epistemic democrats can learn from
wild honeybees
Johanna (Hanne) M. van Beuningen 11750294
1st February 2021 Supervisor: Paul Raekstad Second reader: Enzo Rossi
Democratic Theory
Thesis future planet studies
Major political science
A honeybee colony, both a society and a superorganism
Abstract
This thesis presents the possibility for looking at Honeybee swarms in order to refine the epistemic benefit of democracy. Epistemic democracy by Helene Landemore argues for the importance of cognitive diversity to make smarter decisions. However it has trouble dealing with pluralism and the cognitive limitation of humans within the deliberation phase. The nest selection process of a honeybee swarm can function as a model for human societies because of the scope and the complexity of the swarm. By looking at the swarm and apply system theory, we can learn to approach a collective rational decision. Hereby we can overcome human cognitive limitations alike the wild honeybee.
Introduction
Biomimicry can be defined as the process of looking at strategies found in nature (bio) in order to mimic it in human design (mimicry) (Biomimicry Institute, 2020). Using biology as a design tool originates from the industry of product design. As an example, the front of the high-speed Shinkansen train in Japan was designed looking at the beak of a kingfisher and sharkskin was used as inspiration for low resistance swimming suits (Ibid.). In the same field, social biomimicry, looking at animal behaviour to inspire human social relations (especially in relation to insects such as bees and ants) has increased over the past years (Fewell, 2015). Insects as opposed to other animals live together in large, complex communities, which make them suitable for comparison to social and political theories (ibid.). This thesis will focus on the honeybee colony and its way of collecting knowledge. Seeley (2010) observed collective decision-making amongst hundreds of bee scouts that went out to look for a potential new nest. Every scout bee flies out in search of such a new place and reports back to the other scouts with an informational dance on the surface of the swarm, the process continues for one or two days with a decision for the best new nest spot at the end. The process where the honeybees gather information and make a collective decision without an executive order has a similar structure to epistemic democracy. Epistemic democrats argue for a collective intelligence, a functional Legitimation of democracy on the account of the group being smarter than any individual (Landemore, 2012a). This paper investigates how these decision-making processes might enrich the field of epistemic democracy. This results in a research question “What can epistemic democrats learn from honeybee societies? The question will be researched and answered in the following manner. First, the foundation of epistemic democracy will be discussed, with special attention to Helene Landemore’s theory of epistemic democracy, which theory will be prominent throughout this paper. Second, a short summary of the honeybee colony will describe the honeybee swarm and its collective decision-making mechanisms, based mostly on the research of behavioural biologist Thomas Seeley. As both Seeley and Landemore use terminology such as democracy, common interest and intelligence in different forms, it is important to avoid semantic ambiguity. The theory of honeybee decision-making is therefore converted so that it is coherent with the vocabulary of epistemic democrats. Following, the differences and similarities within each part of the decision-making (aggregation of options, distribution of options and conclusion)
are examined. Lastly, system theory is introduced and functions as a framework to portray insights from the honeybee nest search.
First of all, according to Held (2019) democracy is defined as the rule of the people and is built upon the presupposition of self-government of that people but has many different forms. The fundament for this form of government comes from the conception that a person is sovereign and therefore the sovereignty of a state lies with its people. Such people can institutionalize the use of this sovereignty by forming a government. In practice it means that the people have the right to voice their opinion and influence state affairs often though some form of voting procedure (Held, 2019).
Democracy originates from old Athens where the people were gathered roughly forty times a year to vote on public matters (Held, 2019). These ‘people’ only included the free men of Athenian descent, yet this was already significantly different from the then common aristocracies and autocracies (ibid.). Aristotle furthermore argued that democracy should consist of deliberation of all citizens. He did not argue this out of fairness but for its instrumental benefits, as democratic decisions are better in line with the interests of the people than non-democratic ones (Aristotle, 1998 129b12-20). There has been much contestation of the different conceptions of citizenship, authority and the relationship between the two, yet this will not be a central topic. Rather, this paper focuses on the discussion of the democracy’s legitimacy.
Furthermore, democracy can be seen as intrinsically good but more often it is understood as a means to achieve liberty, equality and good decision-making or justice amongst others (Held, 2019; Landemore, 2012a). Amongst the arguments, for the legitimacy of democracy there is a division between instrumental and non-instrumental arguments. Non-non-instrumental arguments understand democracy as intrinsically good since it is a way of treating people as equals (Singer, 1973). Others emphasize it ensures liberty as democracy enables a person to have control over their life (Gould, 1988).
An instrumental argument for democracy is that a government has an incentive to listen to the lower classes in society as they have political power, even when no other
power is available to them (Mill, 1861). Both Rousseau and Mill have argued for the benefits of democracy on individual character development because it makes people more autonomous compared to aristocracies and autocracies (Mill, 1862; Elster, 2002). The instrumental argument presented by Aristotle is the foundation for epistemic democracy: He stated that the governing body ought to include everyone because it would lead to better decisions (Aristotle, 1998 129b12-20). Landemore (2012) sees the epistemic value as a justification for democracy. She argues that under set conditions any group is able to make better decisions than even the smartest individuals within it. This gives democracy an epistemic advantage over autocrats and oligarchs (Landemore, 2014). The epistemic advantage of the many is used by many political theorists but is most prominent amongst epistemic democrats (Mansbridge et al., 2012). This argument could be considered an empty conception of democracy because of the absence of normative claims and ideals. Yet it is also an argument that reaches beyond the border of democratic government, since the instrument of democratic decision-making becomes something that might appeal to those not satisfied with the ideational arguments for democracy (Landemore, 2012a). It becomes an intellectual argument for democracy independently of any moral arguments for or against democracy. The focus on the quality of democracy to make decisions that are better aligned with a common good is usually perceived as epistemic democracy.
Epistemic Democracy
Joshua Cohen (1986) was the first to mention the definition of epistemic democracy; he characterized epistemic democracy as a form of popular democracy focussed on the will of the people. Cohen correctly recognizes several prejudices founded in epistemic democracy. He states that what he calls epistemic populism assumes three things within the voting procedure: (1) an independent standard of correctness, (2) people that vote according to what they think will lead to a common good and (3) decision making as an adjustment of beliefs. Additionally Cohen (1968) states that a critical factor within this theory is the acknowledgement of a common good.
Within the spectrum of theories of epistemic democracy there is a distinction between deliberative and non-deliberative epistemic democracies. Condorcet’s jury theorem is the most important non-deliberative theory of epistemic democracies
(Estlund et al, 1989). Condorcet mathematically argued for the epistemic benefit of numbers over cognitive ability of the few. His argument lays on the preposition that an individual has slightly more than fifty per cent chance to choose the option that leads to the desired outcome. An increase in individuals will exponentially increase the possibility of getting the correct outcome. Condorcet’s jury theorem advocates the importance of numbers for good decision-making.
Deliberative theories of epistemic democracy are more widespread since it is closer related to the more prominent school of deliberative democrats (Estlund, 2008; Landemore, 2012a). In practice, deliberative democracy can strongly resemble epistemic democracy, but the ideal and emphasis differ. Deliberative democrats emphasise the intrinsic value of deliberation regardless of the outcome as opposed to epistemic democrats, who see deliberation as a means (Estlund, 2008). However, both adjoining schools of thought emphasize the importance of the deliberation quality, within a democracy (Landemore, 2014; Mansbridge et al., 2012). In addition, there is discussion about the number of participants and whether or not these individuals have to differ significantly, and if so, what the differing aspect should be (Bohman, 2006; Landemore, 2012; Mill, 1861)
As mentioned earlier, Aristotle argued that within deliberation every citizen had to be included since it would lead to better decision-making. This, however, is unrealistic when we consider modern day societies both because of scope and pluralism. There seems to be a threshold from which the epistemic advantage is replaced by chaos. Amongst epistemic theorists the number is the first issue of contestation besides the critical qualities of the group. Mill (1861) is commonly seen as a deliberative democrat but argues for an epistemic benefit, only, when the full range of opinions is included. The deliberation therefore not necessarily has to include all citizens. Bohman (2006) criticizes Mill in his sole focus on opinions. Instead Bohman (2006) argues that diversity in opinions, perspectives and values ought to be included within the group of deliberators. These conditions are in Bohman’s opinion essential to provide more robustness to the process and especially to the outcome of the deliberation. The technical Diversity Trumps Ability Theorem likewise argues that a randomly chosen group of problem solvers can outperform a group of the brightest problem solvers because of its diversity (Hong & Page, 2004). Landemore builds
upon Hong and Pages (2004), Estlund (1989) and Mill’s (1861) research for her focus on cognitive diversity as a means to achieve better decision-making. She places her theoretical framework between the aggregative model and the deliberative model (Landemore, 2012a). The aggregative model perceives democracy as an aggregation of interests, yet Landemore is solely interested in the aggregation of judgements and predictions. The deliberative model perceives deliberation as the centre of democracy while Landemore perceives deliberation as one of the options in order to achieve collective wisdom (Landemore, 2012a).
Specifically, Helene Landemore’s (2013) notion of epistemic democracy appeals to the claim that the many are smarter than the individual because of the existence of what she calls collective intelligence. Her concept of collective intelligence is the essence of democratic reason necessary for an epistemic democracy. Individuals that unite and achieve collective intelligence are smarter than the smartest individuals within it and are thus able to make better decisions. Collective intelligence is the foundation for democracies’ epistemic advantage and is a function of average ability and cognitive diversity. She builds upon Hong and Page’s theorem (2004) that states that the diversity of the group is more importance than the intelligence of the individuals within such group. According to Landemore (2012a) cognitive diversity is preferable as opposed to other forms of diversity promoted by other democratic theorists. For a people to solve a political issue it is of utmost importance that various people interpret the case in distinct ways. Every person interprets the world in a different way and applies his or her internalized predictive model onto it. To clarify: Predictive models are formed over the full course of someone’s lifetime and represent an individuals’ model of what life entails. Likeminded individuals have similar predictive models; the more two individuals differ the harder it becomes to understand the other’s predictive model. Therefore, diversity is a quality of a group and the same intelligence is impossible to achieve as an individual (Landemore, 2012b).
In addition, Landemore (2012a, 2012b, 2013) distinguishes herself from theorists like Condorcet by expressing the importance of diversity within the decision-making group as opposed to the number of individuals within the group. However, when given the choice to rise the individual ability or the number of people taking a
decision the latter is considered more critical. This results in a hierarchy of group selection criteria; the most important is cognitive diversity because this is the essence of the epistemic advantage of democracy. The second criterion is numbers since it will naturally increase cognitive diversity. The least important of this ranking turns out to be individual ability, and is therefore considered not essential for an epistemic benefit (Landemore 2012a, 2012b, 2013).
Let me elaborate with an example of a decision-making group that has to increase their collective intelligence. In order to achieve maximal cognitive diversity one would have to do a full in depth popular inquiry, which is both expensive and time consuming making it irrational. A random sample would be the next best thing: Although a random sample would have sub-optimal ability, the diversity of the group is likely raised. The least favourable is the selection of the brightest individuals because this will only provide a limited range of predictive models: The brightest are often schooled in similar forms and thus have similar predictive models (Landemore, 2013).
Further, Landemore (2012a) determines three phases in democratic decision-making. She argues that in every stage cognitive diversity is more important than individual ability. The first stage is (1) the aggregation phase where problems and solutions are determined and aggregated. (2) The deliberation phase consists of the exchange of predicative models and knowledge and finally (3) the decision phase is where the actual decision is made. Deliberation can range from informative to persuasive and Landemore’s formulation of deliberation can be placed between the two as consideration. In order to let the decision-making be effective there is a threshold after which more people that are included in the deliberation will lead to chaos and does not provide any material insights anymore. In order to prevent chaos, Landemore proposes representation as a possible solution (Landemore, 2013). Yet, because of the superiority of cognitive diversity over ability she proposes selection by lottery as opposed to an election as we know it in today’s society. A selection by lot will naturally increase cognitive diversity. Yet there are barely any modern day examples of binding deliberative bodies selected by lot, perhaps because it is difficult to accept that diversity will lead to better decision-making and selection by lot can
seem dangerous or irresponsible (Landemore, 2014). This complicates the argument for Landemore.
The base of the epistemic argument for democracy could lie in Nietzsche's perspectivism, despite Nietzsche’s own aversion towards democracy. Nietzsche argues that knowledge is not based upon an absolute objectivity or metaphysical reality but is fragmented and dispersed into perspectives (Anderson, 1998). Anderson (1998) argues that Nietzsche’s perspectivism is placed between strong realism and extensive relativism. Meaning that it rejects the existence of an independent truth but perceives truth can be approached through overlapping perspectives. Meaning, one can only approach objectivity or truth by comparing one’s own perspective with the perspective of someone else (Hales & Welshon, 2000). This implies that every person’s thinking is biased. Everyone has to see the knowledge he or she produces within the light of his or her own perspective. At the same time, Nietzsche states as well that as a biased individual, interaction with others can help us to pursue knowledge (ibid.). Yet, taking his theory as an argument for epistemic democracy is not right, as Nietzsche was a strong opponent of democratic decision-making because he perceived humans as irrational beings (ibid.). Nietzsche’s approach to knowledge is something that epistemic democrats appeal to since it means that more people infer a closer approach of the asymptote that is objectivity. Every individual has a piece of the truth and together people are able to approach the truth from the aggregation of perspectives. Epistemic democracy is a means to aggregate these different perspectives on a larger scale and thus approach a collective decision.
Landemore has also received numerous critiques. First of all there is the question about what number is considered to be the crucial threshold that turns epistemic superiority into uncoordinated chaos? And although this is a valid question Landemore (2013) perceives this as unimportant because it is dependent on case specific factors. However, within every case there is a limit to the number. This number could potentially vary when different forms of centred and decentred deliberation are used (Landemore, 2014). The critique that strikes me most about Landemore’s theory of epistemic democracy is the seemingly effortless integration of the wide variety of predictive models within her simple deliberation. A group of
ideally most distinct individuals gather and naturally are able to optimally understand and consider each other’s perspective in order to make a decision. She is not solely talking about homogeneous societies, and formulates her theory for an ideal type complex modern society (Landemore, 2014). The more complex a society becomes the more people you would need to mirror the composition of the population (Landemore, 2013). A more complex society however, possibly complicates respectful deliberation and enforces polarisation within the deliberation process (Landemore, 2014). Although Landemore points out that this is not always the case and the literature is not definitive on this aspect, this leaves a theoretical gap in the deliberative phase (Ibid.). The gap is an opportunity for improvement and further research (ibid.).
Likewise, epistemic democrats generally assume a certain goodness and justness within a state. Condorcet mathematically argues that usually humans only have a higher than fifty per cent chance of getting it right, which is the foundation of his jury theorem. He likewise explains that if something happens and this value (the chance of getting it right) lowers below fifty per cent more people will only decrease the chances of getting it right. Landemore has a complementary remark of her theory. She notes that once a state suffers from a society wide bias such as racism, her theory reverses and the many are dumber than the individual (Landemore, 2012a). She likewise argues that the reason why there is little empirical evidence for epistemic democracy is because it is difficult to meet the requirements for ideal political deliberation (Landemore, 2013). These requirements are for example full-informed participants, equality amongst them and a certain resistance towards social pressure. The underlying tendency seems to be that the state needs some sort of stability in order for this mechanism to work (Landemore, 2014). Landemore, however, does not provide a mechanism to uncover a state’s instabilities and systemic biases. The question could be asked whether there is any state that does not suffer from a nation wide bias. It is therefore not feasible to legitimise a decision made within her proposed epistemic governing body.
Epistemic democracy, especially with the addition of a selection of a governing body by lot, has earned a lot of academic attention, but has likewise received scepticism because of the lack of empiric examples and a continuous debate about forms of
deliberation (Landemore, 2012a). Collective decision-making as described by Landemore is rare. This paper therefore suggests to investigates beyond the world of human societies into the realm of other living species, such as insects. Social biomimicry is an emerging chapter within the interdisciplinary field of biomimicry (Fewell, 2015; Grüner et al., 2015). A behavioural biologist might be able to enrich the field of epistemic democracy and provide insights in collective decision-making processes of non-human organisms. Therefore next to considering Landemore’s conception of epistemic democracy I want to provide a dominant understanding of honeybee collective decision-making (Seeley, 2010). Amongst animal societies there are only a few insect colonies that match our modern urban society, in resource and energy distribution as well as scale (Fewell, 2016). The most prominent is the honeybee swarm (Grüner et al., 2015; Fewell, 2016; Seeley, 2012).
Honeybee collective decision-making
Thomas D. Seeley is professor neurobiology and behaviour (Cornell University, n.d.). He has devoted his professional career to researching the behaviour of honeybees and specifically what he calls swarm intelligence (SI) of honeybees (Seeley, 2010). It is of utmost importance to understand the different lenses used for the researches used within this paper. Seeley is a biologist and a behavioural scientist and not a democratic theorist. Biomimicry is the application of biologist insights into a different field of research and in this case democratic theory. A potential pitfall within multidisciplinary research is confusion over the definitions. Especially since Seeley (2010) has applied his insights from his research to the political world though the lens of a biologist within his book where he summarizes his findings, Honeybee Democracy.
The title Honeybee Democracy connects animal behaviour to a human made concept, which literally translated means the rule of the people. The term honeybee democracy therefore could be seen as a form of anthropomorphism. And although opposed to the dominant conception of the ruling queen bee, the workers cooperatively shape the swarm. This being the case, we cannot immediate draw the conclusion that this would infer the existence of a democracy. There is difference within homogeneity of the society, cognitive capacities and the importance of the selves as opposed to the group’s interests (Fewell, 2015; Seeley, 2010). Therefore the
term democracy or democratic decision-making as used by Seeley will not be used to define the collective decision-making by honeybees. Instead within this paper a more informative term will be used, suitable for both humans and animals: collective decision-making. Each individual honeybee gathers incomplete information with the aim to aggregate this information and make a decision. To call it a democracy would only confuse and complicate the discussion.
Common interest is likewise terminology used by both Seeley and Landemore. There is however a difference between the common interest of a group of humans, and that of house hunting honeybees. Seeley argues how bees are able to make good collective decisions in the common interest. The common interest of the honeybee society is defined as the decision leading to the survival of the swarm on the short and long term. Within the nest hunting quest that was mentioned in the introduction of this paper, this infers choosing the nesting site that is most ideal for the hive: A large space with a small place to enter on a latitude that prevents most other animals from attacking their potential home (Seeley, 2010). The common interest is strongly correlated with the individual interest as the swarms’ survival is strongly correlated with the survival of the individual honeybee. Within epistemic democracy as described by Landemore the common interest, which leads to the common good, is not clearly defined. As I interpret the literature, Landemore wants to refrain from defining the common good. She does this because it is irrelevant for achieving an epistemic benefit as long as there is a mutual understanding of the common good amongst the decision-makers (Landemore, 2012b). Besides, she does conclude that a sovereign can chose to either act in line with the common good or solely out of self-interest (Althaus et al., 2014; Landemore, 2012b). This implies a duality between common interests and self-interest as concepts that cannot be aligned. The common good Seeley describes for the bee swarm is more consequence driven. The human concept of common good is less based upon consequentialist principles and more value driven. As both concepts of common interest are linked to the survival or succession of the individual and the group, I will continue to use common good for both humans and honeybees.
The inner workings of the honeybee hive
Honeybees are eusocial insects; insects that have several generations of reproductive and non-reproductive castes cooperating in nurturing the young (Nowak et al., 2010). The queen bee is the mother of all the individuals but has no other executive power over the other bees. Information is dispersed throughout the hive using their antennas, their body movement and piping signals (Frisch & Lindauer, 1956). The most prominent example of collective decision-making happens when a daughter swarm splits off from the mother swarm and has to choose a new nesting site. The daughter swarm ascends from their original home, flies for a few meters before the swarm rest and the decision process starts (Seeley, 2010).
The process begins by a selection of nest scouts from the thousand honeybees within the swarm. The scouts often consist of the older worker bees (the females), which usually work as foragers for the swarm (Seeley, 2010). This is not surprising, the nest scouts and the foragers perform very similar tasks with long travel distances and specific measuring methods. The scouting for alternative nest locations is done over many hours and sometimes even many days by more than hundreds scouts. Every scout flies out on its own in order to look for an option that fulfils the list of requirements it has learned to look for. After they have found a possible option they report back to the swarm by communicating the location, smell and quality assessment of the nest site with a dance on the swarm. Note that the information of any single honeybee is incomplete as it has only searched in one direction and is not aware of the full range of possible nesting sites (Grüner et al., 2015; Seeley, 2010). The neighbouring bees follow this so-called waggle dance in order to receive the information needed to make an independent assessment of the advertised nesting site (Frisch & Lindauer, 1956). If a neighbouring scout is convinced she will ascend and travel to the advocated location and make her own independent assessment. When convincing enough the second scout will also advertise their findings back with a dance on the swarm. This means that every possible site brought into the mix is considered and checked by multiple scouts. This lowers the possibility of assessment errors made by an individual scout (Grüner et al., 2015). This gradually discloses the phase of consensus building amongst the scouts, consensus; meaning an overall accepted opinion or decision within the group. Over time fewer new options are proposed and bees are primarily assessing the nesting sites found by other scouts.
The dancing honeybees aim to persuade the neutral honeybees to examine their option of choice. The process of building consensus can be broadly separated into two parts; the first part consists of the accumulation of a wide variety of possible nesting sites. In the second part only rarely new options are advertised and the main task of the bees is to compare and contrast the findings collectively.
Surprisingly the success of a nesting site is not determined by persuading every bee within the group of scouts to dance for the same location simultaneously; the success is rather determined by the amount of time any bee is dancing for a nesting site. Seeley (2010) found that a single scout bee dances for her discovered nesting site only for a few hours, regardless of the quality of the site. If she succeeds to convince neighbouring scouts her discovery will be revisited, if not the site is abandoned. This way the decision process is carried on from first generation scouts to the second-generation scouts and onward until all the active scouts are dancing for the same option. This works as a negative feedback loop as unconvincing options are automatically filtered out. Besides, it leaves room to process good quality nesting sites that reveal themselves later in the debate because of for example a longer travel distance (Grüner et al., 2015).
Finally the decision is made once no other nesting site(s) but one is advertised. Yet, scouts do not have a synoptic overview over the discussion, they solely know what their direct adjoining bees are advertising. The bees use a quorum in order to overcome their limitation of not having the cognitive ability to keep count and poll the dances on the swarm. This infers that about 20 to 30 bees, depending on the size of the swarm, have to be present at the nesting site for the decision to be made (Franks et al., 2002). This amount represents a quorum and is therefore a threshold after which the bees sense that a decision has been made. Once this happens the scouts return to the swarm and send out a signal. This piping signal will alert the other bees that the decision is made which leads them to ascend and travel to the chosen destination (Franks et al., 2002; Seeley, 2010).
The realm of eusocial insects has received increasing interest in the field of social biomimicry due to their capability to make group decisions with more than hundreds of individuals. Honeybee swarms, alike ant colonies are often determined as
superorganisms due to the insect’s exquisite capability to work together as if they are one (Grüner et al., 2015). Theorists have been exploring the possibility of applying behavioural patterns of insects to the social sciences. Seeley has attempted to distract insights from his lifelong work in the field of honeybee behaviour (Seeley, 2010). He has however only marginally linked his knowledge in this field to behavioural and political sciences. Grüner, Fietz and Jantsch (2015) have placed the honeybee swarm on the spectrum of the rational human to the bounded rational and the emotional human and place the swarm close to the rational human. They argue that humans are modelled as rational individuals with a rational decision-making process (ibid.). He additionally introduces the concept of the bounded rational human, which states that humans make non-optimal choices that are satisfactory as opposed to rational (Simon, 1990). For a rational decision the decision-maker is in need of full information, static interests and the capability to overlook all the consequences of a decision. None of these requirements are feasible according to Simon (1990), which results in the bounded rational human. From here Grüner et al. (2015) makes the assumption that we can learn from the bees in order to approach the rational human, economist are modelling and make better decisions. This could infer that when we apply the model of the bee swarm we might make better decisions. Fewell (2015) has investigated the application of Seeley’s (2010) theory and determined the differences and similarities between insect and human behavioural theories. Regardless of the simplicity of the bees’ cognitive activity and the obvious presence of a common good within their colony, their way of decision-making possibly works as a model. Bees have a homogenous society largely focussed on group success. Human societies, however, consist of a spectrum of individuals pursuing their own combination of both individual and group success (Fewell, 2015; Grüner et al., 2015). Still alike honeybees the human’s individual prosperity is heavily intertwined with the success of its environment (Fewell, 2015). Although it is acknowledged that the relationship between individual and group prosperity might be more complex due to the difference in homogeneity, group cohesion and individualisation, the resemblance remains (Fewell, 2015; Grüner et al., 2015).
Although direct comparison cannot be achieved due to substantial differences between human and insect societies, there are some similarities to be touched upon (Fewell, 2015). I will do this by dividing the decision-making process into three
phases earlier determined by Landemore, the first being the aggregation of option in the aggregation phase. The second, the deliberative phase, containing the distribution of options amongst de decision makers and, at last, the actual decision-making procedure in the decision phase.
1. Aggregation phase
Aggregation of options within Landemore’s epistemic democracy infers constituting a decision-making group with much cognitive diversity. This cognitive diverse group is constituted through a selection by lot, which is both efficient and economically favourable. This form of representation however implies as well that the decision body can consist of inexperienced people. Interestingly honeybees alike what Landemore argues have a selection of decision makers that have no decision-making experience. As any bee lives less than half a year and a swarm normally choses a new home site only once a year. Landemore argues that experience is not essential, as ability is placed third below (1) diversity and (2) numbers. Instead, diversity leads to a wide range of predictive models that can be applied to the problem (2012). Landemore’s argument is that applying, as much predictive models to a situation, will provide the group or governing body with a variety of scenarios, unimaginable for a single individual. The different perspectives that Nietzsche formulates are, for humans, represented within the diversity of the group. Amongst the bee scouts the perspectives are embodied within the search-direction of the individual location scouts. Every bee has limited information yet as time passes the nest scouts are capable of collectively disclosing the full range of nesting options in the area. Time and the capabilities of a single bee to discover all the possible nest sites in the surrounding region are limited. The time pressure is high because the swarming phase is dangerous for a hive. Every scout flies out in a different geographical direction and discovers one or several possible nesting spaces. It is thus of importance that the bees work together in order to aggregate the nest locations faster.
2. Deliberation phase
For humans, the most compelling form of distribution of options or predictive models would be through speech. Landemore proposes to use deliberation in a manner that lets every individual consider the other predictive models. This is a way
to test both problem definition and solution proposals to different predicative models and provide robustness to the decision. Amongst the scout bees the robustness of the nest location is provided by a mutual inquiry of each nest site. Every scout checks the allegations of the other scout before she accepts the claim. Landemore pays attention in her description of deliberation but limits her theory to simple face-to-face deliberation (2014). The deliberation is centralized in a way that everyone shares their ideas within a group as opposed to having multiple one to one conversations. She admits that this infers small groups but sees research perspective for new forms of epistemically optimal deliberation (Landemore, 2014). For honeybees the distribution of ideas is through a decentralized network (Grüner et al., 2015). When a honeybee comes back to the swarm hanging from a branch it excitingly promotes its found potential nesting site yet only the surrounding bees are able to receive the information. If one of the surrounding honeybees is convinced it will check on the nesting site and come back to inform the next surrounding group about the quality of the nesting site. What is interesting is that honeybees do not reach consensus on the swarm but reach a quorum at one of the possible options of choice. There is no leader aggregating the opinions or accounting the dances on the swarm. Humans on the other hand usually are able to interpret the general opinion of the group within the debate and the direction it is heading. Being able to account the general opinion is convenient but it can also discourage honest contribution, aimed at manipulating the deliberation.
3. Decision phase
The decision on one of the options for political deliberation is often a majority vote although smaller groups often seek consent (absence of objection) or consensus (mutual agreement). Honeybees work with a quorum, which implies there is a threshold on the choice for a nesting site, and when the threshold is reached (meaning that sufficient honeybees are at the nesting site) the decision is final. Although Seeley’s conceives majority vote a sign of division, Landemore argues that the majority vote will cancel out individual faults and biases. Landemore sees strength in the differences between people because it is able to better internalize the complexity of politics. Seeley’s sequential analysis of the honeybee movement on the swarm can here provide insights. Favourable nest locations are passed though the location scout group, meaning that the scout that discovers the site is not present at
the location when the quorum is sensed. This differs remarkably from humans’ central decision procedure. Central deliberation and decision-making could emphasize the perspective of the more outspoken individuals and encourage a tactical vote.
Throughout the three steps the significant difference between the two ways of making collective decisions is the framework from which the individuals communicate and aggregate information. Honeybees communicate through a decentralized network where one honeybee transfers its information to the next. The bees act upon a change in this information stream and accordingly know how to proceed. Communication within an epistemic democracy is centralized (although Landemore identifies room for different forms of deliberation). The diverse group of individuals selected by lot gather and consider the perspectives on the issue at hand and subsequently make a decision. The interesting element of honeybee societies is that they appear to work together in a self-regulating decentralized system; a balanced system that manages to execute complex tasks with cognitively limited individual bees (Grüner, 2015). Each honeybee operates according to a very limited and simple set of rules. When a threshold is reached the scout starts producing signals and when the signals resonate the scout ascends (Franks et al., 2002). Or when the neighbour bee dances vigorously, another scout reads the information from the dance and when convinced investigate the site. However all these separate actions interact and almost always guide the swarm into the most fit nesting site possible. This is at least the system that biologists present to the world from their perspective. Grüner argues for the investigation of the application of such a decentralized system into human collective decision-making (2015).
Interestingly, both Seeley and Landemore argue for a form of intelligence within the group. Seeley is talking about swarm intelligence when he explains how a group of individual honeybees is able to overcome its individual cognitive deficiency. The abilities of bees to aggregate information help them to reduce the trade off between speed and accuracy, which is essential within decision-making. Landemore likewise argues that intelligence is overcoming one’s own shortcomings. Landemore even states that collective intelligence is more than the sum of the intelligence of the individuals within it. This is the potential what she is striving towards and also the
purpose of the framework she theorized. The framework is her argument for why democracies have an epistemic edge over other forms of government. Both theorists define intelligence as the capability of a group to become more than the sum of its parts. Seeley has witnessed this empirically within the swarm. Landemore potentially aims to theorize a similar framework but designed for the cognitively diverse human kind. A systematic analysis might improve most important deliberative phase, as it is a method to overcome the individual constraints. Thereby it can improve the group’s intelligence (Meadows, 2008).
System Thinking
Seeing the world as a system is something that originates from multidisciplinary research but is commonly used in environmental and behavioural sciences (Haraldsson, 2004). System thinking works from the assumption that all of the scientific disciplines are non-linear. This means that any linear process discovered or researched happens within a context within a field of other interacting processes (Meadows, 2008). The interaction between the elements of a system can produce feedback loops, a loop of elements that once a change is made in one element all the other elements start enforcing or stabilizing the change made earlier. The loop that enforces a change in information is called a reinforcing feedback loop. Contrary, the loop that stabilizes the change made earlier is called a balancing feedback loop. These feedback loops present themselves in many variations; they can be either dominant or inferior throughout working of the system (Meadows, 2008). These interactions are complicated to comprehend for a human mind because we are used to think in linear relations (Meadows, 2008). As humans, we can only keep an eye on several of the numerous variables within a system. We either have biased information, which leads to false conclusions, or we have relatively accurate information but draw the wrong conclusion (Meadows, 2008). This phenomenon of making acceptable decisions with little information is theorized as bounded rationality (Simon, 1990). Systematic analysis is a means to overcome our own cognitive limitations alike the wild bee (Grüner et al., 2015). This cognitive limitation is our bounded rationality and Grüner et al. (2015) argues that humans can come closer to rational, good decisions when we look at the honeybee swarm. The honeybee swarm is a simplified version of individuals within a complex and ever changing world. Thinking in
systems is a way to visualize, organize and integrate information in order to understand patterns and relations within complex problems (Haraldsson, 2004). Landemore has likewise proposed a systemic approach to deliberation as a possible enrichment of epistemic democracy (2014). These deliberative democrats propose a decentred system of smaller deliberative groups in order to divide the requirements for optimal deliberation over several institutions (Chambers, 2017; Mansbridge et al., 2012). System theory within this thesis is, however, used for its epistemic benefit. It is used as a means to gather and organize diverse interpretive models as opposed to deliberation groups. In the following paragraphs, system theory will provide an epistemic answer to the information gap within the deliberation phase of Landemore’s theory.
Meadows determines three things necessary for a system: elements, interconnections and a purpose or function (2008). Within the swarm looking for a nesting site, the individual bees are the elements. The interconnections are the rules according to which they act. The purpose of the swarm can largely be seen as survival, corresponding to the common interest that was described earlier. But when you reduce the system solely to the swarming period, the purpose of the system is to find the best home in their reachable surrounding. Within a beehive the honeybees that advertise for favourable nesting sites produce reinforcing feedback loops as more and more bees start dancing and advertising the site which leads to exponential growth of bees dancing for that site. Complementary, bees dancing for a less favourable site will dance less convincing which produces a stabilizing or balancing feedback loop as perhaps only one bee is convinced to check the medium quality nesting site but feels not passionate enough to advertise it which cancels out the less optimal decision option. Honeybees live in societies with more than thousands of individuals who independently need to assess their course of action based upon what is needed within the hive. These basic rules internalized by the bees help them to deal with this complexity and perhaps can enrich humans to deals with our complex societies (Fewell, 2015).
Of course humans have societies that are much more dynamic and every individual does not act according to a standard set of rules. We are much more likely to change
our behaviour in favour of our social relationships (Fewell, 2015). However ideally governments are able to construct a balanced society where resources are naturally distributed and every individual is encouraged to develop and grow. We can see this for example within many progressive tax systems. When a household has a consistently increasing income, they will have to pay relatively more taxes since they likely profited from the current governmental and societal framework. This income tax consequently can be utilized to improve the societal framework in order to raise the lowest societal class (Ministry of Finance, 2020). A government is created to interfere with society only when it appears to fall out of balance, not when it is completely balanced. Liberals and socialists define out of balance in a different way, but the general principle is widely accepted among theorists (Schwartz, 2010). Seeing a state as a deliberative system has similarities to Mansbridge et al. (2012), but without the focus on the epistemic benefit. There is complexity within human societies that is much harder to grasp and leaves much room for chance compared to honeybee societies. The message here is that complexity can seem overwhelming but by visualising the system it can become easier to understand and comprehend complexity (Meadows, 2008).
The approach of thinking in systems can fulfil an epistemic addition to deliberation as described within Landemore’s epistemic democracy. The aggregation and integration of predictive models as described by Landemore can be more effective when theorized within the framework of system thinking. The framework can assist the group to grasp the non-linearity and the interconnection, difficult to disclose in either written or spoken language. Language is constructed within a linear essence, building on logic and consequential relationships (Meadows, 2008). Nietzsche states that knowledge is constructed from dialogue; system analysis can be seen as an approach to deepen dialogue. Visualising language by the means of a system helps to overcome this shortcoming of spoken and written language. It is difficult to filter out the differences in deliberative talent amongst the individuals in an open deliberation. There is pluralism in personalities, which can lead to an unequal consideration of the insights brought to the table by the individuals. By visualising the system the chances are higher that every variable is actively considered within everyone’s final individual vote. This addresses the question many theorists have appointed (Landemore, 2014)
To summarise, using a system to integrate the predictive models of the members present in a decision group does not reject the theory of Landemore. It rather builds upon her theory and perceives thinking in systems as a way to epistemically enrich the deliberation phase of Landemore’s theory. System theory can face the information gap in the deliberative phase as it addresses both pluralism and the human cognitive limitation. Alike the honeybee, system theory helps to overcome individual cognitive limitations and approach the epistemic rational choice. Thinking in systems has the potential to uncover hidden system traps and system malfunctions. System mechanisms that inevitably lead to problematic escalating behaviour whether it is a race to the bottom or destructive growth can be uncovered (Meadows, 2008). Besides it trespasses the issue of blame and aims at understanding issues such as institutional racism, climate policy and poverty (idem.).
Let me explain what the application of system thinking might look like. The decision group has to understand the intention and structure of the decision making process, however this is proportional to explaining what respectful deliberation entails. The group is informed about the way a system can be constructed, the purpose, elements and the interconnections. The purpose of epistemic democracy is to make a decision that benefits the common good. Further the issue at hand is determined, from there on elements are (visually) added to the system in order to deepen the understanding of the systematic problem behind the visual symptoms.
There are many formats to portray a system, ranging from theoretical to mathematical (Meadows, 2008). Further research might dive deeper into the pragmatics of system thinking. For now the causal loop diagram functions as an example, as it is the simplest version and most suitable for qualitative data instead of quantitative research. A causal loop diagram maps the elements and determines the direction and character of the relations between them. It can be used to analyse the workings of policies or separate decisions made. Usually arrows are marked with a plus or a minus, which indicates either a positive or a negative correlation. When variables are connected in a circle with arrows pointing the same direction we are speaking about a feedback loop. The plusses and minuses indicate the nature of the feedback loop. Zero or an even amount of minuses indicates a reinforcing feedback loop. An uneven amount of minuses indicate a balancing feedback loop. The causal
loop diagram is an accessible approach to system theory, easily explained to the decision-making group.
Another important factor within system analysis is the integration of delays; especially economists tend to neglect the effect of a delay within a system. Everything takes time, it takes time for information to be transferred, and it also takes time for information to be processed. It takes time for money to be transferred and for norms to be changed. When a delay occurs within a system the output of the system starts to oscillate. Oscillation is not necessarily a negative effect but does influence the behaviour of the system.
A great difficulty with this method is setting the limit of the system. Systems for political use are supposed to be an abstract reflection of the relevant part of the state. This requires systems to find the balance between too abstract and too detailed. A system too abstract loses it relevance to real life decisions. A system too detailed loses its relevance because the chaos makes it impossible to draw any conclusion for real life. Setting the limit of a system is therefore of great importance for the value of a system. It is important to understand however that boundaries within a system are man made for functional purposes. When we look at reality there are not much actual boundaries and systems are often interconnected with other systems (Meadows, 2008). One might argue that this difficulty, essential to system thinking, is a challenge for using systems all together. Which is a valid point to discuss since this proposal is meant to improve deliberation, not complicate it. The representatives are however selected by lot and have not been operating within the political sphere their entire life. This aspect gives them an advantage and a disadvantage of having no prior experience. A disadvantage manifests because of the lack of a learning curve, the participants are not able to learn from past decisions made with the identical group. Landemore (2012b; 2013; 2014) addresses this objection to her theory in several papers. The advantage of having no prior experience is a decrease of prejudices towards groups, institutions and mechanisms. In the case of drawing boundaries it is rather useful. People mentally accustom to boundaries they are interacting with however for different types of issues a different boundary setting is required. This gives the inexperienced an advantage over an experienced group of representatives.
Conclusion
Humans are learning much from the natural world as they have been, for a long time. Honeybees present us with a form of collective decision making that surprises and astonishes. Bees are able to aggregate information collected by more than hundreds of bees and find the option best fit for their swarm within days. Hereby they are capable of overstepping their individual cognitive shortcomings and work together in a coordinated, efficient and effective system. Humans have likewise known the fundaments of epistemic theory since the emergence of democratic theory in historic Athens. Aristotle first argued for the intelligence of the group over the knowledge of the individual. Ever since epistemic theory has developed, Landemore has provided the world with an appealing proposal to fully exploit the wisdom of the crowd. She argues that for the epistemic benefit to exhibit itself, cognitive diversity is critical. She bases this on the Diversity Trumps Ability Theorem. According to Landemore, cognitive diversity will provide the decision-making process with widely varying predictive models, internalized in the diverse individuals. In order to easily establish a high level cognitive diversity she proposes to use political representatives, selected by lot as opposed to a regular election used in the current system. Following she describes how inclusive and respectful deliberation results in a collectively wise choice, leading to the common good. The deliberative phase is where honeybee wisdom can enrich the field of epistemic democracy. It is also the field that Landemore proposes as an area of further research due to issues of pluralism and the limitations of the human cognition. The system that bees appear to interact with can help humans to map and relate the intelligence established from the predictive models present within the group. Likewise Fewell and Grüner et al. propose the opportunities for honeybee intelligence to refine social sciences. System thinking is an approach to visualize the interconnected variables that shape a problem. It can be a means to collectively acquire and shape knowledge in the tradition of Nietzsche’s perspectivism. This is necessary in order for a group to overcome their own cognitive limitations alike the wild honeybees. Cognitive limitations such as our human bounded rationality, the human difficulty to keep track of multiple variables simultaneously and the difficulty we face to to comprehend non-linear relationships. Using system theory as a foundation for democratic deliberation might enhance the epistemic edge democracy has over other forms of government. Within these systems, reinforcing and balancing feedback loops can be defined which can help to
analyse the structural problems within a society. It provides a smarter decision framework for deliberation within Landemore’s theory of epistemic democracy.
Bibliography
Althaus, S., Bevir, M., Friedman, J., Landemore, H., Smith, R., & Stokes, S. (2014). Roundtable on Political Epistemology. Critical Review, 26(1-2), 1-32.
Anderson, R. L. (1998). Truth and objectivity in perspectivism. Synthese, 115(1), 1-32. Aristotle. 1998.Politics.Translated by C.D. Reeve. Indianapolis: Hackett Publishing (IV 14).
Biomimicry Institute. (2020). The Biomimicry Institute - Examples of nature-inspired sustainable design. Consulted on op 14 January 2021, from
https://biomimicry.org/biomimicry-examples/
Bohman, J. (2006). Deliberative democracy and the epistemic benefits of diversity. Episteme: A Journal of Social Epistemology, 3(2), 175-191.
Chambers, S. (2017). Balancing epistemic quality and equal participation in a system approach to deliberative democracy. Social Epistemology, 31(3), 266-276.
Cohen, J. (1986). An epistemic conception of democracy. Ethics, 97(1), 26-38. Cornell University. (n.d.). Thomas Seeley | Department of Neurobiology and Behavior Cornell Arts & Sciences. Consulted on 7 January 2021, from
https://nbb.cornell.edu/thomas-seeley
Dicks, H. (2016). The philosophy of biomimicry. Philosophy & Technology, 29(3), 223-243.
Elster, J. (1997). The market and the forum: three varieties of political theory. Deliberative democracy: Essays on reason and politics, 3, 18.
Estlund, D. M., Waldron, J., Grofman, B., & Feld, S. L. (1989). Democratic theory and the public interest: Condorcet and Rousseau revisited. The American Political Science Review, 1317-1340.
Estlund, D. (2008). Introduction: Epistemic approaches to democracy. Episteme, 5(1), 1-4.
Fewell, J. H. (2015). Social Biomimicry: what do ants and bees tell us about organization in the natural world?. Journal of Bioeconomics, 17(3), 207-216.
Franks, N. R., Pratt, S. C., Mallon, E. B., Britton, N. F., & Sumpter, D. J. (2002). Information flow, opinion polling and collective intelligence in house–hunting social insects. Philosophical Transactions of the Royal Society of London. Series B:
von Frisch, K., & Lindauer, M. (1956). The" language" and orientation of the honey bee. Annual review of entomology, 1(1), 45-58.
Gould, C. C. (1990). Rethinking democracy: Freedom and social co-operation in
politics, economy, and society. Cambridge University Press.
Grüner, S., Fietz, A., & Jantsch, A. (2015). Float like a butterfly, decide like a bee. Journal of Bioeconomics, 17(3), 243-254.
Hales, S. D., & Welshon, R. (2000). Nietzsche's perspectivism (Vol. 22). University of Illinois Press.
Haraldsson, H. V. (2004). Introduction to system thinking and causal loop diagrams (pp. 3-4). Department of Chemical Engineering, Lund University. Hong, L., & Page, S. E. (2004). Groups of diverse problem solvers can outperform groups of high-ability problem solvers. Proceedings of the National Academy of Sciences, 101(46), 16385-16389.
Landemore, H.E. (2012a) Why many are smarter than the few and why it matters, Journal of Public Deliberation, 8(1).
Landemore, H. (2012b). The Mechanisms of Collective Intelligence in Politics. Collective wisdom: Principles and mechanisms, 251-89.
Landemore, H. (2013). Deliberation, cognitive diversity, and democratic inclusiveness: an epistemic argument for the random selection of
representatives. Synthese, 190(7), 1209-1231.
Landemore, H. (2014). Yes, we can (make it up on volume): Answers to critics. Critical Review, 26(1-2), 184-237.
Mansbridge, J., Bohman, J., Chambers, S., Christiano, T., Fung, A., Parkinson, J., ... & Warren, M. E. (2012). A systemic approach to deliberative democracy. Deliberative
systems: Deliberative democracy at the large scale, 1-26.
Meadows, D. H. (2008). Thinking in systems: A primer. chelsea green publishing.
Mill, J. S., (1861), Considerations on Representative Government, Buffalo, NY: Prometheus Books, 1991.
Ministry of Finance. (2020). Soorten inkomstenbelasting. Consulted on 17 January 2021, from
https://www.rijksoverheid.nl/onderwerpen/inkomstenbelasting/soorten-inkomstenbelasting#:%7E:text=Het%20tarief%20voor%20box%201,(was%2040%2 C85%2).
Nowak, M. A., Tarnita, C. E., & Wilson, E. O. (2010). The evolution of eusociality. Nature, 466(7310), 1057-1062.
Schwartz, H. M. (2010). States versus markets. Palgrave Macmillan.
Seeley, T. D., & Visscher, P. K. (2004). Quorum sensing during nest-site selection by honeybee swarms. Behavioral Ecology and Sociobiology, 56(6), 594-601.
Seeley, T. D. (2010). Honeybee democracy. Princeton University Press.
Simon, H. A. (1990). Bounded rationality. In Utility and probability (pp. 15-18). Palgrave Macmillan, London.
