs by Grim (2009) Grim, Singer,Fisher, and colleagues (2013), and Angere & Olsson (2017) producedsimilar findings based on different modeling frameworks.^[8] These models show that in highly connected communities earlyerroneous results may spread quickly among scientists, leading them toinvestigate a sub-optimal line of inquiry. As a result, scientists mayprematurely abandon the exploration of different hypotheses, andinstead exploit the inferior ones. In light of these findings, Zollman(2010) emphasized that for an inquiry to be successful it needs aproperty of “transient diversity”: a process in which acommunity engages in a parallel exploration of different theories,which lasts sufficiently long to prevent a premature abandonment ofthe best theory, but which eventually gets replaced by a consensus onit. Besides the result that connectivity can be harmful, it has alsobeen shown that learning in less connected networks is slower, whichindicates a trade-off between accuracy and speed in the context ofsocial learning (Zollman 2007; Grim, Singer, Fisher, et al. 2013).

Subsequent studies showed, however, that the Zollman effect is notvery robust within the parameter space of the original model(Rosenstock et al. 2017). In particular, the result holds for thoseparameters that can be considered characteristic of difficult inquiry:scenarios in which there is a relatively small number of scientists,the evidence is gathered in relatively small batches, and thedifference between the objective success of the rival hypotheses isrelatively small. Moreover, additional models showed that if thediversity (and hence, exploration) of pursued hypotheses is generatedin some other way, more connected communities may outperform lessconnected ones. For instance, Kummerfeld and Zollman (2016) showedthat relaxing the trade-off between exploration and exploitation byallowing agents to occasionally gain information about the hypothesisthey are not currently pursuing is a way to generate diversity,leading a fully connected community to perform better than lessconnected ones. Another way of generating diversity was examined byFrey and Šešelja (2020): they show that if scientistshave a dose of caution or “rational inertia” when decidingwhether to abandon their current theory and start pursuing the rival,a fully connected community gets a sufficient degree of exploration tooutperform less connected groups. Similar points have been made withABMs based on other modeling frameworks, such as the boundedconfidence model by Douven and Hegselmann (2022), or anargumentation-based ABM by Borg, Frey, Šešelja, andStraßer (2018), each of which shows a different way ofpreserving transient diversity in spite of a high degree ofconnectivity.

4.2.2 The spread of disinformation

ABMs studying epistemically pernicious strategies in scientificcommunities have largely employed network epistemology bandit models(seeSection 3.2). For instance, a model by Holman and Bruner (2015) examines how aninterference of industry-sponsored agents may impact the informationflow in the medical community, and which strategies scientists couldemploy to protect themselves from such a pernicious influence. Forthis purpose, they consider a scenario in which medical doctorsregularly communicate with industry-sponsored agents about theefficacy of a certain pharmaceutical product as a treatment for agiven disease. Since the industry-sponsored agents are motivated byfinancial rather than epistemic interests and they are unlikely tochange their minds no matter how much opposing evidence they receive,they are not merely biased, but “intransigently biased”.Their simulations indicate two ways in which a scientific communitycan protect itself from the pernicious influence of the intransigentlybiased agent: first, by increasing their connectivity, and second, bylearning how to reorganize their social network on the basis oftrustworthiness, which leads them to eventually ignore the biasedagent. In a follow-up model, Holman and Bruner (2017) also show howthe industry can bias a scientific community without corrupting any ofthe individual scientists that compose it, but by simply helpingindustry-friendly scientists to have successful careers.

Using a similar network-epistemology approach, Weatherall,O’Connor and Bruner (2020) developed a bandit model to study the“Tobacco strategy”, employed by the tobacco industry inthe second half of the twentieth century to mislead the public aboutthe negative health effects of smoking (analyzed in detail by Oreskes& Conway 2010). In particular, the model examines how certaindeceptive practices can mislead public opinion without eveninterfering with (epistemically driven) scientific research. Theauthors look into two such propagandist strategies: a “selectivesharing” of research results that fit the industry’spreferred position, and a “biased production” of researchresults where additional research gets funded, but only suitablefindings get published. The results show that both strategies areeffective in misleading policymakers about the scientific output undervarious examined parameters since in both cases the policymakersupdate their beliefs on the basis of a biased sample of results,skewed towards the worse theory. The authors also look into strategiesemployed by journalists reporting on scientific findings and show thatincautiously aiming to be “fair” by reporting an equalnumber of results from both sides of the controversy may result in thespread of misleading information.

Another example of ABMs developed to examine deception in science isan argumentation-based model by Borg, Frey, Šešelja, andStraßer (2017, 2018; seeSection 3.1.3). Assuming the context of rival research programs where scientists haveto identify the best out of three available ones, deceptivecommunication is represented in terms of agents sharing only positivefindings about their theory, while withholding news about potentialanomalies. The underlying idea is that deception consists in providingsome (true) information while at the same time withholding otherinformation, which leads the receiver to make a wrong inference(Caminada 2009). Unlike the previous two models discussed in thissection, where not all agents display biased or deceptive behavior,Borg et al. study network effects in a population consisting entirelyof deceptive scientists. Such a scenario represents a community thatis driven, for instance, by confirmation bias and an incentive toshield one’s research line from critical scrutiny. Thesimulations show that, first, reliable communities (consisting of nodeceivers) are significantly more successful than the deceptive ones,and second, increasing the connectivity makes it more likely thatdeceptive populations converge on the best theory.

4.3 Cognitive diversity

As we have seen inSection 2.1, the problem of cognitive diversity concerns the relation between thediversity of cognitive features of scientists (including theirbackground beliefs, reasoning styles, research preferences, heuristicsand strategies) and the inquiry of the scientific community.Philosophers of science have especially been interested in how thedivision of labor across different research heuristics impacts theperformance of the community.

A particularly influential ABM tackling this issue is the epistemiclandscape model by Weisberg and Muldoon (2009). The model examines thedivision of labor catalyzed by different research strategies, wherescientists can act as the “controls”,“followers” or “mavericks” (see aboveSection 3.1). In view of the simulations, Weisberg and Muldoon argue that, first,mavericks outperform other research strategies. Second, if we considerthe maverick strategy to be costly in terms of the necessaryresources, then introducing mavericks to populations of followers canlead to an optimal division of labor. While Weisberg andMuldoon’s ABM eventually turned out to include a coding error(Alexander et al. 2015), their claim that cognitive diversity improvesthe productivity of scientists received support from adjusted versionsof the model, albeit with some qualifications.

First, Thoma’s (2015) model showed that cognitively diversegroups outperform homogeneous ones if scientists are sufficientlyflexible to change their current approach and sufficiently informedabout the research conducted by others in the community. Second,Pöyhönen (2017) confirmed that if we consider the maverickstrategy to be slightly more time-consuming, mixed populations ofmavericks and followers may outperform communities consisting only ofmavericks in terms of the average epistemic significance of theobtained results. According to Pöyhönen, another conditionthat needs to be satisfied if cognitive diversity is to be beneficialconcerns the topology of the landscape: diverse populations outperformhomogeneous ones only in case of more challenging inquiries(represented in terms of rugged epistemic landscapes), but not in caseof easy research problems (represented by landscapes such as Weisbergand Muldoon’s one,Figure 2). The importance of the topology of the landscape was also emphasizedby Alexander and colleagues (2015) who use NK-landscapes to show thatwhether social learning is beneficial or not crucially depends on thelandscape’s topology. Finally, there are other researchstrategies (such as Alexander and colleagues’“swarm” one, seeSection 3.1.1) that outperform Weisberg and Muldoon’s mavericks, while smallchanges in the follower strategy may significantly improve itsperformance (Fernández Pinto & Fernández Pinto2018).

Another aspect of cognitive diversity that received attention in themodeling literature concerns the relationship between diversity andexpertise. This issue was first studied in economics by Hong and Page(2004). The model examined how heuristically diverse groups,consisting of agents with diverse problem-solving approaches, comparewith groups consisting solely of experts with respect to finding asolution to a particular problem. Hong and Page’s originalresult suggested that “diversity trumps ability” in thesense that groups consisting of individuals employing diverseheuristic approaches outperfom groups consisting solely of experts,that is, agents who are the best “problem-solvers”. Whilethis finding became quite influential, subsequent studies showed thatit does not hold robustly once more realistic assumptions aboutexpertise are added to the model (Grim, Singer, Bramson, et al. 2019;Reijula & Kuorikoski 2021; see also Singer 2019).

The problem of cognitive diversity and the division of labor inscientific communities was also studied by Hegselmann andKrause’s bounded-confidence model (Hegselmann & Krause 2006;see aboveSection 3.3). The ABM examines opinion dynamics in a community that is diverse inthe sense that only some individuals are active truth seekers, whileothers adjust their beliefs by exchanging opinions with those agentswho have sufficiently similar beliefs to their own. Hegselmann andKrause investigate conditions under which such a community can reachconsensus on the truth, combining agent-based modeling and analyticalmethods. They show that, on the one hand, if all agents in the modelare truth seekers, they achieve a consensus on the truth. On the otherhand, if the community divides labor, its ability to reach a consensuson the truth will depend on the number of truth seeking agents, theposition of the truth relative to the agents’ opinions, thedegree of confidence determining the scope of opinion exchange and therelative weight of the truth signal (in contrast to the weight ofsocial information). For instance, under certain parameter settingseven a single truth-seeking agent will lead the rest of the communityto the truth.

4.4 Social diversity

As we have seen inSection 2.4, ABMs were introduced to study two issues that concern socialdiversity in the context of science: first, epistemic effects ofsocially diverse scientific groups, and second, factors that canundermine social diversity or disadvantage members of minorities inscience.

Concerning the former question, Fazelpour and Steel (2022) developed anetwork epistemology bandit model (seeSection 3.2) to study effects of social (or demographic) diversity on collectiveinquiry. The model examines how different degrees of trust betweensocially distinct sub-groups impact the performance of the scientificcommunity. The simulations show that social diversity can improvegroup inquiry by reducing the excessive trust scientists may place ineach other’s findings. This is particularly relevant in cases ofdifficult inquiry in tightly connected communities, where a highdegree of trust may lead to a premature endorsement of an objectivelyfalse hypothesis. The authors also demonstrate how social diversitycan counteract negative effects of conformity in the choice ofresearch paths, under the condition that the tendency to conformisn’t too high. Using a similar bandit framework, Wu (2023)examines how a marginalized group learns in environments in whichmembers of the dominant group ignore or devalue results coming fromthe former. She shows that in such situations, members of themarginalized group can achieve an epistemic advantage in the sense offorming more accurate beliefs than the members of the dominant group.In this way, the model aims to explain the “inversionthesis” from standpoint epistemology, according to whichmarginalized groups may have an epistemically privileged status (Wylie2003).

With regard to the latter question, we previously mentioned the modelby Rubin and Schneider (2021), which shows how minority members canbecome disadvantaged due to the priority rule in science and thestructure of the scientific community (Section 4.1). In a follow-up model, Rubin (2022) demonstrates how theorganizational structure of the scientific community can alsoinfluence the emergence of “citation gaps”, wherepublications by members of underrepresented and minority groups arecited less than those by members of the majority. The impact of one’sminority status on one’s position in scientific collaborations wasmainly studied by means of evolutionary game-theoretic models ofbargaining (seeSection 3.4). The models show different ways in which minorities can becomediscriminated against merely due to their minority status and theimpact of such a status on the bargaining and collaborative practices(Rubin & O’Connor 2018; O’Connor & Bruner 2019;Ventura 2023). Moreover, the results show that such an outcome mayhave a negative impact on scientific progress, and it may help toexplain why gender and racial minorities tend to cluster in academicsubdisciplines (O’Connor & Bruner 2019; Rubin &O’Connor 2018). We will revisit these models below inSection 4.7, devoted to the issue of scientific collaboration.

4.5 Peer-disagreement in science

As we have seen inSection 2.5, peer disagreement debate highlighted different norms that can guidescientists. ABMs were introduced to examine the impact of these normson the communal inquiry. To this end, Douven (2010) and De Langhe(2013) enhanced the bounded confidence model by Hegselmann and Krause(2006; seeSection 3.3) to study the impact of the Conciliatory and Steadfast Norms on thegoals of inquiry.^[9]

Both models suggest that the impact of these norms is contextdependent. In Douven’s model, when data received through inquiryis not very noisy—in the sense that it is indicative of thetruth rather than resulting from measurement errors—conciliatingpopulations will be faster in discovering the truth than the steadfastones. Nevertheless, if the data becomes noisy, the simulations show atrade-off between accuracy and speed. While steadfast populations getwithin a moderate distance of the truth relatively quickly, theydon’t improve their accuracy in the subsequent rounds of thesimulation. In contrast, conciliating populations end up closer to thetruth but it takes them relatively longer to do so. This indicatesthat whether a norm is optimal for the communal goals may largelydepend on contextual issues of inquiry, such as obtaining noise-freeevidence. De Langhe comes to a similar conclusion, using a model thatrepresents longstanding scientific disagreements and an inquiryinvolving multiple rivaling epistemic systems. The model is inspiredby Goldman’s (2010) idea that even though disagreeing peers mayshare the evidence concerning the issue in question, they may notshare the evidence for the epistemic system within which they evaluatethe former kind of evidence. His simulations suggest that whileconciliating within one’s epistemic system is beneficial, thatis not the case when it comes to disagreements between differentepistemic systems.^[10]

4.6 Scientific polarization

The initial work on ABMs of polarization in scientific communities isbased on Hegselmann and Krause’s bounded confidence model(Hegselmann & Krause 2006; seeSection 3.3).^[11] The model shows that a community can polarize when some agents formtheir opinions by disregarding the evidence coming from the world andby instead taking into account only what they learn from others, whohave sufficiently similar views. Subsequent models, based on differentframeworks, sought to examine whether polarization can emerge even ifall agents are truth-seekers, that is, rational agents who form theiropinions only in view of epistemic considerations.

For example, a model by Singer et al. (2019) shows the emergence ofpolarization in groups of deliberating agents who share reasons fortheir beliefs, and who merely use a coherence-based approach to managetheir limited memory (by forgetting those reasons that conflict withthe view supported by most of their previous considerations). Usingthe Bayesian modeling framework Laputa, Olsson (2013) shows howpolarization can emerge over the course of deliberation if agentsassign different degrees of trust to the testimony of others,depending on how similar views they hold. In another Bayesianframework based on bandit models, O’Connor and Weatherall (2018)show how a community of scientists who share not only their testimony,but unbiased evidence, can end up in a polarized state if they treatevidence obtained by other scientists, whose beliefs are too far offfrom their own, as uncertain. Moreover, Michelini, Osorio andcolleagues (forthcoming) combine bandit models with thebounded-confidence framework to study polarization in communities thatdisagree on the diagnostic value of evidence for a certain hypothesis(modeled by the Bayes factor). Their results indicate that an initialdisagreement on the diagnosticity of evidence can lead topolarization, depending on the sample size of the performed studiesand the confidence interval within which scientists share theiropinions.

Polarization in epistemic communities has also been studied by ABMssimulating argumentative dynamics. For example, using the framework byBetz (2013), Kopecky (2022) proposes a model that shows how certainways in which rational agents engage in argumentative exchange in anopen debate forum can result in polarized communities, even if agentsdon’t have a preference for whom they argue with.

4.7 Scientific collaboration

As discussed inSection 2.7, ABMs have been used to study when collaborating is beneficial and howit impacts inquiry. To examine why collaborations emerge and when theyare beneficial, Boyer-Kassem and Imbert (2015) developed acomputational model in which scientists collaborate by sharingintermediate results. They study collaborative groups in a largercompetitive community, driven by the priority rule, wherecollaborators have to equally share the reward for making a discovery.Their simulations suggest that collaborating is beneficial both forcommunal and individual inquiry. When scientists collaborate, theyproceed faster with their inquiry, making it more likely that theywill belong to a team that is first to make a discovery. Moreover,while the scientific community may profit from all scientists fullycollaborating (since solutions found by some scientists will be sharedwith everyone), for individual scientists other collaborativeconstellations may be better (see also Boyer-Kassem & Imbertforthcoming).

To examine conditions for optimal collaboration Zollman (2017)proposed a network-epistemology model. The model starts from theassumption that by collaborating scientists teach each other differentconceptual schemes aimed at solving scientific problems, and thatcollaboration comes at a cost. His findings suggest that reducing thecosts of collaborations and enlarging the size of the collaborativegroup benefits those involved, though encouraging scientists to engagewith different collaborative groups may not lead to efficientinquiry.

While the above models studies socially uniform communities, ABMs havealso been employed to study socially diverse collaborativeenvironments and discriminatory patterns that can endogenously emergein them. These models are typically based on evolutionarygame-theoretic frameworks examining bargaining norms that accompanyacademic collaborations (seeSection 3.4). Bruner (2019) initially used this framework to show that in culturalinteractions involving a division of resources, minority members canbecome disadvantaged merely due to the smaller size of their group.^[12] O’Connor and Bruner (2019) use this approach to study theemergence of similar patterns in epistemic communities. Thesimulations show that minorities can become disadvantaged inscientific collaborations due to the size of their group, where theeffect is increased the smaller their group is. Rubin andO’Connor (2018) use a similar framework to examine diversity anddiscrimination in epistemic collaboration networks, with a specialfocus on the role of homophily—the tendency to preferentiallyestablish links with members of one’s own social group incontrast to members of the outgroup. Their results suggest thatdiscriminatory norms, likely to emerge in academic interaction, maypromote homophily and decrease social diversity in collaborativenetworks. Furthermore, Ventura (2023) shows how inequality can emergein collaboration networks even in the absence of demographiccategories since scientists can become disadvantaged merely due to thestructure of the collaboration network and their specific position init.

4.8 Additional themes

As mentioned inSection 2.8, ABMs have been applied to the study of many additional themes inphilosophy of science: the problem of allocating funding to researchgrants (Harnagel 2019; Avin 2019a,b), examining Popperian and Kuhnianmodels of scientific progress (De Langhe 2014a), the dynamics ofnormal and revolutionary science (De Langhe 2017), pluralism ofscientific methods (Currie & Avin 2019), the impact of methodsguiding research design and data analysis on the quality of scientificfindings (Smaldino & McElreath 2016), strategic behavior ofscientists as a motivated exchange of beliefs (Merdes 2021),journals’ publishing strategies (Zollman 2009), meta-inductivelearning (Schurz 2009, 2012), the impact of evidential strength on theaccuracy of scientific consensus (Santana 2018), the management andstructure of scientific groups (Sikimić & Herud-Sikimić2022), the reliability of testimonial norms, which guide how oneshould change beliefs in light of others’ claims (Mayo-Wilson2014), the assessment of source reliability (Merdes, Von Sydow, &Hahn 2021), the impact of “epistemic stubbornness” as anunwillingness to change one’s scientific stance in face ofcountervailing evidence (Santana 2021), the impact of differenttheory-choice evaluations on the efficiency of inquiry (Borg et al.2019), and so forth.

5. Epistemology of Agent-Based Modeling

In the previous section we delved into various results obtained bymeans of agent-based modeling. But how seriously should we take eachof these findings? What exactly do they tell us about science? Giventhat one of the main cognitive functions of models in science is tohelp us to learn about the world (cf. the entry onmodels in science), this raises the question: what exactly can we learn with ABMs, andwhich methodological steps do we have to follow to acquire suchknowledge? These issues are a matter of the epistemology ofagent-based modeling.^[13]

5.1 The challenge of abstract modeling

In an early discussion on simulations in social sciences, Boero andSquazzoni (2005) proposed a classification of ABMs into case-basedmodels, typifications, and theoretical abstractions—based on theproperties of the represented target. While a case-based modelrepresents an empirical scenario that is delineated in time and space,typifications represent classes of empirical phenomena. Finally,theoretical abstractions, as their name suggests, abstract away fromvarious features of empirical targets and aim at simplified andgeneral representation of social phenomena.

Following the tradition of abstract ABMs in social sciences, themajority of simulations in philosophy of science have been developedwithin the so-called “KISS” (Keep it Simple,Stupid) approach.^[14] As such, they belong to the third type of ABMs listed above. The mainadvantage of constructing simple models is that they allow for aneasier understanding of the represented causal mechanisms than complexmodels do. However, this also raises the question: what can we learnfrom these models about real science, given their highly idealizedcharacter? For instance, do results of such simulations increase ourunderstanding of scientific communities? Can we use them to providepotential explanations of certain episodes from the history ofscience, or to formulate normative recommendations for how scientificinquiry should be organized? These questions have been a matter ofphilosophical debate, closely related to the discussion on theepistemic function of highly idealized or toy-models across empirical sciences.^[15] In particular, critics have argued that if abstract ABMs are to beinformative of empirical targets, they first need to be verified andempirically validated.

5.2 Verification and validation: from exploratory to explanatory functions of ABMs

In the context of agent-based modeling, verification is a method ofevaluating the accuracy of the computer program underlying an ABMrelative to its conceptual design. Validation, on the other hand,concerns the evaluation of the link between the model and itspurported target. Irrespective of the purpose for which the model wasbuilt, it always requires some degree of verification to ensure thatits simulation code does not suffer from bugs and other unintendedissues. The type of required validation depends on the purpose of themodel and its intended target. As Mayo-Wilson and Zollman (2021) haveargued, validation isn’t always necessary for some modelingpurposes, such as illustrating that certain events or situations aretheoretically possible. Models can instead be justified by“plausibility arguments” and in view of stylizedhistorical case studies. According to Mayo-Wilson and Zollman, ABMsplay a role analogous to thought experiments in that they can be usedto evoke normative intuitions, to justify counterfactual claims, toillustrate possibilities and impossibilities, and so on. Moreover,when it comes to questions concerning the dynamics of social systems,they are more apt for the task than mere thought experiments due tothe complexity of such target phenomena.

Empirical validation is also not required if the function of an ABM isto provide a proof-of-concept or a how-possibly explanation. A modelprovides aproof-of-concept (sometimes also called a“proof-of-possibility” or“proof-of-principle”) if it merely demonstrates atheoretical possibility of a certain phenomenon (cf. Arnold 2008). Thetarget phenomenon may be represented in an abstract, idealized way,disregarding the empirical adequacy of the modeling assumptions.Gelfert (2016: 85–86) distinguishes two specific functions ofthe proof-of-concept modeling. First, a model may exemplify how anapproach or methodology can generate a potential representation of agiven target phenomenon. For instance, if an ABM shows how theframework of epistemic landscapes can be used to represent cognitivediversity, it provides a proof-of-concept in this sense of the term.Second, a model may provide results showing that a certain causalmechanism can be found within the model-world. For instance, if ahighly idealized simulation shows that a certain kind of cognitivediversity results in efficient inquiry of the modeled community, itprovides a proof-of-concept in the latter sense. Whileproof-of-concept modeling is sometimes taken to be just a preliminarystep in the development of more realistic models (see for instanceGräbner 2018), it has also been considered as providing valuablephilosophical insights already on its own (see for exampleŠešelja 2021a).

Similarly, simulations providehow-possibly explanations(HPEs) if they show how a certain target phenomenon can possiblyresult from certain preconditions (Rosenstock et al. 2017;Gräbner 2018; Frey & Šešelja 2018;Šešelja 2022b). In contrast to how-actuallyexplanations, or explanations simpliciter, which are accounts of howphenomena actually occur, HPEs (sometimes also called “potentialexplanations”) cover accounts of possible ways in whichphenomena can occur.^[16] According to a broad reading of the notion proposed byVerreault-Julien (2019), HPEs are propositions that have the form“It is possible that ‘p because q’.”, wherep is the explanandum,q the explanans, and thepossibility may refer to various types of modalities such asepistemic, logical, causal, and so forth.

The above epistemic functions—providing a proof-of-concept or anHPE—are often considered a kind of exploratory modeling, wherethe represented target can be an abstract, theoretical phenomenon(Ylikoski & Aydinonat 2014; Šešelja 2021a). Incontrast, to provide explanations of empirical phenomena, many haveargued that ABMs need to be empirically validated. This has beenemphasized as important in case models are supposed to explain certainpatterns from actual scientific practice, to provide evidencesupporting existing empirical (including historical) hypotheses, or toprovide suggestions for interventions through science policy (Arnold2014; Martini & Fernández Pinto 2017; Thicke 2020;Šešelja 2021a).

Validation of ABMs consists in examining whether the model is areasonable representation of the target, where“reasonable” may refer to different aspects of the model(Gräbner 2018). For instance, we may test how well mechanismsrepresented in the model conform to our empirical knowledge aboutthem, whether the exogenous inputs for the model are empiricallyplausible, to what extent the output of the model replicates existingknowledge about the target, or whether it can predict its futurestates. Different authors have suggested different elaborations ofthese points with respect to ABMs in philosophy of science (see Thicke2020; Šešelja 2022b; Bedessem 2019; Politi 2021; Pesonen2022).

While some of these steps may be challenging, others may be morefeasible. For instance Martini and Fernández Pinto (2017) arguethat ABMs of science can be calibrated on empirical data. Harnagel(2019) exemplifies how this could be done by calibrating an epistemiclandscape model with bibliometric data.

Others have argued that results of simulations should at least beanalyzed in terms of theirrobustness. Robustness analysis(RA) includes:

1. parameter RA, examining the stability of results with respect tochanges in parameter values of the model, usually studied by means ofsensitivity analysis;^[17]
2. structural RA, which focuses on the stability of results underchanges in structural features of the model and its underlyingassumptions;
3. representational RA, which studies the stability of the resultswith respect to changes in the representational framework, modellingtechnique, or modelling medium (Weisberg 2013: Chapter 9; Houkes,Šešelja, & Vaesen forthcoming).

The main purpose of RA in the context of ABMs is to help withunderstanding conditions (in the model world) under which results ofthe model hold: whether they depend on specific parameter values or onspecific assumptions about the causal factors involved in therepresented target, or whether the result is just an artifact ofcertain idealizing assumptions in the model. This can benefit ourunderstanding of the model and its results in at least two ways. Onthe one hand, if an RA shows that the result holds under variouschanges in parameter values, structural and representationalassumptions, this may increase our confidence that the result is not amere artifact of idealizations in the model (see, e.g., Kuorikoski,Lehtinen, & Marchionni 2012).^[18] On the other hand, if the result holds only for specific parametervalues or under specific structural or representational assumptions,this can help to delineate the context of application of the model(for example, the context of difficult inquiry, the context of inquiryinvolving a small scientific community, and so on).

5.3 The “family of models perspective”

The importance of robustness analysis suggests that determining theepistemic function of a single ABM in isolation may be difficult. Inorder to assess whether the results of an ABM are explanatory of acertain empirical target, it is useful to check whether they arerobust under changes in the idealizing modeling assumptions. One wayto do so is to construct a new model by replacing some assumptions inthe previous one. Moreover, using a model based on an entirelydifferent representational framework may additionally help in suchrobustness studies. This is why the value of using multiple models tostudy the same research question has been increasingly emphasized inthe modeling literature.

For instance, Aydinonat, Reijula, and Ylikoski (2021) have argued forthe importance of the “family-of-models perspective”. Theauthors consider models as argumentative devices, where an argumentsupported by a given model can be strengthened by means of analysesbased on subsequent models. The approach according to which phenomenashould be studied by means of multiple models has also been endorsedin the broader context of modeling in social sciences (cf. Page 2018;Kuhlmann 2021).

In sum, even though ABMs in philosophy of science tend to be highlyidealized, they can have exploratory roles, such as identifyingpossible causal mechanisms underlying scientific inquiry, offeringhow-possibly explanations, or providing conjectures and novelperspectives on historical case-studies (cf. Šešelja2022a). In addition, they can play more challenging epistemic roles(such as providing explanations of empirical phenomena or evidence forempirical hypotheses) if accompanied by validation procedures, whichmay benefit from looking into classes of models aimed at the sametarget phenomenon.

6. Conclusion and Outlook

This entry has provided an overview of applications of agent-basedmodeling to issues studied by philosophers of science. Since suchapplications are primarily concerned with social aspects of scientificinquiry, ABMs in philosophy of science are mainly developed within thesubfield of formal social epistemology of science. While agent-basedmodeling has experienced rapid growth in this domain, it remains tostand the test of time. In particular, foundational issues concerningthe epistemic status of abstract models and their methodologicalunderpinnings (for instance, the KISS vs. the KIDS approach, seefootnote 14) remain open. Having said that, theoretical achievements made with theexisting models have been remarkable and they provide a firm groundfor further growth of agent-based modeling as a philosophicalmethod.

One avenue that has received relatively little attention is thecombination of agent-based modeling and empirical methods. Althoughearly philosophical discussions on ABMs emphasized the fruitfulcombination of experimental and computational methods (see, e.g.,Hartmann, Lisciandra, & Machery 2013), this area has not beenexplored extensively, leaving ample opportunities for future research.For instance, highly idealized ABMs can benefit from validation interms of experimental studies (as exemplified by Mohseni et al. 2021).Moreover, qualitative research, such as ethnographic studies ofscientific communities, can help to empirically inform assumptionsused to build models (see, e.g., Ghorbani, Dijkema, & Schrauwen2015).

Another emerging line of inquiry comes from the recent developments inartificial intelligence. In particular, big data and machine-learningmodels can be a fruitful way of generating input for ABMs, such as thebehavior of agents (see, e.g., Kavak, Padilla, Lynch, & Diallo2018; Zhang, Valencia, & Chang 2023). For instance, naturallanguage processing technology has been used to represent agents whoexpress their arguments in terms of natural language (Betz 2022). Suchenhancements provide novel opportunities for highly idealized models:from their robustness analyses in which certain assumptions arede-idealized to the exploration of novel questions and phenomena.

Bibliography

• Adami, Christoph, Jory Schossau, and Arend Hintze, 2016,“Evolutionary Game Theory Using Agent-Based Methods”,Physics of Life Reviews, 19: 1–26.doi:10.1016/j.plrev.2016.08.015
• Alexander, J. McKenzie, 2013, “Preferential Attachment andthe Search for Successful Theories”,Philosophy ofScience, 80(5): 769–782. doi:10.1086/674080
• Alexander, Jason McKenzie, Johannes Himmelreich, and ChristopherThompson, 2015, “Epistemic Landscapes, Optimal Search, and theDivision of Cognitive Labor”,Philosophy of Science,82(3): 424–453. doi:10.1086/681766
• Alexandrova, Anna, 2008, “Making Models Count”,Philosophy of Science, 75(3): 383–404.doi:10.1086/592952
• Andersen, Hanne, 2010, “Joint Acceptance and ScientificChange: A Case Study”,Episteme, 7(3): 248–265.doi:10.3366/epi.2010.0206
• Anderson, Elizabeth, 1995, “Knowledge, Human Interests, andObjectivity in Feminist Epistemology”:, PhilosophicalTopics, 23(2): 27–58. doi:10.5840/philtopics199523213
• Angere, Staffan and Erik J. Olsson, 2017, “Publish Late,Publish Rarely!: Network Density and Group Performance in ScientificCommunication”, in Boyer-Kassem, Mayo-Wilson, and Weisberg 2017:34–62 (ch. 2).
• Arnold, Eckhart, 2008,Explaining Altruism: A Simulation-BasedApproach and Its Limits, (Practical Philosophy 11), Frankfurt:Ontos Verlag. doi:10.1515/9783110327571
• –––, 2014, “What’s wrong with socialsimulations?”,The Monist, 97(3): 359–377.doi:10.5840/monist201497323
• Avin, Shahar, 2019a, “Centralized Funding and EpistemicExploration”,The British Journal for the Philosophy ofScience, 70(3): 629–656. doi:10.1093/bjps/axx059
• –––, 2019b, “Mavericks andLotteries”,Studies in History and Philosophy of SciencePart A, 76: 13–23. doi:10.1016/j.shpsa.2018.11.006
• Axelrod, Robert M., 1984,The Evolution of Cooperation,New York: Basic Books.
• –––, 1997,The Complexity of Cooperation:Agent-Based Models of Competition and Collaboration, (PrincetonStudies in Complexity), Princeton, NJ: Princeton UniversityPress.
• Axelrod, Robert and William D. Hamilton, 1981, “TheEvolution of Cooperation”,Science, 211(4489):1390–1396. doi:10.1126/science.7466396
• Axtell, Robert L, Joshua M. Epstein, and H. Peyton Young, 2001,“The Emergence of Classes in a Multi-Agent BargainingModel”, inSocial Dynamics, Steven N. Durlauf and H.Peyton Young (eds.), (Economic Learning and Social Evolution 4),Cambridge, MA: MIT Press, 191–211 (ch. 7).
• Aydinonat, N. Emrah, Samuli Reijula, and Petri Ylikoski, 2021,“Argumentative Landscapes: The Function of Models in SocialEpistemology”,Synthese, 199(1–2): 369–395.doi:10.1007/s11229-020-02661-9
• Bala, Venkatesh and Sanjeev Goyal, 1998, “Learning fromNeighbours”,The Review of Economic Studies, 65(3):595–621. doi:10.1111/1467-937X.00059
• Balietti, Stefano, Michael Mäs, and Dirk Helbing, 2015,“On Disciplinary Fragmentation and Scientific Progress”,PLOS ONE, 10(3): e0118747.doi:10.1371/journal.pone.0118747
• Bedessem, Baptiste, 2019, “The Division of Cognitive Labor:Two Missing Dimensions of the Debate”,European Journal forPhilosophy of Science, 9(1): article 3.doi:10.1007/s13194-018-0230-8
• Berry, Donald A. and Bert Fristedt, 1985,Bandit Problems:Sequential Allocation of Experiments, (Monographs on Statisticsand Applied Probability), Dordrecht: Springer Netherlands.doi:10.1007/978-94-015-3711-7
• Betz, Gregor, 2013,Debate Dynamics: How Controversy ImprovesOur Beliefs, (Synthese Library : Studies in Epistemology,Logic, Methodology and Philosophy of Science 357), Dordrecht:Springer. doi:10.1007/978-94-007-4599-5
• –––, 2022, “Natural-Language Multi-AgentSimulations of Argumentative Opinion Dynamics”,Journal ofArtificial Societies and Social Simulation, 25(1): article 2.doi:10.18564/jasss.4725
• Boero, Riccardo, and Flaminio Squazzoni, 2005, “DoesEmpirical Embeddedness Matter? Methodological Issues on Agent-BasedModels for Analytical Social Science”,Journal of ArtificialSocieties and Social Simulation, 8(4). [Boero and Squazzoni 2005 available online]
• Bokulich, Alisa, 2014, “How the Tiger Bush Got Its Stripes:‘How Possibly’ vs. ‘How Actually’ ModelExplanations:”,The Monist, 97(3): 321–338.doi:10.5840/monist201497321
• –––, 2017, “Models and Explanation”,inSpringer Handbook of Model-Based Science, Lorenzo Magnaniand Tommaso Bertolotti (eds.), Cham: Springer InternationalPublishing, 103–118. doi:10.1007/978-3-319-30526-4_4
• Bolton, Patrick and Christopher Harris, 1999, “StrategicExperimentation”,Econometrica, 67(2): 349–374.doi:10.1111/1468-0262.00022
• Borg, AnneMarie, Daniel Frey, Dunja Šešelja, andChristian Straßer, 2017, “Examining Network Effects in anArgumentative Agent-Based Model of Scientific Inquiry”, inLogic, Rationality, and Interaction: 6th International Workshop,LORI 2017, Sapporo, Japan, September 11-14, 2017, AlexandruBaltag, Jeremy Seligman, and Tomoyuki Yamada (eds.), (Lecture Notes inComputer Science 10455), Berlin/Heidelberg: Springer BerlinHeidelberg, 391–406. doi:10.1007/978-3-662-55665-8_27
• –––, 2018, “Epistemic Effects ofScientific Interaction: Approaching the Question with an ArgumentativeAgent-Based Model”,Historical Social Research, 43(1):285–309.
• –––, 2019, “Theory-Choice, TransientDiversity and the Efficiency of Scientific Inquiry”,European Journal for Philosophy of Science, 9(2): article 26.doi:10.1007/s13194-019-0249-5
• Boyer-Kassem, Thomas and Cyrille Imbert, 2015, “ScientificCollaboration: Do Two Heads Need to Be More than Twice Better thanOne?”,Philosophy of Science, 82(4): 667–688.doi:10.1086/682940
• –––, forthcoming, “Explaining ScientificCollaboration: A General Functional Account”,The BritishJournal for the Philosophy of Science, first online: August 2021.doi:10.1086/716837
• Boyer-Kassem, Thomas, Conor Mayo-Wilson, and Michael Weisberg(eds.), 2017,Scientific Collaboration and Collective Knowledge:New Essays, New York: Oxford University Press.doi:10.1093/oso/9780190680534.001.0001
• Bramson, Aaron, Patrick Grim, Daniel J. Singer, William J. Berger,Graham Sack, Steven Fisher, Carissa Flocken, and Bennett Holman, 2017,“Understanding Polarization: Meanings, Measures, and ModelEvaluation”,Philosophy of Science, 84(1):115–159. doi:10.1086/688938
• Brock, William A. and Steven N. Durlauf, 1999, “A FormalModel of Theory Choice in Science”,Economic Theory,14(1): 113–130. doi:10.1007/s001990050284
• Bruner, Justin P., 2019, “Minority (Dis)Advantage inPopulation Games”,Synthese, 196(1): 413–427.doi:10.1007/s11229-017-1487-8
• Bueter, Anke, 2022, “Bias as an Epistemic Notion”,Studies in History and Philosophy of Science, 91:307–315. doi:10.1016/j.shpsa.2021.12.002
• Burt, Ronald S., 1992,Structural Holes: The Social Structureof Competition, Cambridge, MA: Harvard University Press.doi:10.4159/9780674029095
• Caminada, Martin, 2009, “Truth, Lies and Bullshit:Distinguishing Classes of Dishonesty”, inProceedings of theSocial Simulation Workshop at the Twenty-First International JointConference on Artificial Intelligence (IJCAI), Pasadena, CA, USA,pp. 39–50.
• Chang, Hasok, 2012,Is Water H₂O? Evidence, Realismand Pluralism, (Boston Studies in the Philosophy of Science,293), Dordrecht ; New York: Springer.doi:10.1007/978-94-007-3932-1
• Christensen, David, 2010, “Higher-Order Evidence”,Philosophy and Phenomenological Research, 81(1):185–215. doi:10.1111/j.1933-1592.2010.00366.x
• Csaszar, Felipe A., 2018, “A Note on How NK LandscapesWork”,Journal of Organization Design, 7(1): article15. doi:10.1186/s41469-018-0039-0
• Currie, Adrian and Shahar Avin, 2019, “Method Pluralism,Method Mismatch, & Method Bias”,Philosopher’sImprint, 19: article 13.
• De Cruz, Helen and Johan De Smedt, 2013, “The Value ofEpistemic Disagreement in Scientific Practice. The Case ofHomofloresiensis”,Studies in History and Philosophy ofScience Part A, 44(2): 169–177.doi:10.1016/j.shpsa.2013.02.002
• De Langhe, Rogier, 2013, “Peer Disagreement under MultipleEpistemic Systems”,Synthese, 190(13): 2547–2556.doi:10.1007/s11229-012-0149-0
• –––, 2014a, “A Comparison of Two Models ofScientific Progress”,Studies in History and Philosophy ofScience Part A, 46: 94–99.doi:10.1016/j.shpsa.2014.03.002
• –––, 2014b, “A Unified Model of theDivision of Cognitive Labor”,Philosophy of Science,81(3): 444–459. doi:10.1086/676670
• –––, 2017, “Towards the Discovery ofScientific Revolutions in Scientometric Data”,Scientometrics, 110(1): 505–519.doi:10.1007/s11192-016-2108-x
• De Langhe, Rogier and Matthias Greiff, 2010, “Standards andthe Distribution of Cognitive Labour: A Model of the Dynamics ofScientific Activity”,Logic Journal of the IGPL, 18(2):278–293. doi:10.1093/jigpal/jzp058
• Deffuant, Guillaume, Frédéric Amblard, GérardWeisbuch, and Thierry Faure, 2002, “How Can Extremism Prevail? AStudy Based on the Relative Agreement Interaction Model”,Journal of Artificial Societies and Social Simulation, 5(4). [Deffuant et al. 2002 available online]
• Deffuant, Guillaume, David Neau, Frederic Amblard, andGérard Weisbuch, 2000, “Mixing Beliefs among InteractingAgents”,Advances in Complex Systems, 3(01n04):87–98. doi:10.1142/S0219525900000078
• DeGroot, Morris H., 1974, “Reaching a Consensus”,Journal of the American Statistical Association, 69(345):118–121. doi:10.1080/01621459.1974.10480137
• Derex, Maxime and Robert Boyd, 2016, “Partial ConnectivityIncreases Cultural Accumulation within Groups”,Proceedingsof the National Academy of Sciences, 113(11): 2982–2987.doi:10.1073/pnas.1518798113
• Derex, Maxime, Charles Perreault, and Robert Boyd, 2018,“Divide and Conquer: Intermediate Levels of PopulationFragmentation Maximize Cultural Accumulation”,PhilosophicalTransactions of the Royal Society B: Biological Sciences,373(1743): 20170062. doi:10.1098/rstb.2017.0062
• Douglas, Heather E., 2009,Science, Policy, and the Value-FreeIdeal, Pittsburgh, PA: University of Pittsburgh Press.
• Douven, Igor, 2010, “Simulating Peer Disagreements”,Studies in History and Philosophy of Science Part A, 41(2):148–157. doi:10.1016/j.shpsa.2010.03.010
• –––, 2019, “Computational Models in SocialEpistemology”, in Fricker, Graham, Henderson, & Pedersen2019: 457–465.
• Douven, Igor and Rainer Hegselmann, 2021, “Mis- andDisinformation in a Bounded Confidence Model”,ArtificialIntelligence, 291: 103415. doi:10.1016/j.artint.2020.103415
• –––, 2022, “Network Effects in a BoundedConfidence Model”,Studies in History and Philosophy ofScience, 94: 56–71. doi:10.1016/j.shpsa.2022.05.002
• Douven, Igor and Christoph Kelp, 2011, “Truth Approximation,Social Epistemology, and Opinion Dynamics”,Erkenntnis,75(2): 271–283. doi:10.1007/s10670-011-9295-x
• Douven, Igor and Alexander Riegler, 2010, “Extending theHegselmann-Krause Model I”,Logic Journal of the IGPL,18(2): 323–335. doi:10.1093/jigpal/jzp059
• Douven, Igor and Sylvia Wenmackers, 2017, “Inference to theBest Explanation versus Bayes’s Rule in a Social Setting”,The British Journal for the Philosophy of Science, 68(2):535–570. doi:10.1093/bjps/axv025
• Dray, William H., 1957,Laws and Explanation in History,(Oxford Classical & Philosophical Monographs), London: OxfordUniversity Press.
• Dung, Phan Minh, 1995, “On the Acceptability of Argumentsand Its Fundamental Role in Nonmonotonic Reasoning, Logic Programmingandn-Person Games”,Artificial Intelligence,77(2): 321–357. doi:10.1016/0004-3702(94)00041-X
• Edmonds, Bruce, 2017, “Different Modelling Purposes”,inSimulating Social Complexity: A Handbook, Bruce Edmondsand Ruth Meyer (eds.), second edition, (Understanding ComplexSystems), Cham: Springer, 39–58.doi:10.1007/978-3-319-66948-9_4
• Edmonds, Bruce and Scott Moss, 2005, “From KISS to KIDS– An ‘Anti-Simplistic’ Modelling Approach”, inMulti-Agent and Multi-Agent-Based Simulation: Joint Workshop MABS2004, Paul Davidsson, Brian Logan, and Keiki Takadama (eds.),(Lecture Notes in Computer Science 3415), Berlin, Heidelberg: SpringerBerlin Heidelberg, 130–144.doi:10.1007/978-3-540-32243-6_11
• Elga, Adam, 2007, “Reflection and Disagreement”,Noûs, 41(3): 478–502.doi:10.1111/j.1468-0068.2007.00656.x
• Epstein, Joshua M., 2006,Generative Social Science: Studiesin Agent-Based Computational Modeling, (Princeton Studies inComplexity), Princeton, NJ: Princeton University Press.
• Epstein, Joshua M, 2008, “Why Model?”,Journal ofArtificial Societies and Social Simulation, 11(4): article 12. [Epstein 2008 available online]
• Epstein, Joshua M. and Robert Axtell, 1996,Growing ArtificialSocieties: Social Science from the Bottom Up, (Complex AdaptiveSystems), Washington, DC: Brookings Institution Press.
• Fang, Christina, Jeho Lee, and Melissa A. Schilling, 2010,“Balancing Exploration and Exploitation Through StructuralDesign: The Isolation of Subgroups and Organizational Learning”,Organization Science, 21(3): 625–642.doi:10.1287/orsc.1090.0468
• Fazelpour, Sina and Daniel Steel, 2022, “Diversity, Trust,and Conformity: A Simulation Study”,Philosophy ofScience, 89(2): 209–231. doi:10.1017/psa.2021.25
• Feldman, Richard, 2006, “Epistemological Puzzles aboutDisagreement”, inEpistemology Futures, StephenHetherington (ed.), Oxford/New York: Oxford University Press,216–326.
• Feldman, Richard and Ted A. Warfield (eds.), 2010,Disagreement, Oxford/New York: Oxford University Press.doi:10.1093/acprof:oso/9780199226078.001.0001
• Fernández Pinto, Manuela and Daniel Fernández Pinto,2018, “Epistemic Landscapes Reloaded: An Examination ofAgent-Based Models in Social Epistemology.”,HistoricalSocial Research/Historische Sozialforschung, 43(1): 48–71.doi:10.12759/HSR.43.2018.1.48-71
• Feyerabend, Paul K., 1975,Against Method, London: NewLeft Books.
• Forber, Patrick, 2010, “Confirmation and Explaining HowPossible”,Studies in History and Philosophy of Science PartC: Studies in History and Philosophy of Biological and BiomedicalSciences, 41(1): 32–40.doi:10.1016/j.shpsc.2009.12.006
• –––, 2012, “Conjecture and Explanation: AReply to Reydon”,Studies in History and Philosophy ofScience Part C: Studies in History and Philosophy of Biological andBiomedical Sciences, 43(1): 298–301.doi:10.1016/j.shpsc.2011.10.018
• French, John R. P. Jr, 1956, “A Formal Theory of SocialPower.”,Psychological Review, 63(3): 181–194.doi:10.1037/h0046123
• Frey, Daniel and Dunja Šešelja, 2018, “What Isthe Epistemic Function of Highly Idealized Agent-Based Models ofScientific Inquiry?”,Philosophy of the SocialSciences, 48(4): 407–433. doi:10.1177/0048393118767085
• Frey, Daniel and Dunja Šešelja, 2020,“Robustness and Idealizations in Agent-Based Models ofScientific Interaction”,The British Journal for thePhilosophy of Science, 71(4): 1411–1437.doi:10.1093/bjps/axy039
• Fricker, Miranda, Peter J. Graham, David Henderson, and Nikolaj J.L. L. Pedersen (eds.), 2019,The Routledge Handbook of SocialEpistemology, New York: Routledge. doi:10.4324/9781315717937
• Fumagalli, Roberto, 2016, “Why We Cannot Learn from MinimalModels”,Erkenntnis, 81(3): 433–455.doi:10.1007/s10670-015-9749-7
• Gelfert, Axel, 2016,How to Do Science with Models: APhilosophical Primer, (SpringerBriefs in Philosophy), Cham:Springer International Publishing. doi:10.1007/978-3-319-27954-1
• Ghorbani, Amineh, Gerard Dijkema, and Noortje Schrauwen, 2015,“Structuring Qualitative Data for Agent-Based Modelling”,Journal of Artificial Societies and Social Simulation, 18(1):article 2. doi:10.18564/jasss.2573
• Gilbert, Margaret, 2000,Sociality and Responsibility: NewEssays in Plural Subject Theory, Lanham, MD/Oxford: Rowman &Littlefield Publishers.
• Gilbert, Nigel, 1997, “A Simulation of the Structure ofAcademic Science”,Sociological Research Online, 2(2):91–105. doi:10.5153/sro.85
• Gilbert, G. Nigel and Klaus G. Troitzsch, 2005,Simulation forthe Social Scientist, second edition, Maidenhead/New York: OpenUniversity Press.
• Goldman, Alvin I., 2010, “Epistemic Relativism andReasonable Disagreement”, in Feldman and Warfield 2010:187–215. doi:10.1093/acprof:oso/9780199226078.003.0009
• –––, 1999,Knowledge in a Social World,Oxford: Clarendon Press. doi:10.1093/0198238207.001.0001
• Goldman, Alvin I. and Moshe Shaked, 1991, “An Economic Modelof Scientific Activity and Truth Acquisition”,PhilosophicalStudies, 63(1): 31–55. doi:10.1007/BF00375996
• Graebner, Claudius, 2018, “How to Relate Models to Reality?An Epistemological Framework for the Validation and Verification ofComputational Models”,Journal of Artificial Societies andSocial Simulation, 21(3): article 8. doi:10.18564/jasss.3772
• Granovetter, Mark S., 1973, “The Strength of WeakTies”,American Journal of Sociology, 78(6):1360–1380. doi:10.1086/225469
• Grasswick, Heidi Elizabeth (ed.), 2011,Feminist Epistemologyand Philosophy of Science: Power in Knowledge, (FeministPhilosophy Collection), Dordrecht/New York: Springer.doi:10.1007/978-1-4020-6835-5
• Grim, Patrick, 2009, “Threshold Phenomena in EpistemicNetworks.”, inComplex Adaptive Systems and the ThresholdEffect: Views from the Natural and Social Sciences, (Papers fromthe 2009 AAAI Fall Symposium 3), 53–60. [Grim 2009 available online]
• Grim, Patrick, Gary Mar, and Paul St. Denis, 1998,ThePhilosophical Computer: Exploratory Essays in Philosophical ComputerModeling, with the Group for Logic and Formal Semantics,Cambridge, MA: MIT Press.
• Grim, Patrick, Daniel J. Singer, Aaron Bramson, Bennett Holman,Sean McGeehan, and William J. Berger, 2019, “Diversity, Ability,and Expertise in Epistemic Communities”,Philosophy ofScience, 86(1): 98–123. doi:10.1086/701070
• Grim, Patrick, Daniel J. Singer, Steven Fisher, Aaron Bramson,William J. Berger, Christopher Reade, Carissa Flocken, and Adam Sales,2013, “Scientific Networks on Data Landscapes: QuestionDifficulty, Epistemic Success, and Convergence”,Episteme, 10(4): 441–464. doi:10.1017/epi.2013.36
• Grimm, Volker and Steven F. Railsback, 2005,Individual-BasedModeling and Ecology, (Princeton Series in Theoretical andComputational Biology), Princeton, NJ: Princeton University Press.doi:10.1515/9781400850624
• Grüne-Yanoff, Till, 2009, “Learning from MinimalEconomic Models”,Erkenntnis, 70(1): 81–99.doi:10.1007/s10670-008-9138-6
• Haraway, Donna Jeanne, 1989,Primate Visions: Gender, Race,and Nature in the World of Modern Science, New York/London:Routledge. doi:10.4324/9780203421918
• Harnagel, Audrey, 2019, “A Mid-Level Approach to ModelingScientific Communities”,Studies in History and Philosophyof Science Part A, 76: 49–59.doi:10.1016/j.shpsa.2018.12.010
• Hartmann, Stephan, Chiara Lisciandra, and Edouard Machery, 2013,“Editorial: Formal Epistemology Meets ExperimentalPhilosophy”,Synthese, 190(8): 1333–1335.doi:10.1007/s11229-013-0269-1
• Hartmann, Stephan, Carlo Martini, and Jan Sprenger, 2009,“Consensual Decision-Making Among Epistemic Peers”,Episteme, 6(2): 110–129.doi:10.3366/E1742360009000598
• Heesen, Remco, 2019, “The Credit Incentive to Be aMaverick”,Studies in History and Philosophy of Science PartA, 76: 5–12. doi:10.1016/j.shpsa.2018.11.007
• Hegselmann, Rainer, 2017, “Thomas C. Schelling and James M.Sakoda: The Intellectual, Technical, and Social History of aModel”,Journal of Artificial Societies and SocialSimulation, 20(3): article 15. doi:10.18564/jasss.3511
• Hegselmann, Rainer and Ulrich Krause, 2002, “OpinionDynamics and Bounded Confidence Models, Analysis, andSimulation”,Journal of Artificial Societies and SocialSimulation, 5(3). [Hegselmann and Krause 2002 available online]
• –––, 2005, “Opinion Dynamics Driven byVarious Ways of Averaging”,Computational Economics,25(4): 381–405. doi:10.1007/s10614-005-6296-3
• –––, 2006, “Truth and Cognitive Divisionof Labor: First Steps towards a Computer Aided SocialEpistemology”,Journal of Artificial Societies and SocialSimulation, 9(3): article 10. [Hegselmann and Krause 2006 available online]
• Hempel, Carl G., 1965,Aspects of Scientific Explanation: AndOther Essays in the Philosophy of Science, New York: FreePress.
• Holman, Bennett and Justin P. Bruner, 2015, “The Problem ofIntransigently Biased Agents”,Philosophy of Science,82(5): 956–968. doi:10.1086/683344
• –––, 2017, “Experimentation by IndustrialSelection”,Philosophy of Science: 1008–1019.doi:10.1086/694037
• Holman, Bennett and Kevin C. Elliott, 2018, “The Promise andPerils of Industry-Funded Science”,Philosophy Compass,13(11): e12544. doi:10.1111/phc3.12544
• Holman, Bennett and Torsten Wilholt, 2022, “The NewDemarcation Problem”,Studies in History and Philosophy ofScience, 91: 211–220. doi:10.1016/j.shpsa.2021.11.011
• Hong, Lu and Scott E. Page, 2004, “Groups of Diverse ProblemSolvers Can Outperform Groups of High-Ability Problem Solvers”,Proceedings of the National Academy of Sciences, 101(46):16385–16389. doi:10.1073/pnas.0403723101
• Houkes, Wybo, Dunja Šešelja, and Krist Vaesen,forthcoming, “Robustness Analysis”, inThe RoutledgeHandbook of Philosophy of Scientific Modeling, Natalia Carrillo,Tarja Knuuttila, and Ram Koskinen (eds.). [Houkes, Šešelja, and Vaesen forthcoming available online]
• Hull, David L., 1978, “Altruism in Science: ASociobiological Model of Co-Operative Behaviour amongScientists”,Animal Behaviour, 26: 685–697.doi:10.1016/0003-3472(78)90135-5
• –––, 1988,Science as a Process: AnEvolutionary Account of the Social and Conceptual Development ofScience, (Science and Its Conceptual Foundations), Chicago, IL:University of Chicago Press.
• Jackson, Matthew O. and Asher Wolinsky, 1996, “A StrategicModel of Social and Economic Networks”,Journal of EconomicTheory, 71(1): 44–74. doi:10.1006/jeth.1996.0108
• Kauffman, Stuart A., 1993,The Origins of Order: SelfOrganization and Selection in Evolution, New York: OxfordUniversity Press.
• Kauffman, Stuart and Simon Levin, 1987, “Towards a GeneralTheory of Adaptive Walks on Rugged Landscapes”,Journal ofTheoretical Biology, 128(1): 11–45.doi:10.1016/S0022-5193(87)80029-2
• Kavak, Hamdi, Jose J. Padilla, Christopher J. Lynch, and Saikou Y.Diallo, 2018, “Big Data, Agents, and Machine Learning: Towards aData-Driven Agent-Based Modeling Approach”, inProceedingsof the Annual Simulation Symposium (ANSS ’18), San Diego,CA, USA: Society for Computer Simulation International, article 12 (12pages).
• Kelly, Thomas, 2010 [2011], “Peer Disagreement andHigher‐Order Evidence”, in Feldman and Warfield 2010:111–174. An abridged version was published inSocialEpistemology: Essential Readings, Alvin I. Goldman and DennisWhitcomb (eds.), Oxford/New York: Oxford University Press, 2011,183–217. doi:10.1093/acprof:oso/9780199226078.003.0007
• Kelp, Christoph and Igor Douven, 2012, “Sustaining aRational Disagreement”, inEPSA Philosophy of Science:Amsterdam 2009, Henk W. De Regt, Stephan Hartmann, and SamirOkasha (eds.), Dordrecht: Springer Netherlands, 101–110.doi:10.1007/978-94-007-2404-4_10
• Kitcher, Philip, 1990, “The Division of CognitiveLabor”,The Journal of Philosophy, 87(1): 5–22.doi:10.2307/2026796
• –––, 1993,The Advancement of Science:Science without Legend, Objectivity without Illusions, New York:Oxford University Press. doi:10.1093/0195096533.001.0001
• Klein, Dominik, Johannes Marx, and Simon Scheller, 2020,“Rationality in Context: On Inequality and the EpistemicProblems of Maximizing Expected Utility”,Synthese,197(1): 209–232. doi:10.1007/s11229-018-1773-0
• Konigsberg, Amir, 2013, “The Problem with Uniform Solutionsto Peer Disagreement”,Theoria, 79(2): 96–126.doi:10.1111/j.1755-2567.2012.01149.x
• Kopecky, Felix, 2022, “Arguments as Drivers of IssuePolarisation in Debates Among Artificial Agents”,Journal ofArtificial Societies and Social Simulation, 25(1): article 4.doi:10.18564/jasss.4767
• Kuhlmann, Meinard, 2021, “On the Exploratory Function ofAgent-Based Modelling”,Perspectives on Science, 29(4):510–536. doi:10.1162/posc_a_00381
• Kuhn, Thomas S., 1962,Structure of ScientificRevolutions, Chicago: The University of Chicago Press.
• –––, 1977,The Essential Tension: SelectedStudies in Scientific Tradition and Change, Chicago: Universityof Chicago Press.
• Kukla, Rebecca, 2012, “‘Author TBD’: RadicalCollaboration in Contemporary Biomedical Research”,Philosophy of Science, 79(5): 845–858.doi:10.1086/668042
• Kummerfeld, Erich and Kevin J. S. Zollman, 2016,“Conservatism and the Scientific State of Nature”,TheBritish Journal for the Philosophy of Science, 67(4):1057–1076. doi:10.1093/bjps/axv013
• Kuorikoski, Jaakko, Aki Lehtinen, and Caterina Marchionni, 2012,“Robustness Analysis Disclaimer: Please Read the Manual beforeUse!”,Biology & Philosophy, 27(6): 891–902.doi:10.1007/s10539-012-9329-z
• Laudan, Larry, 1977,Progress and Its Problems: Toward aTheory of Scientific Growth, Berkeley, CA: University ofCalifornia Press.
• Lazer, David and Allan Friedman, 2007, “The NetworkStructure of Exploration and Exploitation”,AdministrativeScience Quarterly, 52(4): 667–694.doi:10.2189/asqu.52.4.667
• Lehrer, Keith and Carl Wagner, 1981,Rational Consensus inScience and Society, (Philosophical Studies Series 24),Dordrecht/Boston: D. Reidel Publishing Company.doi:10.1007/978-94-009-8520-9
• Lewontin, Richard C., 1961, “Evolution and the Theory ofGames”,Journal of Theoretical Biology, 1(3):382–403. doi:10.1016/0022-5193(61)90038-8
• Longino, Helen E., 1990,Science as Social Knowledge: Valuesand Objectivity in Scientific Inquiry, Princeton, NJ: PrincetonUniversity Press.
• –––, 2002,The Fate of Knowledge,Princeton, NJ: Princeton University Press.
• –––, 2022, “What’s Social aboutSocial Epistemology?”,The Journal of Philosophy,119(4): 169–195. doi:10.5840/jphil2022119413
• Mäki, Uskali, 2005, “Economic Epistemology: Hopes andHorrors”,Episteme, 1(3): 211–222.doi:10.3366/epi.2004.1.3.211
• March, James G., 1991, “Exploration and Exploitation inOrganizational Learning”,Organization Science, 2(1):71–87. doi:10.1287/orsc.2.1.71
• Martini, Carlo and Manuela Fernández Pinto, 2017,“Modeling the Social Organization of Science: Chasing Complexitythrough Simulations”,European Journal for Philosophy ofScience, 7(2): 221–238. doi:10.1007/s13194-016-0153-1
• Mason, Winter A., Andy Jones, and Robert L. Goldstone, 2008,“Propagation of Innovations in Networked Groups.”,Experimental Psychology: General, 137(3): 422–433.doi:10.1037/a0012798
• Mason, Winter and Duncan J. Watts, 2012, “CollaborativeLearning in Networks”,Proceedings of the National Academyof Sciences, 109(3): 764–769.doi:10.1073/pnas.1110069108
• Maynard Smith, John, 1982,Evolution and the Theory ofGames, Cambridge/New York: Cambridge University Press.doi:10.1017/CBO9780511806292
• Mayo-Wilson, Conor, 2014, “Reliability of Testimonial Normsin Scientific Communities”,Synthese, 191(1):55–78. doi:10.1007/s11229-013-0320-2
• Mayo-Wilson, Conor and Kevin J. S. Zollman, 2021, “TheComputational Philosophy: Simulation as a Core PhilosophicalMethod”,Synthese, 199(1–2): 3647–3673.doi:10.1007/s11229-020-02950-3
• Merdes, Christoph, 2021, “Strategy and the Pursuit ofTruth”,Synthese, 198(1): 117–138.doi:10.1007/s11229-018-01985-x
• Merdes, Christoph, Momme Von Sydow, and Ulrike Hahn, 2021,“Formal Models of Source Reliability”,Synthese,198(S23): 5773–5801. doi:10.1007/s11229-020-02595-2
• Merton, Robert King, 1973,The Sociology of Science:Theoretical and Empirical Investigations, Chicago: University ofChicago Press.
• Michelini, Matteo, Javier Osorio, Wybo Houkes, DunjaŠešelja, and Christian Straßer, forthcoming,“Scientific Disagreements and the Diagnosticity of Evidence: HowToo Much Data May Lead to Polarization”,Journal ofArtificial Societies and Social Simulation. [Michelini et al. preprint available online]
• Mill, John Stuart, 1859,On Liberty, London: John W.Parker and son.
• Mohseni, Aydin, Cailin O’Connor, and Hannah Rubin, 2021,“On the Emergence of Minority Disadvantage: Testing the CulturalRed King Hypothesis”,Synthese, 198(6):5599–5621. doi:10.1007/s11229-019-02424-1
• Muldoon, Ryan and Michael Weisberg, 2011, “Robustness andIdealization in Models of Cognitive Labor”,Synthese,183(2): 161–174. doi:10.1007/s11229-010-9757-8
• Nash, John F. Jr, 1950, “The Bargaining Problem”,Econometrica, 18(2): 155–162. doi:10.2307/1907266
• Nguyen, James, 2020, “It’s Not a Game: AccurateRepresentation with Toy Models”,The British Journal for thePhilosophy of Science, 71(3): 1013–1041.doi:10.1093/bjps/axz010
• O’Connor, Cailin, 2017, “The Cultural Red KingEffect”,The Journal of Mathematical Sociology, 41(3):155–171. doi:10.1080/0022250X.2017.1335723
• –––, 2019, “The Natural Selection ofConservative Science”,Studies in History and Philosophy ofScience Part A, 76: 24–29.doi:10.1016/j.shpsa.2018.09.007
• O’Connor, Cailin and Justin Bruner, 2019, “Dynamicsand Diversity in Epistemic Communities”,Erkenntnis,84(1): 101–119. doi:10.1007/s10670-017-9950-y
• O’Connor, Cailin and James Owen Weatherall, 2018,“Scientific Polarization”,European Journal forPhilosophy of Science, 8(3): 855–875.doi:10.1007/s13194-018-0213-9
• Olsson, Erik J., 2011, “A Simulation Approach to VeritisticSocial Epistemology”,Episteme, 8(2): 127–143.doi:10.3366/epi.2011.0012
• –––, 2013, “A Bayesian Simulation Model ofGroup Deliberation and Polarization”, inBayesianArgumentation: The Practical Side of Probability, Frank Zenker(ed.), (Synthese Library 362), Dordrecht: Springer Netherlands,113–133. doi:10.1007/978-94-007-5357-0_6
• Oreskes, Naomi and Erik M. Conway, 2010,Merchants of Doubt:How a Handful of Scientists Obscured the Truth on Issues from TobaccoSmoke to Global Warming, New York: Bloomsbury Press.
• Page, Scott E., 2017,The Diversity Bonus: How Great Teams PayOff in the Knowledge Economy, Princeton/Oxford: PrincetonUniversity Press.
• –––, 2018,The Model Thinker: What You Needto Know to Make Data Work for You, New York: Basic Books.
• Payette, Nicolas, 2011, “Agent-Based Models ofScience”, inModels of Science Dynamics. UnderstandingComplex Systems, Andrea Scharnhorst, Katy Börner, and Petervan den Besselaar (eds.), Berlin/Heidelberg: Springer, 127–157.doi:10.1007/978-3-642-23068-4_4
• Perović, Slobodan, Sandro Radovanović, VlastaSikimić, and Andrea Berber, 2016, “Optimal Research TeamComposition: Data Envelopment Analysis of Fermilab Experiments”,Scientometrics, 108(1): 83–111.doi:10.1007/s11192-016-1947-9
• Pesonen, Renne, 2022, “Argumentation, Cognition, and theEpistemic Benefits of Cognitive Diversity”,Synthese,200(4): article 295. doi:10.1007/s11229-022-03786-9
• Peters, Uwe, 2021, “Illegitimate Values, Confirmation Bias,and Mandevillian Cognition in Science”,The British Journalfor the Philosophy of Science, 72(4): 1061–1081.doi:10.1093/bjps/axy079
• Politi, Vincenzo, 2021, “Formal Models of the ScientificCommunity and the Value-Ladenness of Science”,EuropeanJournal for Philosophy of Science, 11(4): article 97.doi:10.1007/s13194-021-00418-w
• Pöyhönen, Samuli, 2017, “Value of CognitiveDiversity in Science”,Synthese, 194(11):4519–4540. doi:10.1007/s11229-016-1147-4
• Reijula, Samuli and Jaakko Kuorikoski, 2021, “TheDiversity-Ability Trade-Off in Scientific Problem Solving”,Philosophy of Science, 88(5): 894–905.doi:10.1086/714938
• Reiss, Julian, 2012, “The Explanation Paradox”,Journal of Economic Methodology, 19(1): 43–62.doi:10.1080/1350178X.2012.661069
• Reutlinger, Alexander, Dominik Hangleiter, and Stephan Hartmann,2018, “Understanding (with) Toy Models”,The BritishJournal for the Philosophy of Science, 69(4): 1069–1099.doi:10.1093/bjps/axx005
• Reydon, Thomas A.C., 2012, “How-Possibly Explanations asGenuine Explanations and Helpful Heuristics: A Comment onForber”,Studies in History and Philosophy of Science PartC: Studies in History and Philosophy of Biological and BiomedicalSciences, 43(1): 302–310.doi:10.1016/j.shpsc.2011.10.015
• Riegler, Alexander and Igor Douven, 2009, “Extending theHegselmann–Krause Model III: From Single Beliefs to ComplexBelief States”,Episteme, 6(2): 145–163.doi:10.3366/E1742360009000616
• Robbins, Herbert, 1952, “Some Aspects of the SequentialDesign of Experiments”,Bulletin of the AmericanMathematical Society, 58(5): 527–535.
• Rolin, Kristina, 2015, “Values in Science: The Case ofScientific Collaboration”,Philosophy of Science,82(2): 157–177. doi:10.1086/680522
• –––, 2019, “The Epistemic Significance ofDiversity”, in Fricker, Graham, Henderson, & Pedersen 2019:158–166.
• Rosenstock, Sarita, Justin Bruner, and Cailin O’Connor,2017, “In Epistemic Networks, Is Less Really More?”,Philosophy of Science, 84(2): 234–252.doi:10.1086/690717
• Rubin, Hannah, 2022, “Structural causes of citationgaps”,Philosophical Studies, 179(7):2323–2345.
• Rubin, Hannah and Cailin O’Connor, 2018,“Discrimination and Collaboration in Science”,Philosophy of Science, 85(3): 380–402.doi:10.1086/697744
• Rubin, Hannah and Mike D. Schneider, 2021, “Priority andPrivilege in Scientific Discovery”,Studies in History andPhilosophy of Science Part A, 89: 202–211.doi:10.1016/j.shpsa.2021.08.005
• Rueger, Alexander, 1996, “Risk and Diversification in TheoryChoice”,Synthese, 109(2): 263–280.doi:10.1007/BF00413769
• Sakoda, James M., 1971, “The Checkerboard Model of SocialInteraction”,The Journal of Mathematical Sociology,1(1): 119–132. doi:10.1080/0022250X.1971.9989791
• Saltelli, Andrea, Marco Ratto, Terry Andres, Francesca Campolongo,Jessica Cariboni, Debora Gatelli, Michaela Saisana, and StefanoTarantola, 2008,Global Sensitivity Analysis: The Primer,Chichester: John Wiley & Sons.
• Santana, Carlos, 2018, “Why Not All Evidence Is ScientificEvidence”,Episteme, 15(2): 209–227.doi:10.1017/epi.2017.3
• Santana, Carlos, 2021, “Let’sNot Agree toDisagree: The Role of Strategic Disagreement in Science”,Synthese, 198(S25): 6159–6177.doi:10.1007/s11229-019-02202-z
• Schelling, Thomas C., 1971, “Dynamic Models ofSegregation”,The Journal of Mathematical Sociology,1(2): 143–186. doi:10.1080/0022250X.1971.9989794
• Schurz, Gerhard, 2009, “Meta-Induction and SocialEpistemology: Computer Simulations of Prediction Games”,Episteme, 6(2): 200–220.doi:10.3366/E1742360009000641
• –––, 2012, “Meta-Induction in EpistemicNetworks and the Social Spread of Knowledge”,Episteme,9(2): 151–170. doi:10.1017/epi.2012.6
• Šešelja, Dunja, 2021a, “Exploring ScientificInquiry via Agent-Based Modelling”,Perspectives onScience, 29(4): 537–557. doi:10.1162/posc_a_00382
• –––, 2021b, “Some Lessons from Simulationsof Scientific Disagreements”,Synthese, 198(Suppl 25):6143–6158. doi:10.1007/s11229-019-02182-0
• –––, 2022a, “Agent‐based Models ofScientific Interaction”,Philosophy Compass, 17(7):e12855. doi:10.1111/phc3.12855
• –––, 2022b, “What Kind of Explanations DoWe Get From Agent-Based Models of Scientific Inquiry?”, inProceedings of the 16th International Congress of Logic,Methodology and Philosophy of Science and Technology, HanneAndersen, Tomávs Marvan, Hasok Chang, Benedikt Löwe, andIvo Pezlar (eds.), London: College Publications.
• Sikimić, Vlasta and Ole Herud-Sikimić, 2022,“Modelling Efficient Team Structures in Biology”,Journal of Logic and Computation, 32(6): 1109–1128.doi:10.1093/logcom/exac021
• Singer, Daniel J., 2019, “Diversity, Not Randomness, TrumpsAbility”,Philosophy of Science, 86(1): 178–191.doi:10.1086/701074
• Singer, Daniel J., Aaron Bramson, Patrick Grim, Bennett Holman,Jiin Jung, Karen Kovaka, Anika Ranginani, and William J. Berger, 2019,“Rational Social and Political Polarization”,Philosophical Studies, 176(9): 2243–2267.doi:10.1007/s11098-018-1124-5
• Skyrms, Brian, 1990,The Dynamics of RationalDeliberation, Cambridge, MA: Harvard University Press.
• –––, 1996,Evolution of the SocialContract, Cambridge/New York: Cambridge University Press.
• Smaldino, Paul E. and Richard McElreath, 2016, “The NaturalSelection of Bad Science”,Royal Society Open Science,3(9): 160384. doi:10.1098/rsos.160384
• Solomon, Miriam, 2006, “Groupthink versusTheWisdom of Crowds: The Social Epistemology of Deliberation andDissent”,The Southern Journal of Philosophy, 44(S1):28–42. doi:10.1111/j.2041-6962.2006.tb00028.x
• Squazzoni, Flaminio (ed.), 2009,Epistemological Aspects ofComputer Simulation in the Social Sciences, (Lecture Notes inComputer Science 5466), Berlin/Heidelberg: Springer.doi:10.1007/978-3-642-01109-2
• Steel, Daniel, Sina Fazelpour, Kinley Gillette, Bianca Crewe, andMichael Burgess, 2018, “Multiple Diversity Concepts and TheirEthical-Epistemic Implications”,European Journal forPhilosophy of Science, 8(3): 761–780.doi:10.1007/s13194-018-0209-5
• Strevens, Michael, 2003, “The Role of the Priority Rule inScience”:, Journal of Philosophy, 100(2): 55–79.doi:10.5840/jphil2003100224
• –––, 2011, “Economic Approaches toUnderstanding Scientific Norms”,Episteme, 8(2):184–200. doi:10.3366/epi.2011.0015
• Thagard, Paul, 1988,Computational Philosophy of Science,Cambridge, MA: MIT Press. doi:10.7551/mitpress/1968.001.0001
• Thicke, Michael, 2020, “Evaluating Formal Models ofScience”,Journal for General Philosophy of Science,51(2): 315–335. doi:10.1007/s10838-018-9440-1
• Thoma, Johanna, 2015, “The Epistemic Division of LaborRevisited”,Philosophy of Science, 82(3):454–472. doi:10.1086/681768
• Tiokhin, Leonid, Minhua Yan, and Thomas J. H. Morgan, 2021,“Competition for Priority Harms the Reliability of Science, butReforms Can Help”,Nature Human Behaviour, 5(7):857–867. doi:10.1038/s41562-020-01040-1
• Ventura, Rafael, 2023, “Structural Inequality inCollaboration Networks”,Philosophy of Science, 90(2):336–353. doi:10.1017/psa.2022.73
• Verreault-Julien, Philippe, 2019, “How Could Models PossiblyProvide How-Possibly Explanations?”,Studies in History andPhilosophy of Science Part A, 73: 22–33.doi:10.1016/j.shpsa.2018.06.008
• Wagenknecht, Susann, 2015, “Facing the Incompleteness ofEpistemic Trust: Managing Dependence in Scientific Practice”,Social Epistemology, 29(2): 160–184.doi:10.1080/02691728.2013.794872
• Wagner, Elliott and Jonathan Herington, 2021, “Agent-BasedModels of Dual-Use Research Restrictions”,The BritishJournal for the Philosophy of Science, 72(2): 377–399.doi:10.1093/bjps/axz017
• Weatherall, James Owen and Cailin O’Connor, 2021a,“Conformity in Scientific Networks”,Synthese,198(8): 7257–7278. doi:10.1007/s11229-019-02520-2
• –––, 2021b, “Endogenous EpistemicFactionalization”,Synthese, 198(S25): 6179–6200.doi:10.1007/s11229-020-02675-3
• Weatherall, James Owen, Cailin O’Connor, and Justin P.Bruner, 2020, “How to Beat Science and Influence People:Policymakers and Propaganda in Epistemic Networks”,TheBritish Journal for the Philosophy of Science, 71(4):1157–1186. doi:10.1093/bjps/axy062
• Weisberg, Michael, 2013,Simulation and Similarity: UsingModels to Understand the World, (Oxford Studies in Philosophy ofScience), Oxford/New York: Oxford University Press.
• Weisberg, Michael and Ryan Muldoon, 2009, “EpistemicLandscapes and the Division of Cognitive Labor”,Philosophyof Science, 76(2): 225–252. doi:10.1086/644786
• Wray, K. Brad, 2002, “The Epistemic Significance ofCollaborative Research”,Philosophy of Science, 69(1):150–168. doi:10.1086/338946
• –––, 2007, “Who Has ScientificKnowledge?”,Social Epistemology, 21(3): 337–347.doi:10.1080/02691720701674288
• Wright, Sewall, 1932, “The Roles of Mutation, Inbreeding,Crossbreeding, and Selection in Evolution”,Proceedings ofthe Sixth International Congress on Genetics, 1:356–366.
• Wu, Jingyi, 2023, “Epistemic Advantage on the Margin: ANetwork Standpoint Epistemology”,Philosophy andPhenomenological Research, 106(3): 755–777.doi:10.1111/phpr.12895
• Wylie, Alison, 1992, “The Interplay of EvidentialConstraints and Political Interests: Recent Archaeological Research onGender”,American Antiquity, 57(1): 15–35.doi:10.2307/2694833
• –––, 2002,Thinking from Things: Essays inthe Philosophy of Archaeology, Berkeley, CA: University ofCalifornia Press.
• –––, 2003, “Why Standpoint Matters”,inScience and Other Cultures: Issues in Philosophies of Scienceand Technology, Robert Figueroa and Sandra Harding (eds.), NewYork/London: Routledge, 26–48.
• Ylikoski, Petri, 2014, “Agent-Based Simulation andSociological Understanding”,Perspectives on Science,22(3): 318–335. doi:10.1162/POSC_a_00136
• Ylikoski, Petri and N. Emrah Aydinonat, 2014, “Understandingwith Theoretical Models”,Journal of EconomicMethodology, 21(1): 19–36.doi:10.1080/1350178X.2014.886470
• Zamora Bonilla, Jesús P., 1999, “The ElementaryEconomics of Scientific Consensus”,Theoria: AnInternational Journal for Theory, History and Foundations ofScience, 14(3): 461–488.
• Zhang, “Synergistic Integration Between Machine Learning andAgent-Based Modeling: A Multidisciplinary Review”,IEEETransactions on Neural Networks and Learning Systems, 34(5):2170–2190. doi:10.1109/TNNLS.2021.3106777
• Zollman, Kevin J. S., 2007, “The Communication Structure ofEpistemic Communities”,Philosophy of Science, 74(5):574–587. doi:10.1086/525605
• –––, 2009, “Optimal PublishingStrategies”,Episteme, 6(2): 185–199.doi:10.3366/E174236000900063X
• –––, 2010, “The Epistemic Benefit ofTransient Diversity”,Erkenntnis, 72(1): 17–35.doi:10.1007/s10670-009-9194-6
• –––, 2013, “Network Epistemology:Communication in Epistemic Communities”,PhilosophyCompass, 8(1): 15–27.doi:10.1111/j.1747-9991.2012.00534.x
• –––, 2017, “Learning toCollaborate”, in Boyer-Kassem, Mayo-Wilson, and Weisberg 2017:65–77 (ch. 3).
• –––, 2018, “The Credit Economy and theEconomic Rationality of Science”,The Journal ofPhilosophy, 115(1): 5–33. doi:10.5840/jphil201811511

Academic Tools

	How to cite this entry.
	Preview the PDF version of this entry at theFriends of the SEP Society.
	Look up topics and thinkers related to this entry at the Internet Philosophy Ontology Project (InPhO).
	Enhanced bibliography for this entryatPhilPapers, with links to its database.

Other Internet Resources

• Angere, Staffan, 2010, “Knowledge in a Social Network”, Lund University, unpublished manuscript.
• the CoMSES Model Library, a digital repository of ABMs
• NetLogo

Related Entries

computational philosophy |epistemology, formal |epistemology: social |game theory |game theory: evolutionary |models in science |scientific knowledge: social dimensions of

Acknowledgments

Many thanks to Wybo Houkes, Matteo Michelini, Samuli Reijula,Christian Straßer, Krist Vaesen and Soong Yoo for valuablediscussions and comments on earlier drafts. I am also grateful to theanonymous reviewer for the helpful comments and suggestions. Theresearch for this paper is supported by the DeutscheForschungsgemeinschaft (DFG, German Research Foundation)—projectnumber 426833574.