Disinformation Society , Communication and Cosmopolitan Democracy

This paper argues that ‘fake news’ is endemic to ‘information society’ as a whole, not just the internet or news media. It is part of daily experience, generated by established patterns of communication, social group categorisation, framing, and patterns of power. These disruptions are intensified through interacting with the dynamics of information capitalism, which values strategic effectiveness more than accuracy. Assuming democratic cosmopolitan society must have good communication, this paper explores the factors which produce obstacles to such communicative processes, as the patterns which support bad communication and disinformation must be understood before they can be dealt with.


Introduction
Assuming democratic cosmopolitan society must have good communication (Hosseini 2013;2015: 58-59), this paper explores the obstacles to such communicative processes.Processes of communication, information and reasoning are here seen as primarily concerned with maintaining group bonding, group category boundaries, status within groups, and selfidentity, rather than about conveying accurate information or finding rational solutions to problems (Mercier and Sperber 2017;Malinowski 1949, pp. 312ff.).Consequently, I argue 'fake news' is generally common, but endemic to 'information society' as a whole not just to news media.It is part of daily experience, generated by the patterns of communication and magnified by information capitalism.In making this argument, I do not distinguish between 'fake news', 'lies ', 'disinformation', 'misinformation' and 'misunderstanding', as differentiating between those terms requires understanding the intentions of others, which is fraught.Furthermore these categories overlap, and are affected by similar processes.
Firstly, I describe some patterns which produce normal miscommunication and information distortion, such as: complexity, categorisation, framing, power relations, standardised lying, information groups, hierarchy, and siloing.I then discuss how capitalist 'information society', intensifies these processes through data smog, orders of doubt, markets, precarious information work and distributed governance.I avoid mono-causality, agreeing with Luhmann (1989, pp. 8-9) that single causality primarily operates to allocate blame or innocence, although I part from Luhmann by suggesting that systems work differently depending on their 'components', and the 'components' work differently depending on the system.We cannot understand disinformation or 'fake news' without understanding how humans communicate, and we cannot understand how humans communicate without understanding the effects of social patterns.Finally I discuss the ways some of these problems might be diminished.
Fake news could easily be explained as propaganda.Nearly all media sources are owned and controlled by the corporate sector.It might seem reasonable to argue that the media supports or justifies that sector, its profitability, ideology and power and to point out that there is almost no mainstream 'left-wing' media independent of capitalism.Since the 1980s neoliberal, pro-corporate, policies have been implemented.These have been sold with undeliverable promises of liberty and prosperity, consequently neoliberal politicians and their corporately sponsored supporters distort reality and distract the populace with fake news.While there is no denying propaganda happens, it acts in a context which naturalises it and gives it force.This paper studies the context, not the propaganda or its makers.
The paper necessarily takes a broad span to make its case, it aims to gather some general principles, and make a general argument.As such, it summarises work in many fields, being exploratory and suggestive of further research, rather than definitive.

Reflexivity and social science
Human beings are social and self-reflexive creatures.Consequently, within limits, they change their behaviour in response to the behaviour of others, and in response to their theories and information about others (Bryant 2002).Self-reflexive methodology assumes that social scientists (at least) can change their behaviour and understanding with experience, while politics assumes that influencing information can influence behaviour.Theories direct attention to events and interpret what is happening, so that changing theories changes what people perceive, or interpret, of their world (Chalmers nd).Consequently, social theory can never be complete, because if some people accept a theory they may take advantage of it, and thus change the dynamics of society.For example, Marxist theory promoted revolution in some countries, and 'reform' in others, modifying the dynamics of capitalism.After the fall of Communism in Europe, neoliberal theories about capitalism's success led to revocation of this reform, producing a return to economic, social and ecologic crises.
Human reflexivity is one factor that makes society what is known as a 'complex system', which interacts with other complex systems.

Complexity
A 'complex system' is a collection of interacting nodes with multiple connections, which modify each other, or which self-modify in response to stimuli from the system (Marshall et al 2015, pp. 10-14;Incropera 2016;Holland 2014).Economies, information and communication systems, and ecologies are different complex systems which are not independent or easily separated (Luhmann 1989).Different systems can interact to balance or to disrupt each other, or both.The communication and information system may disrupt the ecological system or the economic system and vice-versa (Marshall et al 2015, pp 14-15).
Complex systems are usually in flux, even within relatively stable recurring states, but even when complex systems are relatively stable, they are not entirely predictable.Surprises happen all the time (Doak et al 2008).Gertjan Vlieghe, of the Monetary Policy Committee of the Bank of England, has argued (2017) that although completely accurate economic forecasting is impossible, it can be useful: trends may be predictable, if not in detail.We can, for example, predict that unless human systems change, the world will get warmer.We cannot predict what the weather will be like in ten years in Sydney.Similarly, we can describe those circumstances which lower the possibility of productive communication, but cannot predict what will be communicated.
Setting aside disputes about 'truth', human attempts to understand, model, or order, complex systems can never be complete or entirely accurate; they always have to acknowledge some degree of uncertainty and lack.Models of lesser complexity than the events being modelled are misrepresentations and partly fictions.In complex systems we can only discover the relative accuracy of our models, by observing what follows from our actions and paying attention to unexpected happenings.Without attending to how processes regularly become disordered, it is likely our actions will generate more unintended and deleterious effects.
Because of this reflexivity, complexity, uncertainty and unpredictability, nonmisleading social science propositions can rarely be of the form: 'behaviour, structure, or event A results in behaviour B'.They are more accurately of the form: 'A tends to produce B in this context'.The logic required of complex social science is fuzzy and non-deterministic.

2
Information Disorder in General

Communication
Communication and information are entangled with noise, misunderstanding, degrees of inaccuracy, and deception.Without the possibility of miscommunication, there is no possibility of communication.Even relatively non-interpretive signalling systems can involve miscommunication, as when creatures evolve to appear as something they are not, or predators evolve to mimic their prey's mating signals.Evolutionary mimicry and 'deception' even appears to operate at the molecular level (Barbe & Elder 2013).
In human systems, all information is interpreted or misinterpreted, and information is not conveyed from emitter to receiver without the possibility of change (Lakoff 1995).Accurate communication needs effort, and is easily disrupted, unintentionally or otherwise (Marshall 2007, pp. 13-30).These effects can be intensified by inter-relationships, as inaccuracy can be strategically advantageous, helping to produce both social harmony, and victory in conflict.

Categorisation and Social Categories
Linguistic categories are necessary to communication and thought; as Ellis suggests 'categorization not syntax is the most basic aspect of language ' (1993, p. 27).However, categories are disruptive and potentially distorting, as they parcel up the world in ways which may not be absolutely accurate, or completely shared (Marshall 2007, pp. 14-17).They group 'together things that are not the same in order that they will count as the same' (Ellis 1993, p. 25).Categories summarise assumed properties of the events, people and objects 'placed in them' and indicate socially and historically relevant similarities and differences.
Many categories are fuzzy.They often do not refer to things/events which are 'all the same in the same way' but, as Wittgenstein argued (1968, §66ff), they group 'things' or 'events' according to many different principles, as with 'family likeness'.This person has that person's nose, this person's chin looks that that person's, they walk like their uncle, some members have similar talents etc.This fuzziness, and use of varied principles, is especially true of social categories (such as ethnicity, status, occupation, gender, knowledgeability, religion and so on) where 'likeness' grows out of history, social purpose, relationship, usage or conflict.People might demand that as families, ethnicities, or other groupings, are linked by 'blood', 'descent' or 'culture', and serve social functions and connectedness, members should resemble each other and differ from other groups even if they don't (Lakoff 1987;Ellis 1993;Turner et al. 1988).However, such categories are fluid, under construction or subject to competition, as they imply degrees of privilege, status, safety and so on.Social categories are political in nature, as people in categories cooperate and conflict, or try to shift themselves or others from one category to another, or resist category imposition.Stereotypical categories act as heuristics for action in situations of imperfect information, but as heuristics they may not always be accurate (van den Berghe 1997).While stereotypes impose an identity, all social and self-identity and worth depends on social categories; as identity is about 'identification with', and 'identification against', interacting categories of people, despite different groups not always interpreting social categories identically (Turner et al. 1988).
While categories are necessary for information, communication and action, they also produce distortions, incoherencies and inaccuracies, and are subject to struggle and change.

Framing
Accurate interpretation of messages is difficult, because categories are fluid and fuzzy, sentences are not well-formed, and contexts give a range of possible meanings.However, completely accurate interpretation and completely shared meanings is not necessary for most human activities.It only has to be good enough for the purpose (Peckham 1979).Interpretation commonly involves 'framing', or bringing a personal and/or social context, to the message.This framing gives information about the message, its emitter or the situation of emission, allowing ambiguities to be reduced, although not necessarily accurately.Different people or groups may use different framings for the same messages, thus understanding those messages differently (Marshall 2007, pp. 18-23, 311;Lakoff 2014).This difference can drive, and reinforce, the social dynamics between the groups (Barth 1993: 176), as when an intended friendly overture is framed by the recipients as ignorant or hostile, accurately or not.
People frame communication and information by information they already have, so the value and meaning of information depends on other information and is not stable between groups.This information includes interpretations of the history, power relations and struggles between the social categories involved, as well as the properties of the social categories involved.Consequently, established political and social conflict or harmonies, provide important framings for communication.People often attempt to control the framings being used so as to control the argument (Lakoff 2014), and people often act as if they held framings like: 'This person and their information is right or left wing, therefore their information is worthless'.As such, framings can generate 'confirmation bias', or 'my-side bias', (Mercier & Sperber 2017, pp. 211-21) in which information is only accepted if it agrees with what the person/group already believes.
Whilst framing is an essential part of interpretation, it can increase both misinterpretation and rejection of accurate information.

Politics and Communication
Much human communication is inherently political, involving relations of power and using strategic or tactical rhetoric, aiming to persuade other people to act, obey, maintain the relationship, or think in a particular way, etc. (Peckham 1979).However, in complex systems, results of communication, or action, are uncertain.Even a dialogue forms an unpredictable, fluxing and complex system, as people adjust to each other, and attempt to persuade each other.The greater the number of people involved, the more unpredictable, and uncontrollable, the results can be.Force (varying from physical violence to name calling, shaming, and expulsion), can be used to try to reduce uncertainty and maintain clarity, by stopping questioning, removing objections, or punishing particular reactions.Similarly, influencing or disrupting other people's social theories, the social categories they deploy and their imaginings of the future, helps political persuasion and success.This means that lying and hiding information are normal parts of communication, and become more intensely deployed as direct force becomes difficult.
Hannah Arendt remarks that 'truthfulness has never been counted among the political virtues ' (1972, p. 4), and notes that in any political action directed to the future, we have to imagine that things might be different, so that 'deliberate denial of factual truth... and the capacity to change... are interconnected' (ibid., p. 5); 'the lie did not creep into politics by some accident of human sinfulness' (ibid.p. 6).If we add the predictability problems of complexity, then it seems clear that political visions of the future are useable fictions.
Being non-factual is inherent to political and social action, even though it can become disruptive of that action.

Lying as normal
Social psychological experiment shows that in U.S. culture, self-recognised lying is an 'everyday event' (De Paulo et al 1996, p. 984). De Paulo, et al (1996, pp. 984, 989) report that their subjects lied between one in three and one in five of their social interactions and appeared to see those lies as necessary, as they said they would repeat them if the situation could be relived.Taylor and Feldman (2004) observed 78% of their subjects report lying in a 10 minute conversation with a stranger.Rodriquez and Rayne report that despite differences in social class and local cultures amongst their subjects, lies were standardised (ibid, p. 199), largely arising from 'the preference for interactions that promote support, cohesion and solidarity; and the dispreference for actions that undermine such interests' (ibid, p. 208).Honesty was superseded by the imperative 'to maintain affiliation and the identities upon which that affiliation is dependent' (ibid., p. 209 emphasis added).People also lie to gain advantage and this includes putting themselves in the best light as well as manipulating others (DePaulo et al. 2004).'Lying is publically condemned... but practised by almost everybody… to accomplish the most basic social interaction goals' (Kashy & Depaulo 1996, p. 1037).
In her study of the Pentagon Papers from the Vietnam War, Arendt remarks that the 'policy of lying was hardly ever aimed at the enemy... but was destined chiefly... for domestic consumption ' (1972, p. 13), or for keeping unity amongst the policy makers themselves (ibid., p. 34).In a preliminary study of lying in international politics, political scientist John J. Mearsheimer (2011) reinforces this observation, by arguing that leaders are more likely to lie to their own people, than to the leaders of other states.He hypothesises this is because leaders know that other leaders will attempt to verify their statements (ibid., p. 28), due to mutual distrust, but this could also arise because leaders put themselves in similar categories, and their people lower in the hierarchy (see section II.6 below).
Lying, is central to the politics of maintaining relationships and persuasion, as well as to manipulating others: it is both functional and disruptive.

Information Groups
Social power is often based in maintaining, expanding or limiting group membership, shaping group framings or categorisation, and categorising opponents as belonging to despised groups (Marshall 2007, Lakoff 2014).We can label the groups that form around, or cultivate, particular kinds of knowledge or culture as 'information groups' (Marshall et al. 2015).These can be 'on-the-ground' groups, but in information society they are frequently disembedded 'imagined communities' with relatively shared values and categories.Indeed, Anderson's (1991) original formulation implied that the 'imagined communities' of nations are created through shared media, and the resulting creation of shared selfcategorisation/identities, information and values.
The basic dynamic of information groups expands the observation that the more people of similar opinions talk to each other the more similar their opinions become, and the more distant they become from what they interpret to be the opinions of those categorised as opposing them (Sunstein 2009).Repetition of information by others in the same, or valued, group(s) reinforces its truth.People may approve information to keep their position in the group, persuade others or smooth group relations.Ideally, the information group helps tell people who and what to accept and who to distrust through the experience and judgement of others; as, in daily life, people rely on the knowledge of other people especially in information society where we cannot know how to do everything, or know how everything works (Giddens 1991).We are used to being ignorant but capable of operating and we depend on the knowledge of other people in our 'community'.The information group reinforces what Rozenblit and Keil (2002) call the 'illusion of explanatory depth', in which people feel they understand complex phenomena with much greater precision, coherence, and thoroughness than they do, by helping to filter out counter-information, while providing assurance that some group members really do understand.However, no group members may understand properly, and the group will reinforce and pass on mistakes as well as accuracy (Sloman & Fernback 2017).
Members of information groups tend to disparage categorised outgroups (especially those perceived as using different information).This enforces boundaries, inhibits exploration of other sources, and foreshadows what might happen to people if they step out of their group.Perception of group membership, is also connected to power and status within the group, as the more message receivers frame and categorise the emitter as an exemplar of their particular group, the more persuasive the message becomes.The more receivers are persuaded that an emitter exemplifies an outgroup, the less persuasive they become and the more likely their information will be rejected (Haslam et al 1996;Hopkins & Reicher 1996).Similarly, people seem better at spotting errors in the arguments of others than they are in detecting their own or their group's errors (Mercier and Sperber 2017, pp. 213ff.).It can, therefore, be useful to some people, to attempt to reinforce those boundaries and close off cross-group conversation to maintain power.
Group positions can also become emphasised, when such positions distinguish high status ingroup members from outgroup members, and so shared and differentiating cultures can arise between information groups.For example, it appears Republicans and Democrats read different science books (Shi et al 2017), or share statistically different first names (Verdantlabs 2016), and hold radically different views on the competence of the President, and the veracity of news about him.Jones (2017) reports a Gallup poll stating that while 87% of Republicans approved of President Trump's performance, only 9% of Democrats did.Differences between groups can increase, as when during the period of climate change recognition, and neoliberal triumph, US 'conservatives' moved from having the strongest faith in science to the lowest, while Democrats remained stable (Gauchat 2012).
Information groups can be relatively open or closed in specific ways, depending on context.Thus, in the US, Republican information groups encompass libertarians, religious fundamentalists, conservatives and white supremacists, while attacking Republicans who engage with Democrats as 'Republicans in Name Only'.Some evidence suggests 'right wing' information groups cluster around fairly few sources and are more closed to outgroup information than 'left' groups.Pro-Trump Republicans particularly cluster around Brietbart, owned by Steve Bannon, part of Trump's administration (Benkler et al. 2017), which indicates the unity of identity and information arising from rejecting and distrusting outgroups.When boundaries are tight, members of one information group may be completely unaware of the existence of particular other information groups or their informational focus, having little information exchange with these groups, even when in the same organisation (Marshall et al 2015, pp. 82-3).
Information groups provide filters, development and background but they make information accuracy secondary to group and identity dynamics, and to relationships with other groups.

Hierarchy
Hierarchies can further restrict information flow as people can learn to expect punishment for giving unwanted but accurate information to those above them in the hierarchy, and rewarded for giving less accurate, but wanted, information.'Deception is a crucial element in avoiding and managing hierarchical impediments, from incompetent bosses to exemplary ones' (Shulman 2007, p.158).'Higher-ups' can be ignorant of basic facts about the organisation they supposedly control (Mezias & Starbuck 2003).The more punitive or violent the hierarchy, the more that accurate information flow from below upwards is disrupted, and the more deceit is likely to be common.A shallow hierarchy, with large differentials in power, can be as destructive to information flow as a deep hierarchy, and higher-ups can emphasise the status and category difference between them and underlings through furniture or the position of their offices (Detert & Burris 2016).Attempts by higher-ups to circumvent this problem can be self-defeating through emphasising the dangers of communicating upwards (Detert & Burris 2016).Similarly, those higher up may not share their information with those lower down, especially when position and status is justified by a reputation for knowledge or intelligence and could be threatened if information, especially of mistakes, was available to those below.Similarly, higher-ups may lie, or hide information, strategically, when their planned actions will affect people in their organisation deleteriously (Marshall 2012, pp 295-8).These processes generate the 'power/ignorance nexus' (Graeber 2004, pp 72-3) and the nexus works both ways.
Higher-up 'ignorance', means that alternate, informal, or unauthorised, networks can arise in a workspace to get work done, or to share information and learning.These networks may hold the unofficial and necessary knowledgebase of the organisation (Marshall 2012, pp. 293ff.;Shulman 2007, p. 84).Higher-ups, not part of these networks, can interpret them as distracting, a challenge to power, or as a waste of resources, and attempt to destroy them, throwing functional work into difficulty (Sachs 1995;Hall et al. 2009).The more authoritarian or punitive the hierarchy, the more likely that higher-ups may not like invisible or informal channels of communication to exist, as they fear being undermined.
Higher-ups may have the ability to survey what those beneath them do, as Foucault (1979, pp.170ff) suggests in his account of the fictional panopticon, or the shifting of large scale observation from military camp to other institutions.However, while 'superiors' may observe inferiors, they still have to frame and interpret that observation.Accurate interpretation requires prior knowledge, which may have no basis in the life they are observing, or is muddled by the conviction that they are superior and less prone to mistakes.Inferiors knowing they are observed, dissemble in turn.Surveillance does not necessarily disrupt the power/ignorance nexus.Surveillance may also render it harder for an organisation to adapt to local conditions, as flexible and useful responses are inhibited by commands from the centre (Marshall et al. 2105, p. 65).Individuals, at all levels in the organisation, deceive in order to gain more autonomy and functionality than their workplace conditions allow (Shulman 2007, p. 157).
Furthermore, many people in a hierarchy may have ascended to their level of incompetence (Peter & Hull 1969;Lazear 2003).If so, these higher-ups may not be capable of understanding, or implementing, either what is required of them or the tasks of their underlings; they may favour people who emit desired and reassuring information, reinforcing the power/ignorance nexus.To obtain adequate information, people must know when they have inadequate information.People who are incompetent or ill-informed in a domain, often lack the skills to evaluate competence in that domain, and tend to think they are competent, while being unable to recognise competence in others (Kruger and Dunning 1999;Erhlinger et al 2009).This ignorance of incompetence can be socially reinforced by the agreement of other people in their information group.Dominant people who are informationally challenged, or threatened by hostile information, may attempt to reinforce their information groups and weaken those groups interpreted as hostile through promoting their own members, defunding opponents, sacking opposition, appointing 'true believers' to run enquiries, making disruptive rules, restructuring and so on.This extension of group ideology or practice may be destructive, as with the extension of corporate style management and values into universities (Collini 2017) or government (Hood and Dixon 2015) with little evidence of savings or improved service for all the upheaval.The new US Presidency defends its knowledge/ignorance by deleting scientific data from the Web (Hermann 2017) while appointing people hostile to environmental protection to administer environmental protection.

Organisational Silos
Different parts of an organisation can form their own information groups, and develop their own cultures.Such groups are often called 'silos'.Organisations may encourage silos for efficiency as gathering specialists together seems to support their ability to process information quickly, and reinforces skill-based community.However, it also breaks communication and knowledge transfer within the organisation (Tett 2015).
Different silos can be in relationships of rivalry, or conflict because they value different functions, information and theories: the informational worlds of accounts may not resemble those of research.It cannot be assumed that any one part of the organisation understands the information, realities and problem fields of any other part of the organisation, and so 'structural secrecy' develops.This can lead to blockage of information, misdirection, misunderstanding and projection (Vaughan 1999).
Structural distinction and hierarchy may help organisational efficiency but they create patterns of ignorance which lower that efficiency.
While these patterns of disinformation can exist in most complex societies, patterns in information capitalism both extend and intensify these communicative problems, and these patterns can be relatively easily manipulated by those who find hostility, separation and miscommunication useful.

Data Smog
Ease of information production generates the problem of 'data smog' which is fundamental for information society (Shenk 1997).The easier it is to create and store information, disinformation, or noise; the more will be created.'Information' becomes so plentiful, it is difficult to find (or sort out), what is accurate or relevant.Filtering that information by what the searcher already knows, or social pressure, becomes even more likely (Marshall 2013b, p. 5-6).
As information production becomes a mark of competence, academics and scientists can publish before ready to maintain their threatened jobs.There is less time to carefully referee or replicate research (and replicating research may be seen as low status).'Predatory' or 'pseudo' journals flourish, entrapping inexperienced writers who need to meet publication scores (Marshall et al. 2015, pp. 207ff), and enter library catalogues in certain subjects (Nelson & Huffman 2016).People promoting dubious medicine, or anti-climate change arguments, can now find journals to give their disinformation apparent credibility.Media commentators, likewise produce articles to keep their names before the public to bolster their celebrity and earnings, lowering accuracy in favour of attention and consistency.
As data and arguments supporting almost any position can be found, self-confirmation is easy.Decisively refuted information stays available and can be spread again if it confirms a bias, as the refutation and counter-information may not be as easily visible or may not come from a group valued by its readers (Shenk 1997).Information which is simple, easily recalled, dramatic and appeals to already accepted information, is likely to spread faster in and between connected information groups, than information which is complicated, boring, uncertain or contradicts bias, even if that information is more accurate.In information society, bad information can drive out good (Marshall et al 2015, pp. 100-102).Data smog intensifies the importance and effect of information groups, as information groups help filter and frame the smog, reducing the quantity of information to save time and allow social action and support identities.
Accuracy of information can be undermined by strategies to cope with quantity.

Orders of Doubt: Trust vs. Distrust
In data smog, as reinforced by group separation, it appears that people in information outgroups are hopelessly biased, or deliberately trying to deceive (Sunstein 2009;Marshall 2013b).Their news is easily labelled 'fake news' because it does not agree with the groups' existing information, and its accuracy does not have to be checked.Arendt notes a similar phenomenon in the 1960s US administration's dismissive attitude to accurate intelligence reports on events in South East Asia.Preserving the administration's world-view, status and course of action, was more important than accuracy, or adjustment (1972, pp. 13-22).Similarly, the current President of the USA is known for condemning 'hostile' news as 'fake news', while apparently allying with Fox News more than his intelligence services (Tovey 2017a).
Such processes are reinforced when a person's status and competence depend on informational mastery.In an information hierarchy, it may be difficult to have recognition as being high status, authoritative and wrong (Marshall 2013b, p. 8).This makes it harder for high-ups to learn from mistakes.In my experience of political discussion online, people resent being told they need specialised knowledge to understand something, apparently hearing such statements as part of an outgroup's power grab.People appear to routinely think of themselves as independent thinkers and not easily mislead; this is a mark of cultural status, and reinforces both ignorance of outgroup information, and allegiance to a misleading information group.
Pronouncements that other groups are lying, save time, protect group allegiances and remove any need to put energy into interpreting the messages of outgroups.However, this ongoing process of doubt leads to 'information paranoia'; the suspicion that counterinformation is likely wrong or deceitful (Marshall 2013b;Marshall et al 2015).
This creates further problems, as no group can avoid mistakes completely, and when inaccurate information from a person's own group is recognised, it is generally classed as unimportant and does not falsify what group members think of as generally correct.When people find out that those on their side actively lie, they can say 'all politicians, or all news sources, lie.The other side is worse', and relax.They can come to doubt everything retrospectively, making an 'order of doubt', in which they can be disparaging, distrustful and paranoid about information without changing information groups.They can still hold that their group's information is better than the outgroup's (Marshall 2013b).Changing their information group, not only risks joining a group they distrust, but risks losing their allegiance and identity categories, being attacked, or dealing with an overwhelming mass of 'unfiltered' and 'hostile' information.Changing might lead them to lose their place in the world.

Doubt and Trust
Trust, is a complicated theoretical issue.Trust may be a beneficial social resource (Giddens 1991;Kramer and Cook 2004, pp. 1-2), but that does not mean it is generally present or always fostered (Kohn 2008, pp. 76ff.).The literature, does not show a unified definition of trust or even an agreed upon procedure for its analysis.Kohn writes of the difficulties of definition, saying that the word has a 'community of meanings ' (2008, p. 8), like Wittgenstein's 'family likeness,' with many fluid possibilities making specification of an 'essence' of trust difficult.The idea of trust clearly implies suspicion or distrust, and contexts may arise in which people operate with levels of relative distrust (A distrusts B, less than A distrusts C), which is not complete trust; people may be trustable in certain ways and not in others.With an order of doubt some unstable and changing level of distrust or trust can always be justified, depending on relevant contexts.This suggests trust may not be a thing, or state, but a shifting process (Moellering 2013), or a cultural construction, different amongst different groups.If so, then what we might call 'trusting' becomes a local cultural way that people frame other people's behaviour and their expectations of that behaviour in different ways.It is an interpretation.Despite these problems of analysis, it is often suggested that levels of trust are low in contemporary social institutions (Hosking 2014, pp. 1-5), although Hosking adds that despite 'the current crisis' 'there has never been a golden age of trust' (ibid., p. 5).Polls attempt to measure trust in contemporary institutions are common.Thus a recent Gallup poll claimed: American's trust and confidence in the mass media 'to report the news fully, accurately and fairly' has dropped to its lowest level in Gallup polling history, with 32% saying they have a great deal or fair amount of trust in the media.This is down eight percentage points from last year… [The number of] Republicans who say they have trust in the media has plummeted to 14% from 32% a year ago (Swift 2016).
This may have something to do with media questioning of Donald Trump during that period, leading to that media being perceived as a hostile outgroup.
Contemporary lack of trust can be said to be generated by neoliberal support for corporately dominated 'free markets' in which the only thing to be trusted is people's monetary self-interest (Hosking,pp. 142ff.),although there are many other causes of distrust (ibid., pp. 9ff.).
For the purposes of this paper, it can be hypothesised that if we frame a person or organisation's information, or their intentions towards us, as relatively 'trustworthy' or 'untrustworthy' (and this depends on the information we, or our group, accept about them), then we are more likely to accept, or reject, further information from them.In this sense the process of trusting depends on experience and alignment with others.This often means we categorise the relatively trusted as similar to us, or as allied with us, as Trump supporters may see him as an ordinary hard working American, being attacked by the elites who attack them (Tovey 2017b).This trust is relative and not always coherent.For example, despite President Trump's high levels of support amongst Republicans, Gore (2017) reports a Quinnipiac University poll finding that a 'total of 73 percent of American voters say President Trump and his Administration make statements without evidence to support them 'very often' or 'somewhat often'.'In an order of doubt, people ally themselves with others, but deny trusting them all the time, to maintain that allegiance.Voters may distrust their side's candidate, (or rather trust they will deceived sometime) but distrust them less than the outgroup's candidate.While there may be ambivalence about one's own groups, distrust in information or communication from framed hostile outgroups seems firm.
Giddens implies that non-face-to-face, or disembedded, societies, have to trust in experts and the dependability of social systems to function.We cannot gain knowledge of everything, so we can remain ignorant of how things work specifically.While peripheral to Giddens' theory, this process produces ambiguities, and he remarks: [A]mbivalence ... lies at the core of all trust relations, whether it be trust in abstract systems or in individuals.For trust is only demanded where there is ignorance... [and] ignorance always provides grounds for scepticism or at least caution (1991, p. 88).
It is also possible that trusting can also distort thinking and communication processes (Erdem 2003), by reinforcing information group distortions.As a report written for the Australian Institute of Company Directors states: 'Many of the factors that support trust simultaneously support the production of biases' (Kay and Goldspink 2017, p. 9).Similarly trusting can be built through collaborative hiding of deceit or corruption and furthers such deceit (Shulman 2007, pp. 94, 157).

Capitalism and Information
Capitalism reinforces the trend for information to be produced and validated by economic, or political, criteria alone.Freely available accurate information (unless it supports corporate power) is either not valuable, or must be 'enclosed' and restricted, so as to become valuable to its holder, or sellable to someone else (Marshall & da Rimini 2015).Relatively accurate information about the economy, the functionality of the company, or about particular economic acts, is restricted as a normal part of the strategic dynamics of that economy.If you know a company, or government's actions before 'the markets' do, then you can buy or sell in anticipation.Misdirecting others with inaccurate information, increases the benefit of any accurate information possessed.The main virtue is private profit and there is no advantage to making all information you have public.
In capitalism, advertising (or producing a front) becomes the model for communication.Few people expect an advertisement to be true but they will accept it if amusing or they like the product.However, the information exists in the order of doubt; it can always be declared false, or exaggerated, if needed.As people rely on the information about their business or capabilities possessed by others, a good front, 'appeal' and strategic effect are vital.The front offers evidence the person or organisation can fulfil obligations amidst uncertainty.Admitting mistakes can be costly, through court action as well as through loss of confidence in the markets.Financial incentives prevent accurate information circulating, or promote false gloss as occurred with the financial products which triggered the global financial crisis of 2008 onwards, when those selling and buying these products did not understand how they worked or the risks involved.The dangers were magnified by organisational silos and hierarchies, which produced lack of communication between traders, risk managers, modellers and software writers (Marshall 2013a).Similarly, if a category of product proves popular, then corporations will attempt to change the category without changing anything else, effectively deceiving for profit.For example, in Australia lobbying by large egg producers increased the density of hens 'free range' categories from a recommended 1500 to 10,000 per hectare (Han 2016).
Commercial in confidence, or private public, partnerships are a similar information arrangement allowing governments to hand taxpayers' money to corporations without revealing details of the agreement to voters.Lack of information about cost, content and responsibility, saves politicians the effort of producing consistent disinformation, and makes the lack of information hard to challenge.For example the information and cost calculations supporting the NSW's government's forced amalgamation of local councils cannot be seen by the public as it remains the property of the company that produced it (Robertson 2016).
Perhaps even more importantly, by filtering and framing all information by its contribution to profit or power, capitalists limit their perception of the problems they face.Price signalling is supposedly the heart of the information contained in a market (Bronk 2013).However, price does not tell purchasers about long term benefits, the trustworthiness of the dealer, or the usefulness or harmful consequences of the product.Focusing on price filters out that information.Immediate profit can be seen as more relevant than long term risk or destruction.
By making artificial boundaries, commodifying information, disrupting accurate information flow, building hierarchies, issuing strategic business enhancing information, and focusing on price, capitalism becomes embedded in disinformation.

Information Workers and Insecurity
In capitalism, workers both expect deception, and may routinely deceive in order to carry out their work (Shulman 2007).People know higher-ups cheat and lie, and may excuse it as normal for themselves (Callahan 2004).As Marx and Engels argued, all virtue and value is reduced to 'callous cash payment' (Marx & Engels 1969, p. 38); there is no 'honesty'.
Information workers are precarious, not only in the normal sense workers are precarious in capitalism, but because information society does not need that many artists, or cultural or informational producers, due to ease of replicating their work.A recording of one famous lecturer can replace thousands of not so well-known lecturers.Opinion columns, news or cartoons, can be syndicated.A few good artists, or information producers, can service millions of people globally.That audience might previously have supported tens of thousands, or more, local artists or information authorities with lesser stature, luck or promotion (Frank & Cook 1995).
It also appears likely that, as well as being replaced, or outsourced, much standard information work will soon be automated well enough not to require human workers; this is a standard news trope (Economist 2016).Academics are becoming more and more casualised, with no security to make 'fearless' or 'unpopular' analyses' and can be denounced if they do so (Collini 2017).Even those people employed to check information, and the accuracy of models, can be perceived as an unnecessary cost and removed, so system errors can mount up (Marshall 2013a, pp. 372-3).
The more precarious everyone's position is in the workplace, the more information distortion is likely to increase.'News' is manufactured by workers in an industry facing decline of income and a diminishing workforce.One journalist writes that the 'journalists' union… says that since 2011, more than 2,000 jobs have been lost in Australian media, or around a quarter of our journalists' (Alcorn 2017).Accurate investigation or reporting of events requires time and generates costs.Costs subtract from profits.These pressures may not allow full research, and encourage recirculation of articles or press releases from other sources with the time, such as PR firms, or corporately sponsored think-tanks.These sources are likely to practice strategic inaccuracy.Not wanting to offend advertisers or shareholders probably leads to self-censorship due to financial insecurity (Davis 2008).
These pressures mean information has to sell well, attract advertisers, and please higher-ups in the media organisation; attracting eyeballs and advertisements is more important than accuracy.This can lead to strange results, showing propaganda is not always intentional, but driven by commercial success and information group strength: [A] Macedonian town of 55,000 was the registered home of at least 100 pro-Trump websites, many of them filled with sensationalist, utterly fake news… These Macedonians on Facebook didn't care if Trump won or lost the White House.They only wanted pocket money to pay for things [and] Trump groups seemed to have hundreds of thousands more members than Clinton groups which made it simpler to propel an article into virality1 (Subramanian 2017; for normal news see Davis 2008).
People know this kind of distortion happens and so their information paranoia increases and believable sources have to be artificially created for commercial purposes.This leads to information fakery such as: viral marketing, buzz marketing, push polling, paid trolls, paid but apparently neutral contributors, think-tanks etc., and produces more distrust (Marshall et al 2015, pp. 102-5).In this situation, there is no neutral source; as accepting an external source as neutral, implies possible admission of error and loss of status.Fact checkers become interpreted as part of the political process (Greenwald 2011).As news becomes less profitable, it becomes the plaything of the hyper-rich, or confederations of industry, geared towards promoting their perceived 'interests' rather than to promoting accuracy (such as Rupert Murdoch, Robert Mercer and Steve Bannon).
In capitalism as workers become precarious, hiding failings, lying, or misdirection becomes habitual in the workplace.Deceit is part of daily survival.

Networking Complexity and Distributed Governance
The more networks of communication and connection, the more complexity is generated.This complexity can involve meshing networks of different types: computer networks, electricity networks, water networks, traffic networks, economic networks, etc. Linkage between different networks, or types of networks, increases the possibility of destabilising complications.The higher the efficiency of transfer in the network, the higher the likelihood of crisis, 'bad information', viruses or worms, moving quickly through the system.Networks also allow maintenance and control to be transferred to distant people, who may not understand the local situation and its requirements.Networks neither guarantee reliability of information, nor stability; they produce complex instability (Marshall et al. 2015, pp. 69-76).
Networks encourage distributed governance.In distributed governance, large numbers of players, with similar and dissimilar interests or opportunities, interact.Power is distributed, responsibility is never clear (or avoided) and unrecognised connections and feedbacks are routine.Power always seems to reside elsewhere, as actions rarely carry through as intended, without resistance, confusion or dispersal (Marshall et al 2015, pp. 58-62).Even the powerful can feel that the world escapes them (Naim 2013), and consequently everyone can feel that they are a victim of spiteful resistance, subversion or obstructive authority.In this situation of uncertain responsibility, avoidance of blame becomes important.People may, as Christopher Hood (2011) argues, try to deflect blame by using such techniques as denial of the consequence, positive spin, shifting the blame elsewhere (usually down the hierarchy), using public/private partnerships, build conflicting jurisdictions, claiming to be following rules, or using suboptimal policies that minimize risk of liability (ibid., pp. 47, 67, 90).As a result, 'loss of accountability comes to be combined with loss of welfare for service users ' (ibid. p. 173).Misinformation becomes standard in government and externally.Hood claims that blame avoidance procedures shape 'the conduct of officeholders, the architecture of organizations, and their operating routines and policies' (ibid., p. 4), but it is equally plausible that these procedures arise and work within the difficult context of distributed governance.Distribution of governance does not mean there are not patterns of dominance.Network capitalism appears to lead to an extreme hierarchy in terms of distribution of wealth and ownership (Marshall et al. 2015, pp. 65-7).This hierarchy is not only likely to disrupt information flow, but as there are only small numbers of people at the upper levels, these people are vulnerable to the quick spread of fads, infections, disease, or information distortion (ibid., pp. 67-9).The malfunction of a widely connected node can have widespread effects, with failure in the system tending to compound.

4
Cosmopolitan Democracy and its Problems These problems impact cosmopolitanism, which becomes a paradoxical process of communication, not a fixed state.It is in flux, like every other complex system, composed of other complex systems, which may not mesh well.It is like most other social processes, fraught and difficult, and recognising the paradoxes may be useful.
With no recognised differences, differences are suppressed, but where differences are recognised, then ethnic (and other) 'communities' form categories and information groups and oppose themselves to, and distinguish themselves from, other categories and information groups making 'good' communication hard.Emphasising these differences, to unite the supressed, is also likely to increase opposition and resistance to change.It may make groups involved more rigid and less fluid.Furthermore, these groups are entangled in the disinformation processes of capitalism, hierarchy, siloing, data smog, information paranoia, and troublesome sources.Distributed governance will tend to mean that people find unexpected resistances to their efforts, with unintended effects likely.Consequently, cosmopolitan aspirations are always subject to the possibility of disruption or failure, and those aspirations may even make failure more noticeable and cause resistance to cosmopolitanism to spread on all sides.
Hierarchical arrangements in and between groups undermine the possibility of equity of exchange, accuracy of interpretation and participation.While it is probably not possible to be without any hierarchies, communication may work better if the hierarchies are low and relatively non-punitive.Yet attacks on hierarchy can be also interpreted as an attack on the cultural tradition which makes the groups, whether this involves hierarchy in the group or between the groups and their outgroups.Dominant subgroups may not voluntarily give up their position in the established hierarchy, especially when they consider the domain as 'theirs' or their power and status as being necessary for group cohesion and safe interaction with other groups.Indeed, hierarchy may be necessary for those very reasons; it can be both an obstacle and a support.
Even without hierarchies, it is hard to communicate across groups when cultures and framings are different.Consequently, we need to understand the differences in the framings people bring to communication, and the events tending to prompt disruptive, or radically divergent, framings.Then it may be possible to help all sides frame events beneficially to the process, although this presumes that some communicators can be recognised as neutral, which may also be difficult.Even if this is possible, then bringing groups which previously ignored each other into communication may increase the possibility of conflict, especially when other factors are involved, such as: silos of occupation or religion, or through framings of a history of conflict, suspicion and so on.Increased communication may not increase harmony, but conflict.Trusting between groups does not have to be absolute for relative trusting to develop.Demanding trust at the beginning may be counter-productive.
Accuracy of understanding may be essential but can be disruptive, as some inaccuracy and falsehood evolves to smooth action between groups and people.Cultivated conscious ignorance could be useful in promoting harmony, as people surrender what they already 'know' about the others and use to frame their communications.However, such ignorance can lead to offence, and may already be part of the problem.
Cosmopolitanism requires people to risk the disapproval of their own groups, and to try and act as bridges between groups, even though they risk being misunderstood on both sides with resulting exile or violence.This is a lot to ask.The solidity of the social categories in play may need to be challenged as, in reality, there may be considerable fuzziness, overlap, intermarriage and so on, which may weaken desired social categories if recognised.This is again difficult to do without offence, as it may challenge ingroup status and relationships.In a company you might break the silos, as Tett states (2015, pp. 246ff.), by moving people between them, but that is difficult in a democratic society, with no recognised neutral actors, and may be interpreted as an attack on culture, which it could well be.Similarly, it will be difficult to set up any information coordination between groups, or to keep boundaries between groups permeable and flexible, if groups have been busy separating.
Peckham implies (1979) that some of these problems can be overcome by people working together on a material project.Material nature and consequences can force people to check whether their intended meaning has been received, and to learn how to correct misunderstandings.Relative equality, so counter information is not penalised, may also help.Perhaps, as suggested by Sloman & Fernbach (2017) it can be useful to ask people to explain how various policy ideas would actually work as this helps open people to their relative ignorance, and can soften political extremism and certainty.It might also be useful to deliberately explore the unintended consequences of policies and actions.An understanding of complexity may also prove useful in stopping expectations that there will be an end to the processes of communication, action, change and cosmopolitan action.In a complex system, we need to accept being unable to predict the detailed results of actions.Increasing, or enforcing, theoretical 'order' or harmony can produce what is perceived as increased disorder and disharmony (Berg & Timmermans 2000;Marshall et al. 2015).This calls for an experimental politics, in which we pursue actions and test the results by asking 'what disorders and unexpected events arose along with the identifiable results of our actions?' or 'how might these disruptive events have arisen through what we have done?' Theories may work up to a point, and then be harmful if extended or intensified beyond that point, and that point can only be established by continuous observation.Not avoiding 'blame' seems central to this process, but may be undermined by techniques developed in distributed governance.Good communication does not arise through calls for trust, praise of hybridity, or requests for dialogue alone; but through greater awareness of the promotors of failure.

Conclusion
Contemporary 'information society' is not 'knowledge society'.Deceit, 'fake news', data smog, and information blockage are part of people's day-to-day life.This arises because communication has social functions other than the transmission of accurate information.In information society, communication becomes primarily strategic and identity based.People invest heavily in information groups around identities and meaning, and develop 'information paranoia', embracing an ordering of doubt or suspicion.They tend to become automatically suspicious of other groups, and even of their own 'side', although the suspicion of their own side largely allows them to continue to support the views and alliances they already have.
Categories, framing, lying, persuasion, group bonding, group identity, and information groups are necessary for smooth social functioning and for interpretation, and selection, of information, however they all produce distortions.These tendencies towards distortion of communication and information seem reinforced by power differentials and other organisational structuring.Good information flow may require relative equality, and lack of existing hostility, but this is not present in the world's larger societies today, if it ever was.Capitalism and the precarious workspace, intensify these problems.However, because of these patterns, we cannot assume that dominant groups are well informed or know what they are doing.The web of power may appear seamless, coherent and impossible to challenge, but it is self-disruptive, disjointed, incoherent and vulnerable.The power/ignorance nexus has two possible consequences.Firstly, the less the dominant really know what they are doing, then the greater the tendency for them to enforce their ideas and generate unintended consequences.Secondly they are potentially vulnerable through that ignorance and the instability it brings.They may be less able to perceive any coming overthrow.
Working with communication requires an awareness of the paradoxical processes that lead to 'bad' and 'good' communication, with an awareness that disorder and unpredictability are inevitable, and that such good communication may require an experimental attitude, with constant checks, as to the actual results.Good communication is difficult and may be undermined by processes developed to help it.