By Yanis Varoufakis, professor of economics at the University of Athens. Cross posted from his blog
In my previous post I explained that, over the past few weeks, I was frantically putting the finishing touches on a hopelessly academic book that reflects on my personal experiences with neoclassical economics; primarily through the lens of my own work which attempted to engage neoclassical colleagues on labour economics, game theory and political economy. Now that the task is finished, and the draft has been sent to my publishers, perhaps it may amuse some of you to read an extract from the concluding chapter; the only part (saving the Preface, which I already uploaded in my previous post) that is not riddled with equations, diagrams etc. In my attempt to sum up the book’s theme, I posed a strange question: If a sophisticated neoclassical economist ever attempted to write a play without violating neoclassicism’s basic tenets, what would that play be like? If interested, read on…
The Social Theorist as Storyteller
Now that all is said and done, and this book on my “personal encounter” with the discursive power economics builds on its gross theoretical failure is complete, perhaps the reader will allow me to conclude on a light-hearted, yet deadly serious, note.
Social theorists, whether we like it or not, are storytellers. We tell meta-stories the purpose of which is to help us understand a social world that is constantly under construction around us, and within which we are both active contributors and incessant interpreters. Incapable of a genuine Archimedean perspective, from which to judge simultaneously both our world and our account of it, we resort to analysing and re-analysing the sort of stories we tell ourselves about our selves.
This book has focused on one particular species of meta-story: that which dominates economic textbooks and discourse from the high school curriculum all the way to finance ministries and the boards of the too-big-to-fail financial behemoths. Having already subjected to critical scrutiny the neoclassical meta-story, and the Dance of the Meta-axioms which keeps it powerful and alive, I shall now end this book with an odd question: Given that drama is the ultimate, and most instructive, form of story-telling, what type of play or novel could a neoclassicist come up with without violating her neoclassical strictures? What would indeterminacy’s role be on her stage, or in the pages of her novel?
The Neoclassical Economist as Playwright[i]
It would not be remiss to imagine that the neoclassical economist tries to be for social theory what the playwright is to theatre: the creator of the plot, the designer of its every twist and turn, the author of a morality tale. At first, she introduces us to the cast of characters, their whims and preferences, their constraints and social location. Next, she sets the scene by developing the intricate web of interdependent decisions that forms their milieu – the grand, usually competitive, ‘game’ they are engaged in.
To pack dramatic punch, this grand game must feature multiple equilibria. Let me explain why: George Bernard Shaw, in the preface to one of his exquisite plays, wrote: “No conflict, no drama!” Hear, hear! To fashion genuine drama, our neoclassical playwright will undoubtedly feel compelled to weave an authentic conflict into its fabric. For this to be so, the ‘game’ that she places her characters in, must, so as to stay faithful to neoclassicism’s tenets, feature more than one equilibrium outcome. Imagine for a moment that it does not, and that each of her characters possesses a dominant strategy. If so, they will be facing no real dilemmas. Their dominant strategy may well instruct them to slaughter and to pillage, to wreak havoc and turn other people’s lives upside down. But the result will be crude, brutal theatre. Conflict contributes to a decent play only by allowing the audience to identify with even the worst villain. Interesting, mesmerising conflict involves a degree of inner turmoil the prerequisite of which is some form of… indeterminacy.
The neoclassicist’s conundrum here is, naturally, that indeterminacy is her sworn enemy. And yet, in attempting to write her captivating ‘play’, she realises that she cannot dispense with at least some indeterminacy. It is for this reason that multiple equilibria are, as I stated above, sine qua non for neoclassical theatre. The question now becomes: What constraints does a reliance on multiple equilibria impose upon the neoclassical playwright? We know that, by the very nature of neoclassical method, once the multiple equilibria are in place, the playwright’s characters have no alternative other than to resolve them by means of some form of randomisation. For if some other non-random mechanism were invoked, which rationalises in front of an audience the selection of one out of multiple equilibria, the whole point of having created a play with multiple equilibria is defeated.
Indeed, if the play’s leading character can reliably select a path among many competing ones without resorting to randomisation, it must be the case (at least in the neoclassical mind-set) that this path is, by definition, optimal and thus consistent not with multiple but with a unique equilibrium. But if this were so, the play is constitutionally dull, featuring characters whose path is predetermined (by the unique equilibrium available to them) and whose choices are, consequently, unrevealing of anything that may cause a ripple of excitement through the audience, or help them re-assess their own humanity.
Creative neoclassical playwrights may attempt to escape this conundrum by allowing their characters occasionally, and in the heat of the moment, to depart from instrumentally rational choices. In the parlance of game theory, the neoclassical playwright may introduce some ‘trembles’, or noise, in the characters’ behavioural pattern, hoping that this will complicate the plot sufficiently for the audience to find the proceedings absorbing. The question is: How absorbing would such a play really be?
I harbour no doubt that many Hollywood flicks are written in more or less this (neoclassical) fashion. In the tradition of John Wayne and Arnold Schwarzenegger movies, the screenplay is mostly predictable and the characters’ actions predetermined, save perhaps for a little randomisation reflecting the possibility that those in the weaker position err randomly into thinking that they stand a better chance than they really do (before being mowed down by the ‘hero’). However, such examples of ‘drama’ would not satisfy a sophisticated neoclassicist playwright or screenwriter who, outside her professional world, knows the difference between a good play (written for the screen or stage) and an impostor. The reason why offers a useful platform from which to reconsider neoclassical economics in general and its highest form, game theory, in particular.
At first, there is the problem with the use of multiple equilibria as a device for raising the dramatic tempo. How does the playwright resolve them? The equilibrium solution, demanded by neoclassicism’s Third Meta-axiom, is to have a unique mixed strategy equilibrium cutting the Gordian knot of indeterminacy by some optimal randomisation rule. Or that protagonists reach a settlement reflecting a Nash bargaining solution that equalises ratios of their utility functions’ first and second order derivatives. It would be as if Shakespeare decided on whether Macbeth would commit murder by tossing a suitably biased coin, or as if Sophocles had Antigone and Creon resolve their ‘disagreement’ by maximising the product of their utilities, and then writing the play according to the outcome. If our playwright rejects the mixed strategy Nash equilibrium solution (as I did in various chapters, beginning with Chapter 3), then the equilibrium story will either be devoid of tension (as the playwright will fall back on dominant strategies) or it will require external input for its completion (e.g. some Deus Ex Machina that introduces reasons, for the outcome, that are external to the ‘game’ the playwright set up). In view of our playwright’s need for tension but also her penchant, as a committed neoclassicist, for endogenous ‘closure’, she is left with no good option.
One escape route, suggested by the last paragraph, might be to consider the neoclassical method as the play’s grammar or background logic while the playwright’s imagination is permitted to look to some external source of inspiration for the actual course that the characters will trace out within that ‘grammar’ or ‘landscape’. Under this partial retention of equilibrium theory, the playwright is at liberty to deploy mixed strategies, Nash bargaining solutions, general equilibria, conventions that evolve in ways Evolutionary Game Theory can understand (recall the last chapter). However, the secret of the play is not to be found in any of these theoretical constructs. Although a play is possible that relies entirely on them, I for one would not enjoy it much. And I do not think that our neoclassicist playwright would either.
Rational Deviations from Equilibrium as the Prerequisite for Good Theatre
So, the identification of taxing dilemmas with multiple equilibria does not have to lead to boring plays, as long as the playwright first sets up the dramatic structure and then augments the equilibrium path that the dramatic structure generates with exogenously determined events. However, I fear that ushering in resolutions which are independent of that structure (as they would have to be, since the equilibrium foundations can neither depend on the resolution nor spawn a resolution) would not work at all well. This kind of play would require that the dramatic tension be structurally independent of its culmination. For those of us who think that a Deus Ex Machina resolution is detrimental to good theatre, the only decent alternative is to abandon the synonymity between hard choices and multiple equilbria, just as I abandoned the identification of the equilibrium strategy in the centipede game of Chapter 4 with the uniquely rational course of action.
Indeed, the discovery of rational non-equilibrium strategic choices in various chapters of this book (i.e. the possibility of perfectly reasonable patterned bluffs, mutations and deviant acts) introduces us to a different class of dramatic devices from which the playwright can build satisfyingly three-dimensional characters. The very possibility that Antigone can defy her dominant strategy rationally is what gave the famous play its genuine dramatic texture, focusing as it does the audience’s attention on not just Antigone’s dilemma but, also, on its own inconsistent views regarding what is rational, ethical and, in the final analysis, ‘right’. On the edge of their seats, the audience suddenly expect the unexpected at every moment since ‘uniqueness’ no longer guarantees determinism. They anticipate dramatic choices not only when the alternatives are equally inviting for the characters but even when they are not!
This is precisely the essence and beauty of deviant, yet fully rational, strategies; of the strategies that neoclassicism attempts to eradicate by means of its Third Meta-axiom. By undermining the characters’ perception of what rational people do, subversive behaviour may or may not reward those who adopt it with superior outcomes. On this account, Aeschylus raised Prometheus from the obscurity and banality of the Hesiodic tale by empowering him to experiment with a subversive, a deviant, an ‘out-of-equilibrium’ strategy. Of course, other more cynical interpretations are always possible. Prometheus could simply be suffering from imperfect information, in which case he would not have repeated his gift to humanity had he been given a second chance. Alternatively, he could have enjoyed martyrdom and accepted pain as a reasonable price for it. However, both these interpretations cheapen the character created by Aeschylus. No tragedy is worth its salt if it is based on the exploits of a hero who simply miscalculated, or who unexpectedly learnt how to derive masochistic pleasure when things turned differently to his original plan.
The above musings lead gradually to the conclusion that high theatre is impossible without radical indeterminacy, i.e. without neoclassicism’s nemesis which economists habitually, mostly unwittingly, try to put away by means of their Dance of the Meta-axioms. But even if we agree that indeterminacy is a prerequisite for a good play, it is certainly insufficient. Indeterminacy prevailed in Chapters 2 to 5, when the agents’ motivation was taken for granted. However, unless their praxes infect their motivation and influence their beliefs, there is no real theatre to speak of. Chapters 5 to 12 took account of this interdependence which, at once, builds on indeterminacy and takes it onto a higher plane of complexity. That escalation and decent into ever more radical indeterminacy was, at least from a self-respecting playwright’s perspective, apt and timely.
As every schoolchild knows, Macbeth adds crime to crime, as a result of successive choices that he ‘fell’ into when ‘murder’ was one of several options. He then emerges defeated while simultaneously victorious. When he achieves clarity toward the play’s end, he tells us that he wished his ‘beliefs’ were expunged:
Canst thou not minister to a mind diseas’d,
Pluck from the memory a rooted sorrow,
Raze out the written troubles of the brain,
And with some sweet oblivious antidote
Cleanse the stuff’d bosom of that perilous stuff
Which weighs upon the heart?
But at the same time he re-discovers his dignity when forced to choose between death and humiliation:
…Lay on Macduff,
And damn’d be him that first cries, ‘Hold, enough!’
The neoclassical interpretation is that Macbeth became involved in conflict due to imperfect information as to his opponents’ preferences and strengths; that he made his choices rationally given his priors of belief; that his choices had been hard because they belonged to sets of equilibrium strategies containing more than one element; that, perhaps, there was an element of the irrational in his deeds that took the form of some random ‘tremble’; some deviation from the optimal strategy which packs no significance in itself; and so on. All these suppositions are contentious because they wilfully ignore the possibility that a rational Macbeth could have chosen from the non-equilibrium strategy set as well (as I have contended in this book). None, however, is as contentious, and downright absurd, as the assertion that Macbeth’s choices left his personality unaffected, or that they affected him in a predetermined (even if stochastically so) manner. Once choice and action contaminate motivation, the latter cannot determine the former. Indeterminacy is, therefore, irrepressible not just as the stuff of good theatre but also as a prerequisite for human development.
Persons and Roles, Impersonators and Actors
The importance of praxis for deciding what kind of theatre (and social theory) we want becomes apparent when the repercussions of the acceptance of the neoclassical equilibrium story are considered. If tensions and dilemmas are the result of multiple equilibria, their resolution cannot be seen as a contributor to the character of agents. Macbeth’s tragedy is not so much about his fate and final destruction (i.e. it is not about the outcome per se) as about the transformation of his self through praxes – a transformation that the neoclassical story cannot even begin to conceive. Even if we approach the matter from an evolutionary game theory perspective (recall the last chapter), the neoclassical insistence that character variety is spearheaded by statistically uncorrelated ‘mutations’, or replication errors, amounts to the same thing: praxes are not causally linked to the evolution of character.
If we are to make room for this two-way process, from self to action and from action to self (the process that I call praxis), theatre regains its potency and meaning, although the neoclassical project sinks without a trace in a sea of radical indeterminacy. But what kind of theatre do we end up with? One possibility is a relativist open-ended script, in which the playwright lets the actors improvise and bring on the stage perspectives from their own lives. Together with neoclassical narratives concerning the outcome, such ‘libertine’ improvisation brings an end to any type of theory about how the play will end. Suddenly we are no longer in the realm of theatre but in something much closer to a multi-player video game in which the game’s author offers the environment but it is the players that collectively script the outcomes.
The input from the social world, the world of actors, or gamers, who have a life outside the ‘game’, or play, decides the outcome and there is no sense in considering the script as anything more than a basic skeleton which cannot be read independently and across different cultural milieus as transcendental (unlike for instance a play authored by Shakespeare or Sophocles). It is this post-modern relativism that neoclassicists of renown (particularly game theorists such as the formidable John Harsanyi) have sought to keep at bay by excluding from games anything that was not in the players’ utility payoff functions (that is, in the ‘script’). But he may have gone too far, killing off the very possibility of good theatre, of wholesome social theory, or even of a believable model of men and women. Is there an alternative to the choice between bad theatre and multi-player video games?
I think so. The claim buried deeply inside this book is that creativity and interest (in theatre as well as in the social sciences) can be restored without going to one of these extremes. There is, I submit, no need either to succumb to deterministic scripts or to go all the way to the other extreme, yielding to relativist, arbitrary, actor-imported narratives. The theory of deviant-cum-rational behaviour that unfolded through various chapters of this book opens up the possibility of having a complete, riveting script without determinism. The key is the admittance to the plot of twists that defy unique equilibria and in so doing capture the imagination. When one is faced with multiple equilibria, any odd choice will do; by observing the outcome you really learn nothing about the chooser’s character. It is only when the character reasonably and purposely deviates from some unique equilibrium path (akin to Prometheus’ theft of fire on behalf of humanity) that the writer is telling us something momentous about both the character and the human condition.
The readiness to infuse rationality with irrationality in the defence of deviance and rebelliousness brings to mind another dimension. In a ‘neoclassical’ theatre production the actor sheds her own personality, steps into the predetermined role, and struggles to impersonate some character whose entire presence flows from the script. The playwright chooses how to weave the plot and the actor attempts to be the person who actually does the weaving. However, in high theatre, rather than impersonating characters, actors personify them.
My preference for scripts which depict and rationalise non-equilibrium choices is well suited to complementing this distinction; to actors that personify rather than merely impersonating. This way, actors can bring on stage their own interpretation of the inner conflict that characterises the choice of some deviant strategy with no need for an open-ended, relativist, multi-player video game-like, script. Since the coexistence of Reason and Unreason in the same behavioural pattern (recall Chapter 6) is inherently confusing, the actor can draw from her own past whatever is necessary to convey the intensity of such choices. Neoclassical theory cannot (because it will not) account for that intensity, thus devaluing, quite inadvertently, the actor’s contribution to the play.
Does any of this matter? I think it does, and that Macbeth offers a good example. For he demonstrates that, often, it is obligatory that in order to understand conflict in the social world around us, we must look into a conflicted heart overseen by a bright mind. If we are to do so without abandoning the realm of rational analysis, the only option is to enrich our perception of it. Rejecting neoclassical equilibrium analysis is an excellent start along this path.
[i] My idea of imagining how a neoclassicist playwright would go about her business of writing her play first occurred to me when I was writing my 1991 book Rational Conflict. See pp. 196-200).