The ethical dimension of complex decision making – a metalogue

Salviati: Hello Simplicio, it is good to see you again, my friend. What are you up to these days?

Simplicio: Good to see you too, Salviati.  I have been doing a class on decision making as part of my MBA degree. It is nice to go over something one already knows well…and even nicer that I was able to convince my employer to pay for the course.

Saliviati: Well done!  What is the most important thing you have learnt so far?

Simplicio: Let me see…um, I think it is that decisions should be based on hard facts and evidence.

Salviati: Hmm. It is true that most decision-making approaches will tell you to focus on facts and evidence – that is, what you know or can find out about the problem. However, it is often difficult to find hard facts and evidence, particularly in situations where facts are scarce or contested. In such situations, you have no choice but to pay attention to what you don’t know, the uncertainties.

Simplicio: That does not make sense. How can you pay attention to something you don’t know?

Salviati: Let me give you an example. You have done projects before, haven’t you?

Simplicio: Of course, you know from our earlier conversation that I’m a project manager by profession.

Salviati: OK, so what are the key variables in a project?

Simplicio: Time and cost, two apexes of the iron triangle.

Salviati: That’s right. Whatever the project might be, you know that time and cost are key variables. However, you do not know their values upfront. Your job as a project manager, is to come up with good estimates for these. Is that right?

Simplicio: Uh huh.

Salviati:  OK, so you need to understand this uncertainty and to do so you must pay attention to it. The interesting thing about uncertainty associated with time and cost is that it is quantifiable. That is, you can develop numerical estimates for these variables using a range of techniques like, say, Monte Carlo simulation. The point I want to make is that regardless of the technique used, quantifying uncertainty involving known variables is a rational and logical process. Would you agree?

Simplicio: Yes, that makes sense.

Salviati: Now consider another problem, that of formulating a business strategy. What are the key variables in this case?

Simplicio: Hmmm…that’s a difficult one. It depends on several factors, the financial position of the organization, market share, the environment, business forecasts…oh, so many things.

Salviati: What information would you need in order to figure out which of these factors is important?

Simplicio: Oh, that’s impossible to tell without knowing more about the situation. There are so many things that could be important. You need to know a lot more about the business and its operating context before you can figure that out.

Salviati: Yes, that’s true, and I should also point out that no amount of data, number crunching or logical analysis is going to get you anywhere until you figure out what is important. Would you agree?

Simplicio: Yes.

Salviati: So, let me ask you: what do people in your organization do when they develop their business strategy?

Simplicio: Ah, they ask the experts of course – they engage Big 4 consultancies.

Salviati: Right…who better than a rank outsider to tell you what to do? At an inflated billing rate too! Surely there is a better way.

Simplicio: Like what?

Salviati: I’ll get to that in a bit, but I first want to make another point. A few minutes ago, I said that dealing with project estimation is essentially a rational and logical process. Let me ask you now: what kind of process do you think strategy formulation is?

Simplicio: What do you mean?

Salviati: Well, it isn’t logical…but it obviously isn’t arbitrary either. So, what is it?

Simplicio: I’m not sure I understand what you are getting at.

Salviati:  This is a difficult question so let me approach it in another way. We agreed that the difficult part in strategy formulation is to figure out what is important, right?

[Simplicio nods]

Salviati: So, let me ask you: important to whom?

Simplicio: to management, of course!

Salviati: Do employees not matter?

Simplicio: They do…but their job is to do as they are told.

Salviati: Really? You think it as Lord Tennyson noted: “theirs not to reason why, theirs but to do and die.”

Simplicio: That is over the top, Salviati. The situation of an employee in a modern-day organization is not comparable to the poor soldiers in the Charge of the Light Brigade. Nobody dies.

Salviati: May be not, but I think the analogy is justified. There are so many cases of ill-fated strategies that could have, should have, been questioned before implementation, but weren’t.  The consequences for employees, though admittedly not fatal, are disastrous…and the point is, employees are rarely given a voice in the decision. It is akin to the Charge of the Light Brigade.

Simplicio: OK, maybe it is, but what is your point?

Salviati: Strategy development and implementation ought to be treated as ethical matters rather than a logical ones.

Simplicio: Yes, I suppose they are. But that begs the question: how does one develop a strategy in an ethical manner?

Salviati: That, my friend, is the key question and it is a difficult one. The difficulty arises from the fact that ethics is hard to talk about meaningfully. Indeed, the philosopher Ludwig Wittgenstein went so far as to state that ethics cannot be articulated. His point is that the term “ethics” is meaningful only in the context of actions, not words. (Editor’s note: see this lecture by Heinz von Foerster for more on Wittgenstein’s take on ethics)

Simplicio: So how does one act ethically in this context?

Salviati: To answer that question, I will turn to Heinz von Foerster’s ethical imperative: act always to increase choices. (Editor’s note: see this lecture for more on von Foerster’s ethical imperative)

Simplicio: Umm…please explain.

Salviati: It is simple: if you do things that increase everyone’s choices then you are behaving ethically.

Simplicio: OK…but how do you “increase everyone’s choices” in practice?

Salviati: By involving them in the decision of course, and there are ways to do that. However, let me be clear, the aim is not to make a decision that satisfies everyone. That is impossible. It is to get as many different perspectives on the problem before making a decision. If you think about it, this is the only way to ensure that you do not miss important factors that could cause your decision to fall apart later.

Simplicio: That sounds idealistic and impractical.

Salviati: It might be idealistic, but it is not impractical. The basic idea is to make the different perspectives on the problem explicit.  This can be done by eliciting the preferred options of the different stakeholder groups and documenting arguments for and against them. There is a visual notation called IBIS (Issue Based Information System) that facilitates this. With IBIS, we can visualize the informal logic of conversation using three types of nodes: questions (or issues), which capture the problem being discussed; ideas, which are options offered by the different stakeholders; and arguments for and against those options (pros and cons).  By making the different viewpoints and the arguments explicit, you set the stage for making an ethical decision. Of course, one must also have in place the conditions that allow for open dialogue, and there are ways to do that as well.

Simplicio: So, the choices are matters of opinion, not fact? That does not sound right.

Salviati: Of course not. If someone makes a claim that needs to be validated or is factually incorrect, it can be challenged by others, even after the debate.

Simplicio: Ah, I see.  This is interesting, I’d love to see how it works.

Salviati: I’ll send you some books, papers and articles on IBIS (Editor’s note: here are some articles and a couple of books.). 

Simplicio: But I still don’t see how this makes the decision ethical.

Salviati: If you think about it, ethics is about doing what is good. The problem is if you try to define what is good and what is not good, you will tie yourself up in knots. It is impossible to come up with a meaningful universal definition of “goodness.” Wittgenstein was right when he claimed that it is impossible to talk about ethics.  If one cannot speak about ethics meaningfully, the only possibility is to do it… and that is what von Foerster’s dictum is about.  It tells us that ethical action is about doing things that increase choices for everyone. By eliciting multiple perspectives, you are increasing choices for the group. If the environment is one in which open dialogue can occur, then the choices can be freely debated by all and a decision reached. Even if the final decision does not make everyone happy – which it won’t, of course – everyone will agree that the process followed was inherently ethical.

Simplicio: OK, I think I see now. Since complex problems are multifaceted, one has to elicit diverse viewpoints on the problem to ensure one has not missed something important…and by doing so, one is also acting ethically.

Salviati: That’s exactly right! Incidentally, such complex, multifaceted problems are often called wicked problems. Much of the literature on wicked problems focuses on the surfacing and debating diverse perspectives, but very few writers (if any) comment on the inherently ethical nature of this process.

Simplicio: This is fascinating Salviati, thank you for broadening my perspective on complex decisions.

Salviati: My pleasure…but you should keep in mind that the process discussed will ensure that you surface and debate options comprehensively. The decision itself is yet to be made.

Simplicio: Oh! So, how does one make the decision.

Salviati: Unfortunately, there is no formula for that Simplicio. As you will appreciate, this is not simply a matter of picking the best option because different stakeholder groups may have different opinions on which one is best.

Simplicio: Hmmm, so what does the decision maker do if the group cannot settle on an option?

Salviati: Well then, the decision maker must make the call.

Simplicio: On what basis?

Salviati: I think I have already answered that. He must choose so as to maximise the number of future choices for all.

Simplicio: Hmm, we have already gone through that…

Salviati: There is no algorithm I can give you for this, if there were it would be a calculation not a decision – a matter of logic rather than ethics. All I can say is that, you, the decision maker must decide how you must act…and that should be in a way that increases choices for the greatest number of stakeholders. It may be further discussion or something else, it depends on the specifics of the situation. Regardless, it is a call you must make…and the choice you make says more about you anything else.

Simplicio: That is an unsatisfying answer.

Salviati: I’m sorry, but its as simple as that…and that is what makes it so hard.

–x–

Notes

A metalogue is a real or imaginary conversation whose structure resembles the topic being discussed. This piece is inspired by Gregory Bateson’s metalogues in Part 1 of his book, Steps To an Ecology of Mind.

The characters in this metalogue are borrowed from Galileo’s  Dialogue Concerning The Two Chief World Systems in which the character Salviati is a proponent of the Copernican “heresy” that the Earth is not at the centre of the universe whereas Simplicio favours the Geocentric view proposed by the Greek philosopher Ptolemy.

The worth of an education – a metalogue

“To be prepared against surprise is to be trained. To be prepared for surprise is to be educated.” – James Carse in Finite and Infinite Games.

Prospective student: How much do your students earn after they graduate?

Teacher: I’m sorry, I don’t know.

Prospective student: What do you mean? Don’t you keep track of what students do after they complete the program?

Teacher: Yes, we do, but we don’t ask them how much they earn. It is not the kind of question one asks….

Prospective student:  I know that, but shouldn’t you be gathering hard data on outcomes?

Teacher:  Of course, and we gather a range of data including the roles they get on graduation and how they progress in their careers.

Prospective student:  OK, but that doesn’t help me. How do I know that your degree is worth the hefty fees you charge?  If I’m investing that much, I must be sure of a decent return.

Teacher:  One cannot be certain about anything in life, there are never any guarantees. However, it’s perfectly reasonable to expect value from your education, so let me ask you – what would you consider a decent return?

Prospective student:   Hmm, a good job, I guess.

Teacher: Well, you heard from some of our alumni this evening.  They are currently employed as professionals in the field, and most of them got their jobs while studying or soon after completing their degree.

Prospective student: Yes, but you are unlikely to invite those who would say negative things about the program or those who have failed to get jobs.

Teacher: Fair enough, there are a few of those.

Prospective student: Well, that’s just my point. How do I know it will work for me?

Teacher:  You don’t!  Getting a job is an indirect effect of a good education.

Prospective student: I’m not sure I follow.

Teacher: An analogy might help clarify what I mean by indirect, Bill Gates did not become a multi-billionaire by setting out to become one. He became one by following his interests, his passion.

Prospective student: [a tad irritated] I still don’t get it.

Teacher: The objective of education is not to train you for a vocation…although, if you do things right, you are almost certain to get a job, and a good one at that. The aim of education is personal transformation, to broaden your perspective and thus enable you to look anew at the things you do at work and, possibly, even in life. Another aim is to prepare you to become that buzzword: a lifelong learner. No university can teach you those things, but they can help you learn them.

Prospective student: “personal transformation” sounds wonderful but terribly vague. Could you give me an example?

Teacher: Well, you’ve just heard from a few of our alumni and I could tell you many more stories. The thing is, I don’t think it is helpful to hear second-hand stories – their journeys are theirs, not yours. The story I’m interested in right now is yours: what you’re thinking, what you do and where you want to go.

Prospective student: OK. I’m a financial analyst [Editor’s note: feel free to substitute your current profession here] and I want to be a data scientist. Will this course enable me to become one?

Teacher: It could, but whether or not it actually does depends largely on you.

Prospective student: That’s not an answer.

Teacher: It is, and it’s an honest one. No course will make you a data scientist. And if any university tells you they can, they’re lying. What a good university course will do is help you learn the technical and non-technical skills that will enable you to become a data scientist.  Whether you learn or not depends on you. The responsibility for your personal transformation lies largely with you. All we can do is show you the way.

Prospective student: so, do you cover … [student recites a litany of data science languages and techniques].

Teacher: Yes, we cover them.

Prospective student:  Won’t doing those make me a data scientist?

Teacher: No. If tech skills are all you are after, I’d strongly suggest you don’t join our program…or any other university program for that matter. Instead, head off to one of the good online data science education providers and save yourself a whole lot of money.

Prospective student: Huh?

Teacher:  A good face-to-face program at a university covers a whole lot more than tech. For example, there are certain tacit skills and dispositions that are critical to becoming a good data scientist. These skills have to do with problem finding rather than technical adeptness or problem solving.  A good university course will give you opportunities to gain experience in doing that.

Prospective student; Problem finding? What’s that?

Teacher: In university assignments you’re given readymade problems that you can go off and solve. In real life, however, you are rarely given a problem. More often, you are presented with a situation from which you must extract or formulate a problem before you solve it. That’s not always a straightforward process because every situation is unique in its details.

Prospective student: If every situation is unique then there is no formula to deal with it.

Teacher: Exactly! These skills have to be taught indirectly – by putting students in safe-to-fail situations in which they can learn how to deal with the ambiguity inherent in them

Prospective student: But won’t that be throwing students into a situation they’re unprepared for.

Teacher: Although they may not admit it, most consultants – even experienced ones – rarely feel totally in control when dealing with new clients. It’s good to experience that kind of ambiguity early in one’s career, even if it is a second career. Every consulting engagement is a learning experience. This ties in with what I mentioned earlier – becoming a lifelong learner.

Prospective student: So how do you prepare students to deal with these types of scenarios?

Teacher: Through carefully crafted technical and non-technical subjects, with assignments that make them think rather than just do. To do well in the assignments you will have to think things through, try different approaches and even make judgement calls.

Prospective student: Judgement calls?

Teacher:  Yes, that’s right. You will find that the biggest issues when doing data science in the real world are not technical, rather they are about dealing with ambiguous situations in which you don’t have a well-defined problem or adequate data.  Then there are ethical issues that are becoming ever more important today. There are big corporations that completely ignore the ethical implications of what they do. Just because you can do something, it doesn’t mean you should.  All these issues involve judgement calls in which data is of little or no help.

Prospective student: Hmm, I didn’t realise there were so many facets to being a data scientist. Thanks, you’ve given something to think about.

Teacher: No worries…and good luck, I hope you find what you’re looking for.

 

 

Afterword:

HR gurus and consultants continually pontificate about the future of work. The ground reality for many mid-career professionals is that the future of their work is highly uncertain, much of the uncertainty being fuelled by the perception that data-related technologies are going to “disrupt” established industries. Among other things, this has led to an unprecedented demand for courses that teach data-related skills.

What is often left unsaid, however, is the transition to data science – or any profession for that matter – involves more than just picking up technical skills. The biggest missing piece (in my opinion) is the ability to make sense of ambiguous situations. This is a tacit skill that is difficult, if not impossible, to teach but can be learnt given the right environment and attitude. The university ought to provide the environment, the student the attitude.

Note: A metalogue is a dialogue that unfolds in such a way that the structure of the conversation turns out to be illustrative of the issue being discussed. The anthropologist Gregory Bateson coined the term.  Here is a metalogue written by him.

Heraclitus and Parmenides – a metalogue about organizational change

“Organizations are Heraclitian, but Parmenides is invariably in charge.” –Stafford Beer (paraphrased)

Heraclitus: Hello Parmenides, it’s been a while!  What have you been up to since we last met?

Parmenides: Heraclitus, it is good to see you my old friend. You’re not going to believe it, but I’ve been doing some consulting work on managing change in organizations.

Heraclitus:  [laughs] You’re right, that is beyond belief, particularly in view of your philosophical position on change. So, have you recanted? Have you now come around to the truth that everything changes and nothing stands still?

Parmenides: Ah, yes I am familiar with your views on change my friend, but I hate to disappoint you.  My position remains the same as before:  I still believe that the world is essentially unchanging. The key word here is “essentially” – by which I mean that the changes we see around us are superficial and that the essential properties of the world do not change. Indeed, as paradoxical as it may sound, understanding this unchanging essence enables us to manage superficial changes such as those that happen in organizations.

Heraclitus:  I’m not sure I understand what you mean by unchanging essence and superficial change...

Parmenides:  OK, let me try explaining this using an example. Let us consider the case of a physical law and a real world situation to which it applies. A concrete instance of this would be Newton’s Law of Gravitation and the motion of a spacecraft.  The former represents the unchanging essence while the latter represents one of its manifestations. The point is this:  the real world (as represented by a moving spacecraft) appears to be ever changing, but the underlying unity of the world (as represented by Newton’s law) does not change. If one understands the underlying unchanging laws then one has the power to predict or control the superficial changes.

Heraclitus:  Hmm….I don’t see how it relates to organizations.  Can you give me a more down to earth illustration from your work? For example: what is the “unchanging essence” in organizational change?

Parmenides:  That’s easy: the unchanging essence is the concept of an organization and the principles by which they evolve.  Consultants like me help organizations improve performance by influencing or adjusting certain aspects of their structure and interactions. However, the changes we facilitate do not affect the essence of the entities we work with. Organizations remain organizations, and they evolve according to universal laws despite the changes we wrought within them.

Heraclitus: Ah Parmenides, you are mistaken: concepts and principles evolve in time; they do not remain constant. Perhaps I can convince you of this by another means.  Tell me, when you go into an organization to do your thing, how do you know what to change?

Parmenides:  Well, we carry out a detailed study by talking to key stakeholders and then determine what needs to be done.  There are a host of change models that have come out of painstaking research and practice.  We use these to guide our actions.

Heraclitus: Are these models  akin to the physical laws you mentioned earlier?

Parmenides:  Yes, they are.

Heraclitus: But all such models are tentative; they are always being revised in the light of new knowledge. Theory building in organizational research (or any other area) is an ongoing process. Indeed, even physics, the most exact of sciences, has evolved dramatically over the last two millennia – consider how  our conception of the solar system has changed from Ptolemy to Copernicus. For that matter, even our understanding of gravity is no longer the same as it was in Newton’s time. The “unchanging essence” – as you call it – is but a figment of your imagination.

Parmenides:  I concede that our knowledge of the universe evolves over time. However, the principles that underlie its functioning don’t change.  Indeed, the primary rationale behind all scientific inquiry is to find those eternal principles or truths.

Heraclitus: It is far from clear that the principles are unchanging, even in a so-called exact discipline like physics.  For example, a recent proposal suggests that the laws of physics evolve in time.  This seems even more likely for social systems: the theory and practice of management in the early twentieth century is very different from what it is now, and with good reason too – contemporary organizations are nothing like those of a century ago.  In other words, the “laws” that were valid then (if one can call them that) are different from the ones in operation now.

Parmenides:   You’re seduced by superficial change – you must look beneath surface appearances!  As for the proposal that the laws of physics evolve in time, I must categorically state that it is a minority view that many physicists disagree with  (Editor’s note: see this rebuttal for example)

Heraclitus: I take your point about the laws of physics…but I should mention that history is replete with “minority views” that were later proven to be right.  However, I cannot agree with your argument about superficial change because it is beyond logic. You can always deem any change as being superficial, however deep it may be. So let me try to get my point across in yet another way. You had mentioned that you use management principles and models to guide your actions. Could you tell me a bit more about how this works in practice?

Parmenides:  Sure, let me tell you about an engagement that we recently did for a large organization. The problem they came to us with was that their manufacturing department was simply not delivering what their customers expected.  We did a series of interviews with senior and mid-level managers from the organisation as well as a wide spectrum of staff and customers and found that the problem was a systemic one – it had  more to do with the lack of proper communication channels across the organisation  rather than an issue with a specific department. Based on this we made some recommendations to restructure the organisation according to best practices drawn from organisational theory.  We then helped them implement our recommendations.

Heraclitus: So you determined the change that needed to be made and then implemented the change over a period of time. Is that right?

Parmenides: Well, yes…

Heraclitus: And would I be right in assuming that the change took many months to implement?

Parmenides: Yes, about a year actually…but why does that matter?

Heraclitus:  Bear with me for a minute. Were there any significant surprises along the way? There must have been things that happened that you did not anticipate.

Parmenides: Of course, that goes with the territory; one cannot foresee everything.

Heraclitus: Yet you persisted in implementing the changes you had originally envisioned them.

Parmenides: Naturally! We had determined what needed to be done, so we went ahead and did it. But what are you getting at?

Heraclitus: It’s quite simple really. The answer lies in a paradox formulated by your friend Zeno: you assumed that the organization remains static over the entire period over which you implemented your recommendations.

Parmenides:  I did not say that!

Heraclitus:  You did not say it, but you assumed it.  Your recommendations for restructuring were based on information that was gathered at a particular point in time – a snapshot so to speak. Such an approach completely overlooks the fact that organisations are dynamic entities that change in unforeseen ways that models and theories cannot predict. Indeed, by your own admission, there were significant but unanticipated events and changes that occurred along the way.  Now you might claim that those changes were superficial, but that won’t wash because you did not foresee those changes at the start and therefore could not have known whether they would be superficial or not.

Parmenides:   Well, I’m not sure I agree with your logic my dear Heraclitus. And in any case, my approach has the advantage of being easy to understand. I don’t think decision-makers would trust a consultant who refuses to take action because every little detail about the future cannot be predicted.

Heraclitus: Admitting ignorance about the future is the first step towards doing something about it.

Parmenides: Yes, but you need to have a coherent plan, despite an uncertain future.

Heraclitus: True, but a coherent plan can be incremental…or better, emergent –  where planned actions are adjusted in response to unexpected events that occur as one goes along. Such an approach is better than one based on a snapshot of an organisation at a particular point in time.

Parmenides:  Try selling that approach to a CEO, my friend!

Heraclitus: I know, organizations are ever-changing, but those who run them are intent on maintaining a certain status quo. So they preach change, but do not change the one thing that needs changing the most – themselves.

Parmenides: [shakes his head] Ah, Heraclitus, I do not wish to convert you to my way of thinking, but I should mention that our differences are not of theoretical interest alone:  they spell the difference between being a cashed-up consultant and a penurious philosopher.

Heraclitus: [laughs] At last we have something we can agree on.

Further reading:

Beer, Stafford (1997), “The culpabliss error: A calculus of ethics for a systemic world,” Systems Practice, Vol 10, No. 4. Pp. 365-380. Available online at: http://rd.springer.com/article/10.1007/BF02557886

Note: the quote at the start of this piece is a paraphrasing of the following line from the paper: “Society is Heraclitian; but Parmenides is in charge.”