Business as Magic

Sarah Perry is a contributing editor of Ribbonfarm.

---

Vienna, late 18th century, alternate timeline.

In the beginning, there was just the tinkerer. One night, after attending a magic show at the palace, he got drunk and made a bet with a rival courtier. A few months later, he unveiled a marvelous contraption before a palace audience.

The contraption was an automaton, dressed as a Turkish sorcerer. It sat at a desk filled with complicated gears and levers, with a chessboard on top. The Turk played a decent game of chess, beating dukes, princes, and visiting American statesmen. Its reputation spread, and with it the reputation of Vienna as a hub of technological development.

Tinkerers from Austria to America set about making their own Mechanical Turks (as well as Mechanical Russians, Yankees, and chess-playing shepherdesses). Only a few succeeded; most were in Vienna. Of those half-dozen that succeeded, all were known to be acquainted with some down-on-his-luck chess master who, incidentally, was not overly tall or rotund.

Vienna’s reputation continued to spread, and it became fashionable for wealthy patrons to support chess automata and exhibit them. Unfortunately, one of the impoverished chess players got the flu. During his performance, such a loud fit of coughing emanated from his automaton’s desk that the audience was scandalized. The Turk, a hoax! The entire reputation of Vienna was on the line.

To meet the crisis, the wise Empress created a Bureau of Chess Automaton Standards, and appointed a Secretary of Chess Automaton Quality Assurance to head it. He in turn hired many Chess Automaton Inspectors. But both they and the tinkerers faced a problem: all the chess automata were fake. There was a guy inside playing chess. They could not pass a penetrating inspection. In consultation with the tinkerers, they arrived at a solution that met the needs of all parties.

The Automaton Inspectors were taught a ceremonial procedure for verifying the legitimacy of a chess automaton. For the first part, tinkerers were required to present a PowerPoint describing the workings of the automaton to the Automaton Inspectors. The Automaton Inspectors would then perform a cursory inspection of the machine. For the second part, conducted the next day, the Automaton Inspectors would test the chess ability of the automaton to ensure that it had an Elo rating of at least 2000. (This rating would go up a bit every few years, demonstrating the progress of Vienna’s mechanical intelligence industry.)

Meanwhile, savvy firms added sets of noisy gears to their elaborate-looking apparatuses. The noisy gears had no purpose other than to cover any sound the secreted human operators might unintentionally make.

Most of the tinkerers quickly created two new departments: a Department of Mechanical Intelligence, whose job was to maintain and improve the complicated-looking mechanism and explain it to visiting Chess Automaton Inspectors, and a Department of Chess Strategy, whose ostensible job was to continually improve the chess-playing abilities of the automaton to keep up with Bureau standards. The tinkerers (now firms) that did not implement these changes tended to be exposed as frauds, losing their income and status and bringing shame upon their patrons and on Austria.

The firms now had an ostensible (narrative or mythic) function, and a real, practical function. The mythical function was to promote the glory of mechanical intelligence – a strategically important myth to Austria’s international standing. The practical function was to maintain (and improve) an illusion.

These ceremonial inspection practices benefited firms (at least, those who evolved Departments of Mechanical Intelligence and Departments of Chess Strategy) by providing them legitimacy, and benefited the Bureau of Chess Automation Standards by providing income, meaningful-seeming work, and status. However, firms that got on the wrong side of the Empress could be subjected to Extraordinary Review, a more demanding inspection process that necessarily resulted in the scuttling of the false automaton and the disgrace and dissolution of the firm.

Many of the scientists and engineers in the various Departments of Mechanical Intelligence did not even know that the automata were not real. Careful precautions were taken to limit the “testing” of the automaton’s functionality. The secret was kept by only a few insiders who acted as liaisons between the Departments of Mechanical Intelligence and the clandestine human operators.

Sometimes the Departments of Mechanical Intelligence produced real results—quite by accident. One clueless engineer, inspired by his automaton, invented a mechanical weaving machine. Another invented a device that could play voice recordings, enhancing the appeal of the automaton. “Échec!” the automaton would cry. Inspired by the talking machine, an American invented the telephone. The reputation of Vienna was all this time enhanced.

However, the Departments of Mechanical Intelligence mostly produced increasingly elaborate mechanisms that didn’t do anything, and increasingly elaborate theories about how the mechanism might work, if it did work, which it didn’t. Universities began to offer programs in Mechanical Intelligence, with curricula often based on the theories developed in the chess automaton firms. Universities supplied graduates to the firms, professionalized the discipline, and further bolstered the legitimacy of the enterprise.

Everyone lived happily ever after. Perhaps they eventually invented the internet, quite by accident.

What Institutions Are Up To

I have worked in corporations, nonprofit entities, and government agencies. They have always been extremely mysterious to me.

I could never quite figure out what the institutions were up to—their goals, workings, and desired behaviors were never clear to me. Mission statements read at meetings seemed to have little connection to the day-to-day activities we performed. What did they want? What were they doing?

I will share here a sociological theory of institutionalized organizations that helped clear up my confusion. In this model, institutions have two competing interests. On the one hand, they need to be efficient at producing their product or service. On the other hand, they need to be legitimate and to fit into myth-driven regulatory and structural frameworks. In response, institutions decouple their mythical and practical activities. They not only pay lip service to myth, but structure their operations and activities toward signaling conformity with it. Backstage, their members accomplish the necessary tasks using methods often at odds with their institutional myths.

This theory is a bit of a Grand Unified Theory of Everything, and vulnerable to failure modes that such theories invite: grandiosity, incuriosity, certainty. But it also provides a framework for curiosity (like McLuhan’s tetrads, laws in the form of questions): does it apply in this situation? What do the components of the theory correspond to, in concrete terms?

One positive function of a myth is to reduce cognitive complexity and uncertainty. This theory might be one of the very myths it describes. In its favor, it invites criticism and disconfirmation.

But once you see it, you can’t un-see it.

Institutional Myths

I have chosen an instance of active deception as the institutional myth in my imaginary Vienna. But falsity is not the defining characteristic of an “institutionalized myth” or “rationalized myth.” This kind of myth can be identified on four criteria:

1. Language of Morality

Beliefs about trivialities are not myths. Institutional myths “refer to widespread social values and emphatically use semantics coming from those value systems” (Piber & Pietch, 2006). Myths deal with sacred matters and moral consideration, and often use language borrowed from religion.

2. Zone of Ignorance

Myths are protected from falsification or excessive inquiry by social forces. “Follow the sacred and there you will find a circle of motivated ignorance,” says Jonathan Haidt. Linda Degh (1994) describes

legend-like narratives that enforce belief and that deny the right of disbelief or doubt, narratives that express majority opinion and are safeguarded by moral taboos from negation and, what is more, from deviation.

Myths are also often taken for granted, accepted as part of unalterable reality, exempt from questioning because hardly noticed.

3. Idealization and Ambiguity

Myths are not merely falsehoods, but idealized, simplified accounts of complex matters. Myths attach to values that are difficult to precisely define and difficult to measure. A myth is so vague and uncertain that its epistemic status is in some ways unknowable. When government bodies enact legislation concerning these values, it is usually ambiguous: ambiguity is a useful feature, because more voters can agree on an ambiguity than on any particular set of specifics. Institutions cope with the uncertainty of interpretation by evolving orderly ceremonies (often in cooperation with regulatory agencies) and by copying the behaviors and structures of successful firms.

4. Symbolic Evaluation

Since myths deal in subjects that are inherently ambiguous and in some sense unknowable, the evaluation of myths is “restricted to symbolic considerations, which leave many interpretations possible” (Piber & Pietch, 2006). Symbolic ritual activity is directed toward the sacred value, with no demonstrable causal connection to any particular measurable outcome. Empirical falsification is inherently difficult, and therefore can be avoided.

Example: National Security

The cultural myth of national security attires itself in the language of morality: terrorism, the Axis of Evil. It has been protected from scrutiny by the invocation of a sacred event, 9/11. It is ambiguous enough to escape precise definition, and is capable of wide interpretation, from the USA PATRIOT Act to Guantanamo Bay. And it is evaluated symbolically, for instance with adjustments of the “threat level,” that have no clear causal connection to any particular outcome. Are we secure? There’s no way to know.

The benefits of myths are that myths reduce cognitive and coordination costs. Since they are simplified, idealized accounts, they don’t take up much cognitive processing power. They are easy to share, and make excellent targets for coordinated activity.

The drawbacks of myth are that such idealization, simplification, and elision turns the world into black and white. We want to do good, but we can’t tell what that means in a quantifiable and empirically falsifiable way. Myths are targets for wishful thinking. Just as myth can underpin positive coordination, it can facilitate coordination that does not actually serve the interests of the group.

Legitimacy

Selling groceries may make your business a grocery, but selling drugs does not necessarily make your business a pharmacy. So says Mark Suchman (1995), and offers a taxonomy of the types of legitimacy that institutions rely on. Suchman defines legitimacy as

a generalized perception or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs, and definitions.

Myths provide opportunities for institutions to claim legitimacy, and the types of legitimacy have parallels to the four features of myth described above.

Practical Legitimacy

An institution achieves practical legitimacy when its audience (customers, regulatory agencies, etc.) believe that it is acting in their interests. This is a “selfish” basis of legitimacy. People even evaluate the disposition of an institution, as if it were a person, projecting onto it the human capacity for being positively or negatively disposed toward others. Do they like me and my in-group? Are they acting in ways that benefit or hurt us? Institutions give gifts and otherwise signal loyalty to their audience in order to assure them of benevolence and positive regard. They even seek input from their audience, though they usually channel that input into forms that won’t be disruptive. Administrative agencies must solicit public comment on proposed rules, for instance, but are under no obligation to make decisions based on public opinion.

Institutions often have no idea how to achieve practical legitimacy (or the other types). When your calculator asks you to rate it, or when your insurance company sends you a birthday card, they are flailing around in random attempts to establish this kind of legitimacy. (Mimesis, copying what other successful institutions do, is a common response to uncertainty, whether the practice is effective or not.)

Moral Legitimacy

Moral legitimacy comes from less selfish evaluations: the moral rightness of an institution, including its actions, structure, and perceived “character.” Institutions structure themselves in the “correct” manner (for instance, having a Department of Research and Development or a Department of Mechanical Intelligence, or having a Board of Directors or an approval board for research on human subjects). They employ professional staff where appropriate.

If a representative of an institution does something morally reprehensible, the taint may rub off on the institution; to preserve the moral legitimacy of the institution, the offending leader (of the company or the country) must resign. Institutions can signal allegiance to moral values in symbolic ways: donating to charities, advertising allegiance to causes, or altering workplace dress (on occasion).

Cognitive Legitimacy

People have limited cognitive resources. When an institution is easy to understand (at least superficially), as opposed to jarring or mysterious, it seems more worthy of trust. Once an institution has been around for a long time, we may come to take it for granted as part of the background, and not question its legitimacy at all. Even new institutions can seem inevitable and be taken for granted if they array themselves in the structures, ceremonies, and definitions of previously taken-for-granted institutions.

Institutions that seem natural and meaningful are perceived as legitimate. This is the most subtle form of legitimacy: to be comprehensible, to mesh with a larger belief system and the experience of everyday life; and to be taken for granted, so that “for things to be otherwise is literally unthinkable” (Suchman 1995). Older people may grumble privately about rituals of security prior to airline travel (they would be advised not to protest while in line to be groped!), but younger people have no memory of a time when this was not done.

Suchman produces a highly compressed and appealing 2X2X3 (!) that I felt compelled to adapt, if only as a ceremony directed toward the values and aesthetics of the Ribbonfarm readership:

Decoupling

The need for legitimacy is often in direct conflict with the need for efficiency in technical matters. Getting the job done often requires acting contrary to legitimizing myths. Institutions deal with this tension by decoupling their myth-serving ceremonies from their practical, “backstage” actions.

In my alternate Vienna, the proprietors of chess automata decoupled their legitimacy seeking from their practical activities by the creation of specialized departments to interface with the regulatory agency. Behind the scenes, things happened much differently from how they were reported. Most employees were not even aware of this.

Recently, it was reported that the Transportation Safety Administration paid IBM 1.4 million dollars for a randomizer app to tell people which lane of the security check to enter. But they were not paying for a randomizer app. TSA was paying for ceremonially legitimized randomization, in service of the Myth of National Security. IBM, the supplier, provided not only the app, but institutional structure and interfaces: security clearances, regulatory compliance departments, lobbyists, and institutional legitimacy. The decoupling here is so extreme that the service of myth outweighs the technical work by many orders of magnitude.

The decoupling process was first identified by Meyer & Rowan (1977) in their paper “Institutional Organizations: Formal Structure as Myth and Ceremony.” It has over 20,000 citations on Google Scholar, and inspired a whole field (sometimes called Neo-Institutionalism). When decoupling occurs,

Activities are performed beyond the purview of managers. In particular, organizations actively encourage professionalism, and activities are delegated to professionals.

Goals are made ambiguous or vacuous, and categorical ends are substituted for technical ends. Hospitals treat, not cure, patients. Schools produce students, not learning. In fact, data on technical performance are eliminated or rendered invisible. Hospitals try to ignore information on cure rates, public services avoid data about effectiveness, and schools deemphasize measures of achievement.

Integration is avoided, program implementation is neglected and inspection and evaluation are ceremonialized.

Human relations are made very important. The organization cannot formally coordinate activities because its formal rules, if applied, would generate inconsistencies. Therefore individuals are left to work out technical interdependencies informally. The ability to coordinate things in violation of the rules—that is, to get along with other people—is highly valued.

Meyer & Rowan, 1977, p. 375.

When I think of my befuddlement at the behavior of the institutions I have worked in, or the frustration of my engineer friends who work in highly institutionalized environments, this model gives me the sense of dawning comprehension. Institutions have at least two goals, and they are in conflict. The institution tears itself in half in order to simultaneously pursue conflicting goals. This is why their activities and demands seem inscrutable.

When you can identify the sacred myth, and identify the legitimation ceremonies, and notice how they conflict with the practicalities of getting things done, then you can start to make sense of the conflicting and often irrational-seeming actions and demands.

Sociopaths, Losers, and Clueless in Alternate Vienna

I close with an application of a classic theory of Venkat Rao’s: the trichotomy of sociopaths, losers, and clueless. An institution begins with a sociopath; he is the tinkerer in my alternate Vienna. He is conscious of emerging cultural myths, but is not under their thrall. The sociopath then recruits a loser – the chess player who has fallen on hard times. The loser does not believe in the cultural myth, either, but nor does he see a path to personal fulfillment. He does his job in despair.

(By the way, the coughing incident described above actually did happen, to a chess player named Hyacinthe Henri Boncourt. His coughing embarrassed the owner of the Turk, who really did install noisy gears with no purpose other than to cover the sound that future losers might make while performing the magic trick. I think “noisy gears” is an excellent metaphor for institutions installing camouflage to hide potential evidence of impotence, deception, and decoupling.)

As institutionalization occurs, the sociopath must then recruit the clueless – people who to some degree believe the myth, and will act on the mythical interests of the institution. In alternate Vienna, these are the members of the Department of Mechanical Intelligence. The sociopaths act as managers and liaisons between the clueless and the loser, using “people skills” to creatively get things done across the chasm of the decoupling.

Business as Magic

As I have said, the model I outline here inspired an entire field. It is such a fun game that it has been applied to hundreds of purported myths, including salary surveys, informed consent, open source, worker protections,job evaluation, and null hypothesis significance testing. One scholar traces the effort to recouple myth and practice, in the case of an elementary school that had to reorganize its practices based on new legislation about teacher accountability. (Result: everyone was mad.)

How does this theory do, on its own terms? Meyer and Rowan did make predictions in their 1977 paper, although these might be described as vague. For instance:

Organizations which incorporate institutionalized myths are more legitimate, successful, and likely to survive. Here, research should compare similar organizations in different contexts. For instance, the presence of personnel departments or research and development units should predict success in environments in which they are widely institutionalized. Organizations which have structural elements not institutionalized in their environment should be more likely to fail, as such unauthorized complexity must be justified by claims of efficiency and effectiveness.

Experimentally, one could study the size of the loans banks would be willing to provide organizations which vary only in (1) the degree of environmental institutionalization, and (2) the degree to which the organization structurally incorporates environmental institutions. Are banks willing to lend more money to firms whose plans are accompanied by econometric projections? And is this tendency greater in societies in which such projections are more widely institutionalized?

Meyer and Rowan have inspired a major research program. Some of the research uses postmodernist buzzwords (which are, after all, compressed myths), but there is much diversity in method, and both qualitative and quantitative research. The paper claiming myth status for null hypothesis significance testing has a higher density of citations per paragraph than I think I’ve ever seen in a sociology paper; perhaps, in attacking one basis for legitimacy, the author felt it proper to provide his audience with an alternative ritual basis for legitimacy.

Some have described the myth-and-decoupling theory as cynical. I think it gives us an image of entrepreneur-as-magician. This is an ambiguous image: one the one hand, practicing guile and deception for one’s own benefit; on the other, pulling off a complicated magic trick that audiences come back to see. The research and development of noisy gears (and other illusions) is frequently an ignoble profession, but it need not be. Magic is a necessary business skill. And that’s how we got the internet, quite by accident.

---

Degh, Linda. Tape-Recording Miracles for Everyday Living. In American Folklore and the Mass Media, Indiana University Press, 1994.

Meyer, John, and Brian Rowan. Institutional Organizations: Formal Structure as Myth and Ceremony. American Journal of Sociology 83:2:340-363, September 1977.

Piber, Martin, and Gotthard Pietsch. Performance Measurement in Universities: The Case of Knowledge Balance Sheets Analyzed from a New Institutionalist Perspective. In Performance Measurement and Management Control: Improving Organizations and Society Studies in Managerial and Financial Accounting 16:379–401, Elsevier 2006.

Suchman, Mark. Managing Legitimacy: Strategic and Institutional Approaches. Academy of Management Review 20:571-611, July 1995.