Thucydides as a prospect theorist. 2014

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232 brill.com/polis

Thucydides as a Prospect Theorist Josiah Ober

Departments of Political Science and Classics, Stanford University, 100 West Encina Hall, 616 Serra St., Stanford, CA 95405, USA [email protected]

Tomer J. Perry

Department of Political Science, Stanford University, 100 West Encina Hall, 616 Serra St., Stanford, CA 95405, USA [email protected]

Abstract Opposing the tendency to read Thucydides as a strong realist, committed to a theory of behaviour that assumes rationality as expected utility maximization, Ned Lebow and Clifford Orwin (among others) emphasize Thucydides’ attentiveness to deviations from rationality by individuals and states. This paper argues that Thucydides grasped the principles underlying contemporary prospect theory, which explains why people over-weight small probabilities and under-weight near certain ones. Thucydides offers salient examples of excessive risk-aversion in the face of probable gains and excessive risk-seeking by decision-makers faced with high probability losses. Thucydides suggests that in a democracy, leaders’ rhetoric can limit or exacerbate the political effects of bias in risk assessment.

Keywords Thucydides – prospect theory – Melian dialogue – rationality – leadership

In The Humanity of Thucydides, Clifford Orwin proposes that the greatest advantage Thucydides’ ancient readers had over us, as moderns, was that the ancients ‘lived and breathed politics while remaining innocent of “political

© koninklijke brill nv, leiden, ���4 | doi 10.1163/20512996-12340015

Thucydides as a Prospect Theorist

207

science” ’.1 By contrast, we have argued in earlier work that Thucydides was in some ways so precociously modern that he may be said to have invented political science.2 Here we double down on that earlier bet, suggesting, first, that Thucydides was also the inventor (in the sense that he anticipated certain key insights) of an important branch of behavioural economics, and, next, that by reading him as such we may understand him better. Yet there is considerable common ground between Orwin’s Thucydidean humanism and our own understanding of what Thucydides has to offer political theorists today. 1 Introduction A central argument in Orwin’s book is that political thought, as it was practiced by Thucydides and as it may legitimately be practiced by his successors, must seek to understand human nature – to anthropinon, ‘the human thing’. Thucydides is concerned, as we ought to be, with the question of how individuals and states can be expected to behave, both in ordinary and in extraordinary circumstances. Their behaviour is conditioned by nature, in ways that may be modulated by specific political regimes. But some facts about human nature ultimately transcend all regimes.3 We suggest that Thucydides sought to understand the effect of human nature on history by reference to what we would now call behavioural economics. This means that the objects of his study (ancient humans) are fundamentally similar to the objects studied by modern psychologists (modern humans). And so, while it may be modulated by specificities of historical eras (ancient or modern), human nature ultimately transcends time. Thucydides grasped a basic psychological fact that is at the centre of contemporary studies of behavioural economics: although we are capable of employing complex forms of reasoning, we do not typically do so, especially when making choices involving probability and risk. Instead, we tend to misestimate probabilities in predictable ways since emotions like fear and hope make certain scenarios more salient and vivid than others. 1 C. Orwin. The Humanity of Thucydides (Princeton: N.J.: Princeton University Press, 1994), p. 12. 2 J. Ober, ‘Thucydides and the Invention of Political Science’, in A. Rengakos and A. Tsakmakis (eds.), Brill’s Companion to Thucydides (Leiden: E.J. Brill, 2006). 3 Orwin, The Humanity of Thucydides, pp. 10-11.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

208

Ober and Perry

Thucydides recognized that the human mind is wired in certain ways with respect to rationality, desire, judgment, and decision. He understood humans as rational in some important particulars – that is, as beings that are prone to seek their own advantage and, sometimes but not always, their community’s advantage.4 They do so by strategic calculations that take into account the other fairly rational individual and collective actors and their likely preferences over outcomes. In Thucydides, the relevant actors are men and city-states. Yet Thucydides also recognized that there are limits to human rationality, when rationality is understood (as it is for the purposes of this article) as accurate calculation guided by self-interest, and as behaviour that follows logically from that calculation. Human rationality is limited because (among other causes) the emotions of fear and hope affect how people assess and respond to risk and opportunity. Fear and hope are central motivators in Thucydides’ work, as many commentators pointed out.5 We seek to show that Thucydides paid close attention to systematic deviations from rationality occasioned, variously, by fear of loss and fear of disappointment over unrealized gains, and by hope of gain and hope of avoiding loss. Orwin explores both rationality and its limits in Thucydides. The tension between a dream of rational policy-making that would consistently take proper account of risk and opportunity, and the limits of rationality as an instrument of real-world policy-making, are important to Orwin’s discussion of what he calls Thucydides’ ‘Athenian thesis’: the argument that individual and state interests take priority over considerations of justice. The Athenian thesis is elaborated, in conflicted ways, by the speeches of Athenian envoys at Sparta in Book 1 and at Melos in book 5.6 Among the problems inherent in the Athenian thesis is that the effective pursuit of interest demands strict rationality. In practice, individual and collective agents often fail to identify the course of action that would best serve their interests. In particular, chance challenges humans’ ability to correctly evaluate alternatives: players, whether cities or people, tend to make systemic misestimations of the likelihood of possible scenarios and 4 Orwin, The Humanity of Thucydides, chapter 1. 5 On the importance of the emotion of fear in Thucydides, see W. Desmond, ‘Lessons of Fear: A Reading of Thucydides’, Classical Philology, 101 (2006), pp. 359-79. See also R.D. Luginbill, Thucydides on War and National Character (Boulder, Colorado: Westview Press, 1999), p. 65, who says, ‘as the two basic psychological states affecting historical activity, hope and fear are omnipresent’. 6 All English references to Thucydides in this article refer to The Landmark Thucydides: A Comprehensive Guide to the Peloponnesian War, ed. Robert B. Strassler (New York: Free Press, 1996).

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

Thucydides as a Prospect Theorist

209

therefore fail the attempt to promote their own interests. In Orwin’s words: ‘cities are called on to act not on certainties but on probabilities. This very fact fosters the human proclivity for entrapment by improbabilities’.7 We build on Orwin’s argument to suggest that one of Thucydides’ lessons is that, under the impetus of fear and hope, individuals and groups often misestimate their chances of realizing gains or avoiding losses, and they make poorer choices as a result of those mistakes. The results of misestimation, in the high-stakes environment of a great war, often proved catastrophic. We argue that we can better appreciate some details of Thucydides’ understanding of human rationality, and his understanding of reason’s limits in the task of assessing risk and opportunity, by approaching several key passages via prospect theory, developed by the psychologists Daniel Kahneman and Amos Tversky.8 Treating Thucydides as a ‘prospect theorist avant la lettre’ allows us as moderns to leverage the results of contemporary theoretical and experimental social science in order to grasp a bit more of what Thucydides expected his most astute readers – those who are addressed in the famous ‘possession for all time’ passage (1.22.4) – to take away from his text. We assume that Thucydides intended his narrative to instruct future political agents, and that among those agents are statesmen for whom Pericles will be a model – albeit a complex one. Each of the passages we discuss – the narratives of the Melian Dialogue and Sicilian Expedition Debate in books 5 and 6, the ‘Corinthian assessment’ of risk-seeking Athens and risk-averse Sparta in book 1.68-71, and the analysis of Pericles’ leadership and rhetoric at 2.65 – are marked by their striking form and content. We assume, without arguing the point here, that some part of the understanding of human nature that Thucydides hoped his reader would take away from his text, and the justification for his belief that leaders might intervene beneficently in human affairs, are to be found in these well-known parts of the history. 2

Prospect Theory: The Four-fold Pattern

The idea that Thucydides might fruitfully be read as a prospect theorist is not original to us. Well over a decade ago, Ned Lebow reviewed humanistic work on

7 Orwin, The Humanity of Thucydides, p. 202. 8 Summed up, and refined in D. Kahneman, Thinking, Fast and Slow (New York: Farrar, Straus and Giroux, 2011).

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

210

Ober and Perry

Thucydides by political theorists and classicists.9 In an aside, Lebow claimed that, ‘Thucydides had an intuitive grasp of prospect theory (Levy 1992, 1996; Tversky and Kahneman 1992), which is based on the robust psychological finding that people are generally more concerned with preventing loss than they are with making gains’.10 Lebow did not pursue this insight, but we suggest that looking more carefully through the prism of prospect theory at Thucydides’ accounts of how decision-makers assess chances of loss and gain under conditions of extremity may help us see more of what Thucydides himself meant for careful readers to take away from his history. Before proceeding, it is important to note that even though loss aversion is a central idea of prospect theory, the insight that people tend to be more concerned with loss than they are with gain is not original to prospect theory and has long been recognized as a common sense notion. Kahneman notes that his grandmother knew that losses loom larger than gains.11 Bernoulli’s model of diminishing marginal utility, the centrepiece of classical Rational Choice Theory, holds that people value their first dollar more than their 100th dollar, which predicts that at any given wealth level a person would value the money he already has more than any potential gains. What is missing in Bernoulli’s model, according to prospect theory, is the idea of a reference point without which he cannot explain why people who face terrible options, like the Melians in the Melian Dialogue (see below) tend to behave in a risk-loving manner, by taking gambles and rejecting favourable settlements. In Thucydides we find not only the common sense understanding of loss aversion but also a more nuanced understanding of the way circumstances affect people’s attitudes towards risk. In their work on heuristics and prospective judgments, Kahneman and Tversky sought to debunk the strong anti-humanistic form of Rational Choice Theory – that is the version of the theory that says that people actually are fully rational in just the way that contemporary economists say they ought to be. In order to be fully rational, to be an Econ rather than a Human (in the 9

R.N. Lebow, ‘Thucydides the Constructivist’, American Political Science Review, 95 (2001), pp. 547-60. 10 Lebow, ‘Thucydides the Constructivist’, p. 557; J.S. Levy, ‘An Introduction to Prospect Theory’, Political Psychology, 13 (1992): pp. 171-86; J.S. Levy, ‘Loss Aversion, Framing and Bargaining: The Implications of Prospect Theory for International Conflict’, International Political Science Review, 17 (1996), pp. 179-95; A. Tversky and D. Kahneman, ‘Advances in Prospect Theory: Cumulative Representation of Uncertainty’, Journal of Risk and Uncertainty, 5 (1992), pp. 297-323. 11 Kahnemam, Thinking, Fast and Slow, p. 300.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

Thucydides as a Prospect Theorist

211

terms that Kahneman borrows from Richard Thaler), one must consistently choose the course of action that will maximize one’s expected utility.12 Utility symbolizes the psychological benefit of achieving desired results, formalized as a utility function that represents our preferences over possible outcomes. As such, utility may be gained from anything that someone happens to value in a way that can be compared. More precisely, if Jane gains more utility from X that she does from Y, it means she prefers X to Y. Rational calculation of Econs is based on an accurate understanding of statistical probability. Kahneman and Tversky showed that Humans (that is, actual people), unlike Econs (the hypothetical choice-makers of economic theory), commonly make mistakes (that is, their behaviour does not follow from expected utility maximization) when assessing probabilities. Moreover, some kinds of mistakes are quite predictable, based on how the human mind ordinarily functions. The upshot is that Humans make their judgments (and act accordingly) based on ‘prospective utility’ rather than on ‘expected utility’. As a predictable result, when Humans consider uncertain prospects, they stray from the course of strict rationality, typically over-weighting small probabilities and under-weighting near certain ones. Over-weighting possible yet unlikely losses, resulting in risk-aversion, is a key part of prospect theory. Yet under some circumstances, agents will overestimate the chance of achieving very low-probability gains (as, for example when they buy lottery tickets), or of avoiding high probability losses, in either case resulting in risk-seeking behaviour. Indeed, one of the main innovations of prospect theory was to point out that risk-aversion and risk-loving are not, contra traditional economic theory, merely preferences. Instead, the theory predicts risk aversion in certain circumstances and risk loving in others. The point is then that all of us, regardless of preferences, will tend to be risk averse when considering unlikely losses but risk loving in the face of really bad choices.13 The experimental evidence on human behaviour regarding losses and gains, in the face of high- and low-probability prospects, produce what Kahneman 12 Kahnemam, Thinking, Fast and Slow, pp. 269-70; R.H. Thaler and C.R. Sunstein, Nudge: Improving Decisions about Health, Wealth, and Happiness (New Haven: Yale University Press, 2008). 13 ‘Really bad’ choices are choices that yield outcomes far below the reference point. Per below, the reference point is the baseline according to which gains and losses are defined; it is often (though not always) the status quo. This obviously leaves as an open question what it means to be ‘far’ below the reference point, but the important thing to note is that prospect theory views choice as reference-dependent, and that this is an innovative departure from classical economic theory.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

212 Table 1

Ober and Perry Emotions, behaviour, and attitudes under uncertain prospects: the four-fold pattern. After Kahneman 2011 317. 1 = prospect. 2 = focal emotion evoked by prospect. 3 = how most people behave. 4 = expected attitude. Examples from Thucydides in italics Gains

Losses

1. Likelihood of big gain High probability 2. Fear of disappointment ‘Certainty effect’ 3. Risk averse 4. Accept unfavourable settlement (over-insure) Sicilian Expedition Debate

1. Likelihood of big loss 2. Hope to avoid loss 3. Risk seeking 4. Reject favourable settlement

1. Slight chance of big gain Low probability 2. Hope of gain ‘Possibility effect’ 3. Risk seeking 4. Reject favourable settlement

1. 2. 3. 4.

Athenians (per Corinthians)

Melian Dialogue Slight chance of big loss Fear of loss Risk averse Accept unfavourable settlement (over-insure) Spartans (per Corinthians)

and Tversky called ‘the four-fold pattern’;14 this is represented in a 2×2 matrix, adapted here as Table 1. In light of foreseeable behavioural consequences of weighting errors, anyone who, ex ante, assumes that others behave like Econs (by developing a strategic plan based strictly on expectation), is likely to find, ex post, that her prior beliefs about others’ choices, and thus about outcomes, have been proved false. Making plans based on an assumption of full rationality is, therefore, likely to be counterproductive. This, we argue, is a fundamental part of the rationality and probability problem that Thucydides explores in what Orwin calls ‘the Athenian thesis’. Beyond the four-fold pattern, Kahneman and Tversky also found that decisions are greatly influenced by the ways choices are framed: people react differently to seemingly identical choices when the framing is changed. One way to manipulate the framing of a choice is to move one’s reference point, thus making gains into losses or vice versa. Unlike classical rational choice theory, prospect theory views choice relative to a reference point. When people consider choices, they compare possible outcomes to the reference point: 14 Kahneman, Thinking, Fast and Slow, pp. 310-21.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

Thucydides as a Prospect Theorist

213

outcomes that are better than that reference point are counted as gains while the others are losses. Often this baseline is simply the status quo, but it can be some other thing, such as goals we set for ourselves. This notion of framing will be important to our discussion of the Corinthian assessment, below. As Lebow argued, Thucydides has too often been read by political scientists specializing in International Relations as a straight-forward, rationalchoice ‘strong Realist’ – that is, in the terms adopted here, as a scholar who takes people to behave like Econs.15 Instead, we suppose that Thucydides was a highly sophisticated thinker who could reason both as an Econ and as a Human, and who, in his history, assessed the behaviour of Humans against the ideal of Econs’ behaviour. To attend, with Orwin, to ‘the humanity of Thucydides’ is not to suggest that Thucydides was an irrationalist who was unable to think as an Econ. One point of Orwin’s analysis of the Athenian thesis is to show us why and how (in the terms adopted here) ‘assuming that people will reason more like Econs and less like Humans’ will get leaders and states into serious trouble. But in order to develop his critical approach to reason and choice, Thucydides necessarily had to be able to understand rationality as it is (and was) understood by Econs. Prospect theory developed as a reaction to the prominence of the rational choice model, and so its terms would only make sense to Thucydides if he too had an appreciation of rational behaviour. This is why Thucydides’ credentials as a realist are crucial: they are evidence that he recognized the importance of accurate calculation in decision-making. Indeed, optimal policy is, for Thucydides, rational policy insofar as probabilities ought to be rightly weighted. Thucydides supposed that a measure of rationality in policy-making was essential for the success of states. Yet Thucydides appreciated the complexity of such a task and was attentive to the fact that people tend to get this wrong. The error he teaches his readers to avoid is not rationality, but confusing optimal rationality with the actual choices likely to be made by real-world Human decision-makers. As Orwin puts it, although in a different context, ‘For reasons that are wholly rationalistic, Thucydides rejects “rationalism in politics”.’16 Thucydides teaches his reader the difference between rationality and irrationality. He carefully recorded specific deviations from rationality and pointed 15

Lebow, ‘Thucydides the Constructivist’; Cf. J. Ober, ‘Thucydides Theoretikos/Thucydides Histor: Realist Theory and the Challenge of History’ in D.R. McCann, B.S. Strauss (eds.), Democracy and War: A Comparative Study of the Korean War and the Peloponnesian War (Armonk, N.Y. and London: M.E. Sharpe, 2011), pp. 273-306. 16 Orwin, The Humanity of Thucydides, p. 204.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

214

Ober and Perry

to their costs. He also understood, and shows his readers how better to understand, the roots of specific tendencies to deviate from rationality in extreme circumstances. Those roots lie in the way the human mind works – i.e. in human nature. Thucydides instructs his attentive reader about what an effective leader will need to know in order to devise rational policy. But he also offers his reader answers to another set of questions: why and when can people be expected to deviate from rationality? How might they be led, through rhetorical persuasion, to choose a more rational course? How can a community avoid the pitfalls of underestimating risk, while also avoiding the costs of overinsuring against risk? The upshot is that under the right kind of leadership, a community might come to do better, over time, by taking on a reasonable level of risk, and making reasonable gambles on appropriate opportunities. That kind of leadership is a scarce resource in the historical past narrated by Thucydides. We believe that part of his ambition as a writer was to make effective leadership a more abundant resource for the future. 3

Melos: Certainty Effect and Loss

The Melian narrative (including the Melian Dialogue) in 5.84-115 provides Thucydides with a case study in risk, hope, and prospective choices that deviate from rationality with catastrophic consequences. The situation is this: in 416 BCE, after a long hiatus in the Peloponnesian War, the Athenians decided to add the small Aegean island-state of Melos, one of the few Aegean states to have retained its independence, to their empire. Athens dispatched an expeditionary force to take Melos. Before commencing hostilities, the Athenian commanders sought to address an assembly of Melians, in which they would have urged surrender. Melos’ oligarchic rulers denied this request. The Athenian commanders and the Melian oligarchs then have an in camera discussion, which Thucydides (in a striking departure from his normal narrative conventions) offers in the form of a dialogue. The Athenians present a rational argument, centred on Athens’ overwhelming military superiority, in favour of Melian surrender. The Melians refuse to accept that argument, choose to resist, and are subsequently destroyed. Orwin focuses on Thucydides’ treatment of the related themes of risk, fear, and hope in the affair of Melos.17 He notes, rightly, that if we, Thucydides’ readers, are to take away the right lesson from the Melian affair we need to attend to the Melians’ mistake in acting on the basis of their hopes of salvation, by gods 17 Orwin, The Humanity of Thucydides, Chapter 5.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

Thucydides as a Prospect Theorist

215

or Spartans. Yet, as Orwin also emphasizes, we must also attend to the error made by the Athenian envoys in failing to anticipate that their rationalistic arguments will fail with the Melian oligarchs who hold decision authority.18 The Athenians impute too much rationality to the Melians. The Athenian envoys, expecting the Melians to behave like Econs, apparently believed that, once the situation was made clear to them, the Melians would choose to act as did people in early Greek coastal communities. In the ‘Archaeology’ Thucydides notes that the coast-dwellers rationally submitted to the first Greek imperial power, the Cretan King Minos, and, in the long run, they benefitted materially as a result (Thucydides 1.8.3: ‘love of gain would reconcile the weaker to the dominion of the stronger’). What is the aetiology of the Melian failure to reason in this long-term cost-benefit manner? Orwin points to the emotion of shame: the Melians fear the disgrace that submission without fighting would entail and so they pin their hopes of salvation on the gods and the Spartans. It is this emotional commitment to fear and hope that the Athenian envoys fail to shake with their rationalistic argument. In light of Kahneman and Tversky’s prospect theory, we emphasize a somewhat different, but complementary, aspect of the situation: in the face of the Athenian threat, the Melian oligarchs make a gamble that is ill-advised from the perspective of rationality as expected utility maximization. In so doing, they act just as other (modern) humans predictably do when they are driven by the hope of avoiding the high probability of a large loss. They imagine the condition of salvation too vividly; they hope against hope (‘a strange but alltoo-human turn of thought’)19 that they will be able to retain their status quo. In this case the status quo is Melos’ standing as an independent state and the oligarchs’ own position of authority in that state. As the Athenian envoys clearly (and, as it turns out, rightly), point out to them, there is very little chance that the Melians can keep the status quo in the face of greatly superior Athenian power and the firm Athenian determination to incorporate Melos into the Athenian empire. It is striking that the Melians do not dispute this fact: they openly acknowledge that they have slim chances of victory in the case of military confrontation. As their representatives say: ‘you may be sure that we are as well aware as you of the difficulty of contending against your power and fortune, unless the terms be equal’ (5.104). The dispute between the Melians and the Athenians, according to Thucydides’ narrative, was not due to disagreement about the facts; the odds of success were common knowledge in this case. The Athenians also make it clear to the Melians 18 Orwin, The Humanity of Thucydides, pp. 110, 113-7. 19 Orwin, The Humanity of Thucydides, p. 116.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

216

Ober and Perry

that incorporation into the empire will not mean the loss of everything: if the Melians submit without a fight they will be peacefully brought into what amounts to an Athenian co-prosperity sphere. The current ruling oligarchs may lose their monopoly on political power. Taxes will certainly increase, since Melos will pay tribute. And of course Melos will lose the coveted status of independent state. But most Melians will keep their lives and much of their property. For the Melian decision-makers, faced with an uncertain prospect, the situation comes down to a choice between the expected utilities assigned to different possible outcomes, based on utility and probability. The rational (expected utility) choice will be the one with the greatest expectation, once probability has been properly factored in. The greater the probability of an outcome, the greater weight it should have – this is the ‘expectation principle’ of rational choice theory. The Melians must weigh the tiny chance of keeping everything (the status quo, which is their reference point) against the certainty of keeping a fraction of what they currently have (which is counted as a loss relative to the reference point). Whatever probabilities one might assign to the two possibilities, the general situation is clear enough: there is a minute chance of keeping the status quo but it is certain (presuming, as we think we should, that the Athenians’ claim that submission will not entail destruction is credible) that the Melians can keep a sizable part of what they have if they submit. Assume, for the sake of the argument, that the Melians evaluate what they now have (the status quo) as 100 on their utility scale. If they choose to fight there is a 5% chance they can keep all they now have (100) and 95% chance of losing everything (ending up with 0). If they submit, they have a 100% chance of keeping their life, which they value at 15, but not their resources or independence.20 Under these conditions, it is clear that the Melians’ best choice, based on probability and weighting, is to submit. Indeed, any offer that 20

The numbers used in this example are arbitrary, but their meaning is relative: to say that the Melians valued life without their resources and independence at 15 is to say that they value it at around a sixth of their valuation of the status quo which we said they evaluated at 100. It is very difficult to estimate relative valuation of states of affairs, but we assume for the sake of argument that such a relative valuation is possible: That the Melians could not only say that they prefer keeping their independence but also by how much. This is important because the explanation provided by prospect theory only works if there’s some probability at which it would be worth taking the chance of fighting. If the Melians valued their life without independence at substantially more than 15 (relative to their status quo), it would not matter that they had misestimated their probability of success. We later provide some reasons to think the Melians’ great regard for their independence is not totally out of the ordinary.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

Thucydides as a Prospect Theorist

217

will ensure them the value of 5 or greater should be sufficient to tilt the balance in favour of submitting. Yet in fact the Melians choose to fight; predictably they lose and are destroyed. The irrational choice of the Melian rulers seems to take the Athenian envoys by surprise, but it would not be unexpected to anyone familiar with what Kahneman and Tversky call the ‘certainty effect’.21 Kahneman and Tversky showed that most people are risk seekers when they are faced with a prospective choice between a small chance to avoid a great loss and the certainty of retaining a smallish part of what they have (that is, a large loss that is nonetheless not as bad as the loss they might avoid through chance). Offered the choice between a tiny chance to keep everything (avoiding all loss) and the certainty of high costs (losing a lot but not all), most people will irrationally chose to gamble on the tiny chance of avoiding the great loss. As Kahneman notes: ‘people who face very bad options take desperate gambles, accepting a high probability of making things worse in exchange for a small hope of avoiding a large loss. Risk taking of this kind often turns manageable failures into disasters. The thought of accepting the large sure loss is too painful, and the hope of complete relief too enticing, to make the sensible decision that it is time to cut one’s losses’.22 This somber assessment exactly describes the Melians’ situation, their decision, and its consequences. The Athenian speakers in the Melian Dialogue are exasperated because they believe, rightly, that the better choice for the Melians is obvious on the face of it. Moreover, the Athenians rightly believe that they are being more generous than they need to be (5.111.4). They could simply take everything by seizing the island by force. Or they could, based on the hypothetical probabilities above, offer the Melians results which they would value less than 15, yet still be presenting them with an offer that a rational actor would take. This is the substantive (as opposed to merely procedural) reason that the Melian Dialogue can be carried on under the banner of epieikeia – equitability or fairness: the Athenians are in fact making an offer that is more reasonable, more equitable, overall better than what they (as representatives of Athens) could demand.23 Yet the Melians stubbornly refuse, acting entirely in accord with ordinary human nature, turning, in Kahneman’s words (cited above), ‘a manageable failure into disaster’. The Athenians’ exasperation raises an interesting conundrum. Why did the Athenians make such a generous offer, one might ask, if they were assuming 21 Kahneman, Thinking, Fast and Slow, pp. 310-21. 22 Kahneman, Thinking, Fast and Slow, pp. 318-9. 23 Cf. Orwin, The Humanity of Thucydides, pp. 97-8.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

218

Ober and Perry

the Melians would behave like Econs? As we saw, Econs would accept anything over the expected value of the gamble that, in our example, was 5 on the Melians’ utility scale. Why did the Athenians offer the equivalent of 15 instead of, say, 6? Why did they offer to keep Melos intact if it surrendered when anything short of total annihilation would be an improvement to the results of war? The answer to this is that though the Athenians envoys were arguing in the terms of Econs, they actually were, unsurprisingly, behaving as Humans: they too were subjects to the certainty effect, except they were facing a gain, not a loss. In terms of Table 1, they were in the top left corner: faced with a high probability of success, they over-insured. The logic of this phenomenon will be elaborated in more detail in the next section, but we can note for now that the Athenians were predictably risk-averse (against their alleged ‘national character’, which we discuss below in the section on the Corinthian assessment). The interesting confrontation between the risk-averse probable winner and the risk-seeking likely loser is not uncommon. Kahneman cites the work of legal scholars who noted this pattern in contemporary courts, particularly in civil suits involving negotiations of a settlement that is done with uncertainty regarding the result of a trial.24 A defendant with a weak case is likely to lose everything but is nonetheless willing to take the risk. In contrast, the riskaverse plaintiff is eager to ‘lock-in’ a sure gain by accepting a settlement that is ‘less than the statistically expected outcome of the trial’. Like plaintiffs who have a strong case, the Athenians make what they think is a generous offer to settle and avoid the battle, and are shocked when their offer is rejected. The Athenians and Melians fail to reach an agreement not because the Melians were not rational, as neither side was. It is because the Athenians, like many modern plaintiffs, failed to understand how desperate the Melians were and therefore failed to make an offer that was generous enough. It is hard to say how generous the offer should have been for the Melians to accept it (whether or not the Athenians were willing to make such an offer is different matter). Using the estimates provided by Kahneman and the numbers provided above, we would expect that the Melians would weigh the risk of losing the war, which is actually 95%, at 79.3%.25 With these numbers, it is clear why they rejected the Athenian offer which they evaluated at 15: they deemed the prospective value of war at 20.7 instead of its expected value, which is 5. Those numbers are, of course, only suggestive: they merely intend to show that given the extent to which the Melians over-estimated their chances of success, there were offers 24 Kahneman, Thinking, Fast and Slow, p. 320. 25 Kahneman, Thinking, Fast and Slow, p. 315.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

Thucydides as a Prospect Theorist

219

they would have accepted and others that would be rejected. Considering an offer that is only slightly better than the worst possible outcome, desperate people will take their chances, terrible as they may be. Losing their independence was sufficiently awful to the Melians that they rejected the favourable settlement that would have saved their lives. Though it is hard to know without further empirical research, we posit that such a preference is not as outrageous as generations of discussions on the Melian dialogue have suggested; today as well as in ancient times, many people, if not most, would take the dangerous risk of disastrous war when the alternative is loss of independence.26 The point we should take away from Kahneman and Tversky’s discussion of choices made in the face of high probability of large losses, and from Thucydides’ Melian narrative, is that human choice-makers, when faced with very bad options, are not only irrational, they are predictably irrational: anyone aware of this aspect of human nature – which is to say, in the vocabulary of prospect theory, anyone who is aware of the certainty effect – should anticipate a failure to employ the expectation principle when people are confronted with dire prospects. Having paid proper attention to the arguments on both sides of the Melian Dialogue, Thucydides’ careful reader will anticipate an eminently Human failure of reason; the Athenian envoys, making their arguments as if they were facing Econs, did not. In the Econ’s expected-utility world the Athenian envoys are making an offer that a rational actor should accept; they are even more generous than they need to be. But in the world of Humans, they are foolish in imagining that the Melians will think like Econs. Had the Athenian envoys themselves had a more sophisticated idea of the difference between expectations and prospects, they would have taken the likelihood of an irrational Melian choice to resist into account. Kahneman and Tversky demonstrate what Thucydides certainly knew: that the Melians made a perfectly normal (if tragic) human-naturebased misestimate in weighting – the misestimate is foreseeable (although 26

To test our intuition on this matter and with the help of Tomer Ullman of the MIT Computational Cognitive Science Lab we ran a short survey on Amazon’s Mechanical Turk. We recruited 100 participants (36 male, mean age was 34.5) and paid them for their participation. Participants chose between fighting and submitting to a strong invader, in a case where their military advisors say that ‘there is small chance we can win but it is very unlikely’. Submission meant heavy taxes and loss of independence, and losing the fight meant the conquerors would ‘kill every man and sell the women and children as slaves’. Even with these grave consequences, about half of the respondents decided to fight. One respondent’s comment explicitly reflected the part wishful thinking played in his/her decision to fight in order to avoid the terrible result of losing one’s independence, saying: ‘I would want to think I stood a chance at fighting for freedom.’

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

220

Ober and Perry

not certain) and Thucydides’ attentive reader will take that into account if and when she or he is faced with relevantly similar circumstances. 4

Sicilian Expedition Debate: Certainty Effect and Gain

The hypothesis that Thucydides understood the role played by the certainty effect on choices made in the face of high probability events is confirmed by a consideration of the narrative of the Athenian debate on the Sicilian expedition in book 6.8-26. Thucydides pointedly remarks (6.1.1) that this debate was held in the same winter as the annihilation of the Melians. It is a commonplace of Thucydidean scholarship that the Melian Dialogue and the Sicilian Expedition Debate are presented in the history as two sides of a coin – although there is much debate about how to understand the coin itself.27 We suggest that the coin is the certainty effect, and that its two sides are distinguished by the likelihood of great loss and great gain. As in the Melian Dialogue, the issue in the Sicilian Expedition Debate is weighting a highly probable outcome against a certain cost. But, whereas the Melian situation focused on probable loss (Table 1, upper-right box), the Sicilian Expedition Debate centres on the uncertain but probable prospect of a large gain (Table 1, upper-left box). As we have seen, in the face of losses the certainty effect produces irrational risk-seeking. Faced with a high probability of a large gain, however, most people will be irrationally risk-averse: they will tend to over-weight the disappointment they will feel if, contrary to expectations, the gain is not realized. On the basis of experimental evidence, Kahneman and Tversky show that most people, when faced with (for example) a 95% chance to win a large sum of money, say $10,000, will accept the ‘unfavourable settlement’ of a sure-thing payout of less than $9500. Another way of putting this is that most people will choose to pay more than they rationally ought to (in this case, $500) to insure against the small chance of failing to realize the gain. The Athenians were over-insuring when they made a generous proposal to the Melians, and they would do so again in the debate regarding the Sicilian expedition. In the Sicilian Expedition Debate, Nicias, a prominent Athenian leader, tries to persuade the Athenians, voting in a democratic assembly, not to send a military expedition to Sicily. A rival leader, Alcibiades, advocates for the 27 Orwin, The Humanity of Thucydides, pp. 111, 118; Desmond, ‘Lessons of Fear’, p. 366. L. Edmunds, Chance and Intelligence in Thucydides (Cambridge, Mass: Harvard University Press, 1975), p. 186.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

Thucydides as a Prospect Theorist

221

expedition. The background to the Sicilian Expedition Debate is as follows: the Athenians had already decided to authorize a moderate-sized military expedition to Sicily. Based on arguments made by Athenian leaders (including Alcibiades), the Athenians believe that there is a high probability of achieving large gains.28 This was the expeditionary force that Alcibiades, in a speech to the assembly answering Nicias, defended against Nicias’ ‘no expedition’ option. Alcibiades reiterated to his audience that an expedition would probably enable the Athenians to take Sicily and/or achieve other benefits (6.17-18). Alcibiades clearly got the better of this first exchange. Having listened to Nicias and Alcibiades, the Athenians remained convinced that a moderate expedition had a high probability of realizing a big gain. Having failed in his first attempt to scuttle the expedition, Nicias once again attempted, in a second speech to the assembly, to achieve his ‘no expedition’ goal – or, failing that, to eliminate the risk of a great loss (6.24.1). His second speech was based on what Thucydides (6.19.2) points out was a rhetorical ploy: Nicias realized that he could not achieve his best outcome of ‘no expedition’, so long as the choice was between ‘no risk and no gain (because no expedition)’ and the high probability of a big gain from a moderate, and thus not overly costly, expedition. Therefore, Nicias attempted the ploy of inflating, to a prohibitive level, the cost of the alternative to ‘no expedition’. He emphasized the residual risks associated with the moderate expedition (6.20.2-23.2). He told the Athenians that they could eliminate the risk of loss, and thus be certain that they would either realize substantial gains, or at least be sure of returning home safely – but only if they were willing to pay for a much bigger, much more costly expedition (6.23.3). In effect, Nicias kept his original ‘no expedition’ option on the table, but replaced Alcibiades’ moderate expedition option with a new super-sized expedition option, so the choice is now between no expedition and a huge and costly expedition. Nicias’s plan was a clever one. Based on the Athenian demos’ prior decision to approve the expedition and their enthusiastic response to Alcibiades’ defence of the expedition, he understood how most Athenians valued the 28

It should be noted that this belief might well have been a mistake, and a critical one. We do not explore it here, but the Athenians’ failure to appreciate the difficulty of the Sicilian expedition was possibly of great importance in explaining the dire consequences. Nonetheless, our point was that even if they were correct in their assumptions, they made a further mistake that could be attributed to the certainty effect: the mistake of sending the larger, more costly expedition. Thucydides’ focus on Nicias and his ploy to manipulate the discussion in the assembly shows that he thought this further mistake was also an important contributor in the resulting catastrophe.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

222

Ober and Perry

gains from the expedition. In the language of rational choice theory, he took their choice as a revealed preference. Nicias crafted his second proposition according to this preference, cleverly reframing the choice to be between ‘no expedition’ and ‘costly expedition’. If the demos had consistent preferences and was making decisions according to an accurate evaluation of probabilities, increasing the chances of success by a little bit would not have been worth the increase in costs. Like the Athenian envoys at Melos, Nicias made the error of assuming that his audience would think like Econs, properly weighting probabilities based on the expectation principle: Nicias’ rhetorical ploy is predicated on his belief that the Athenians would see that the cost of ‘insuring’ the relatively small probability gain (the gap between Alcibiades’ confident assessment of the likelihood of substantial gains and certainty) was irrationally high. The insurance of the vastly larger expedition was over-priced, in that the difference in cost of a moderate and huge expedition much exceeded the increased (imagined) probability of realizing the gain. If Nicias’ ploy had worked just as he planned, the Athenians would have concluded that the cost of the over-insured expedition exceeded the expected gain. And so, when it is a matter of no expedition or the super-sized expedition, the rational choice is no expedition. Interestingly, Nicias succeeded in persuading his audience regarding the residual costs of a regular-size expedition; they bought his framing of the problem and the proposition that the choice is really between no expedition and the super-sized one. Yet they did not conclude, as Nicias expected, that under such circumstances it is better to give up the expedition. Why was Nicias wrong? Had Nicias thought in terms of the certainty effect, he would have realized that ‘people are averse to risk when they consider prospects with a substantial chance to achieve a large gain’.29 The Athenians now viewed their reference point, the status quo which the no-expedition option would achieve, as far below the gains that they perceived as certain. Nicias unwittingly offered the Athenians the chance to ‘lock in a sure gain’ (or what they believed was a sure gain) by over-insuring against risk. The effect on the audience was not an Econ-like recalculation of the weighted probabilities in favour of the no-expedition option, but rather a cascade of enthusiasm for the sure-thing expedition, over-priced as it was. As Thucydides describes matters, in the aftermath of Nicias’ speech a ‘great eros’ arose among the assemblymen (6.24.3). The eros arose (at least in part) due to Nicias’ offer to guarantee the outcome (realize the gain of conquest or come back safely having done exciting things, i.e. having realized other utilities) if the Athenians will pay the extra premium for the super-sized expedition. 29 Kahneman, Thinking, Fast and Slow, p. 317.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

Thucydides as a Prospect Theorist

223

By paying a big premium to achieve what they (falsely as it turns out) believed to be the certainty of avoiding any bad outcome, and keeping alive the high chance of getting a really big payoff, the Athenians were, ironically in the context of Thucydides’ text, acting unlike their ordinarily risk-seeking selves (below, Corinthian assessment). They acted in the risk-averse manner associated with the conservative Nicias, and unlike the risk-seeking Alcibiades, in their willingness to ‘over-insure’ in favour of certainty. The irony deepens as we continue into Book 7 of Thucydides’ history and learn that the Athenians’ risk-aversion had the unexpected result of taking on a super-risky gamble, one that goes spectacularly badly in the end. Yet we can also now see why the passions ran so high in the assembly held to decide about the expedition: no one saw, ex ante, the residual risks of failure that became so evident ex post. Because of Nicias’ ploy, the most risk-averse among the Athenians (the older men, especially) seem to have got their way – the insurance is paid. Meanwhile, Alcibiades and the younger, more risk-seeking, of the Athenians got their way too – the thrill of the expedition and the chance at a big payoff. So, in the moment of choice, it seemed, everybody won (6.24.3). Or almost everybody. A few doubters remained, but they feared to speak out against the passionate and near-unanimous enthusiasm of the assembled Athenians (6.24.4). By drawing the reader’s attention to the silenced doubters, Thucydides underlines both the danger of cascades of enthusiasm among decision-making collectivities, but also the gap between the rational policy (either no expedition, or a moderate sized expedition with a high probability of success),30 and the actual choice made by the assembly. Once again, the attentive reader has learned an important lesson about how to anticipate decision-making by Humans in the face of high-probability events and under the shadow of the certainty effect. 5

The Corinthian Assessment: Possibility Effect and Loss and Gain31

The closely related Melian and Sicilian Debate narratives juxtapose the two sides of the certainty effect, as described by Kahneman and Tversky: in these two cases, foreseeable weighting errors, and thus irrational (from the point of view of expected utility) behaviours, arise in the face of a high probability of losses and in the face of high-probability gains. This is a marked juxtaposition 30 31

Cf. Orwin, The Humanity of Thucydides, pp. 118-9. This section borrows from Tomer Perry’s work in progress on Thucydides’ account of the Athenian regime and national character.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

224

Ober and Perry

in Thucydides’ text: the events on Melos and Sicily prove to be the hinge of the war, ultimately setting up what Thucydides called the greatest event in all Greek history: the catastrophic Athenian loss on Sicily (7.87.5). The juxtaposition of irrational choices, based on over-weighting options relative to their probability, is, we believe, strong, if not decisive, evidence that Thucydides was in fact thinking, and intended his reader to think, in terms of something like prospect theory in contrasting optimal choices, made on the basis of expectation and statistical probability, with actual choices made under the influence of hope and fear. The linked Melian and Sicilian Expedition Debate narratives allow Thucydides to show his reader important continuities across human societies: the powerful, confident, democratic Athenians in their public assembly calculate their interests in faulty but entirely Human ways, just as do the weak, fearful, oligarchic Melians when meeting in camera with the Athenian commanders. The apparently vast differences between the contexts tend to wash out, if and when we attend to the similarities of the weighting problem faced by the Athenians and the Melians alike. The difference in behaviour turns out not to depend on the assumed ‘national character’ associated with different peoples and regimes: we expect the democratic Athenians to be risk-seeking, the oligarchic Melians to be risk-averse; but they each defy our expectations. Rather, the difference in behaviour turns on whether the high-probability situation concerns a loss or a gain. The ‘human thing’, is, we learn, to make certain specific sorts of error in weighting when faced with uncertain prospects, and to deviate from rationality accordingly. The Melian and Sicilian Expedition Debate narratives offer particularly vivid cases for Thucydides’ exploration of the logic of prospective choices, but examples can be multiplied. The two cases considered above concern the certainty effect: prospective gains and losses in a situation of high probability. It remains to consider, more briefly, weighting errors in the face of the potential gains and losses associated with low probabilities. Once again, particularly obvious examples are available in a strongly marked passage of Thucydides’ text. In a much-studied speech delivered at Sparta by the Corinthians (Thucydides 1.68-71), a speech that Ober has discussed elsewhere under the rubric of the ‘Corinthian assessment’, Corinthian envoys to Sparta lay out to their Spartan audience the differences between the characters of Athenians and Spartans.32 32

Ober, ‘Thucydides and the Invention’; J. Ober, ‘Thucydides on Athens’ Democratic Advantage in the Archidamian War’ in D. Pritchard (ed.) War, Democracy and Culture in Classical Athens (Cambridge: Cambridge University Press, 2010), pp. 65-87. See also

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

Thucydides as a Prospect Theorist

225

The distinctions they set up prove to be generally valid, when tested against Thucydides’ subsequent narrative, although, as we have already seen, there are limits to the predictive power of these ultimately regime-based assessments of the expected behaviours of different societies. On the face of it, there is tension between the discussion of ‘national character’ and prospect theory, as we discussed it so far. The Corinthian assessment seems to suggest that the Athenians and Spartans differ in their attitudes towards risk. But prospect theory rejects the notion that attitudes towards risk are merely a matter of preference, suggesting instead that they are typical responses to specific situations, when people face certain kinds of risks and gambles. Prospect theory departs from rational choice theory on this issue because it insists that choice is made relative to a reference point. In the cases discussed so far, in the Melian Dialogue and the debate concerning the Sicilian expedition, it was unproblematic to define the reference point as the status quo, the state of affairs before the launch of the expedition or prior to the meeting between the envoys. However, city-states may differ in terms of their typical reference points, depending on their regimes. Political structures, decision-making processes, and the related political culture (for example one of participation and deliberation) tend to make certain courses of actions more salient and therefore frame decision-making very differently, leading to distinct decision-making patterns. We argue that the Corinthian assessment of the ‘national characters’ of the Athenians and Spartans should be understood in that light. That is to say, the argument that risk-averse or risk-loving behaviour is connected to a national culture is just the claim that the political cultures and regimes of the different cities tended to present choices in different ways. In democratic Athens the masses participated in deliberation before decision-making, voting for and against possible courses of action. The inclusion of more people and voices increased the variety of propositions brought to the table, resulting in consideration of more unusual and unexpected ideas, of the kind that had low probability of success but very high potential gains. Moreover, in fifth-century Athens the decisions were officially made by the demos and were therefore not attributed to the individuals who originated them, as they later would be. It was only in and after 415 (the date of the Sicilian Debate) that the Athenians passed new laws that subjected the successful proposer of legislation to prosecution if his measures were subsequently deemed

Desmond, ‘Lessons of Fear’, p. 367; Luginbill, Thucydides on War; Edmunds, Chance and Intelligence, pp. 40-1, 89-90.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

226

Ober and Perry

unconstitutional (paranomon) or inexpedient (me epitedeion).33 Through most of the era described in Thucydides’ history, there was less personal cost to raising an unusual idea: if it was really bad, the Athenians reasoned, it would be rejected; if it was adopted, it no longer rested only on the shoulders of the individual who proposed it. In contrast, the oligarchic Spartans made decisions in closed sessions, among themselves. All full Spartan citizens (several thousand men) were entitled to vote, but proposals were made only by those in established leadership positions. Those leaders had a deep vested interest in the status quo which meant they perceived most possible changes as losses (Thucydides 1.79-87). Moreover, the range of opinions included was certainly narrower and unusual ideas were proposed less often than they in Athens. Moreover, as Thucydides noticed, the atmosphere among Greek oligarchs was typically one of mutual fear and suspicion of one another’s ambitions; each oligarch would ideally like to seize power for himself (8.89.3: his example here is ‘The 400’ – an oligarchic junta – at Athens in 411 BC). In this environment, decision-making tended to be made in the shadow of possible threats – constantly considering unlikely yet vivid scenarios of terrible losses.34 Given those political structures, the psychological pattern that Kahneman and Tversky call the ‘possibility effect’ explains the difference in attitudes between the Athenians and Spartans. This is the low-probability level (the two lower boxes of the 2 × 2 matrix: Table 1) of the ‘four-fold pattern’: the one that describes decisions made when faced with low-probability alternatives. The Spartans, as described by Thucydides’ Corinthians tend to occupy the lower-right ‘losses’ box in the ‘low probability/possibility effect’ row (Table 1). The Spartans are said to be generally risk-averse because the emotion driving their key policy decisions is fear of even a small possibility of large losses; they focus on unlikely scenarios that have grave consequences. According to the ‘possibility effect’, in such situation people tend to over-estimate the likelihood of the terrible result and over-insure, for example by accepting unfavourable settlements. Sparta’s fear of loss, when faced with the growth of Athenian power, was according to Thucydides, the ‘truest cause’ of the Peloponnesian

33 34

M.H. Hansen, The Athenian Democracy in the Age of Demosthenes: Structure, Principles and Ideology (Norman, Oklahoma: University of Oklahoma Press, 1999), pp. 205-218. On decision-making by Greek oligarchies, see, further, M.S. Simonton, The Rules of the Few: Institutions and the Struggle for Political Order in Classical Greek Oligarchies (Stanford University: PhD dissertation 2012).

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

Thucydides as a Prospect Theorist

227

War (1.23.6).35 As a result of their fear, the Spartans decide to go to war when their position was stronger since they believed it will insure against any chance that they will lose their leading position in the Greek world. They believe Athens will continue to grow in power and present a greater threat to them in the future. The Corinthians’ rhetorical strategy in the speech mentioned above is based on portraying the increase in Athenian power as a large enough threat to convince the risk-averse Spartans that they must risk war in order to avoid a great loss. They must not, say the Corinthians, make a war-avoiding settlement with Athens, one that they might be prone to take given their risk aversion, because to do so would in fact raise the possibility of great loss (by implication: the end of Sparta’s position as a great polis, of its control over its subject population, and thus of the Spartan way of life). Because the Corinthians are successful in their portrayal of the situation to a risk averse audience, the Spartans come to see the cost of fighting the immediate war as a sure, but small, loss which they prefer to any chance of having to fight a stronger Athens in the future, and maybe losing everything. The Spartans’ risk-aversion is linked by the Corinthians with their habitual slowness to seize a potential opportunity. The Spartans do not change their risk-averse ways quickly. Their habitual slowness is, according to Thucydides in propria persona (8.96.5), what loses the Spartans the opportunity to end the war with a bold attack in 413 BCE, when Athens was still reeling from the catastrophe in Sicily. Rather than take a chance on the big gain of final victory, the Spartans apparently focused on the small chance of the loss they might sustain if their attack were to fail. The Athenians, as described by the Corinthians are the diametric opposite of the Spartans, and as such they occupy the lower-left ‘gains’ box in the ‘low probability/possibility effect’ row (Table 1). The Athenians, according to the Corinthians, are risk-seekers, tending to gamble on long chances in hope of a large gain (like a modern lottery player). As such, when judged by the standard of rationality, they are potentially too quick to act; over-willing to take risks in low probability ventures. One predictable result is the Athenian tendency to reject favourable settlements – as indeed they do in 425, after their victory 35

We differ with Orwin, The Humanity of Thucydides, pp. 32-37 and Appendix 2, in how to interpret the key phrase. Orwin translates prophasis as ‘allegation’ but we do not understand how something can be at once ‘alleged’ and yet ‘least expressed’ (aphanestatên logôi); it seems to us that an allegation is, by definition, expressed rather than hidden, whereas causes, including the truest, are sometimes most obscure. Moreover, at 1.88.1 (with Orwin p. 42), Thucydides in propria persona says that Spartan fear of Athens was their actual (not just declared) motive for declaring war.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

228

Ober and Perry

at Sphakteria. The Spartans sue for peace, but ‘the Athenians, however, kept grasping at more, and dismissed envoy after envoy’ (4.41.3-4). The Corinthian assessment, like the paired Melos and Sicilian Expedition Debate narratives, show us two sides of a specific psychological effect. The Corinthian assessment is not, however, strictly parallel: the Corinthians’ point, made to their Spartan audience, is that the Athenians are having the better part of it. It is the Spartans, they claim, who must change their habitual riskaversion if they are to compete with the risk-seeking Athenians. They are, of course, wrong. Long term deviation from rationality, as Kahneman notes with regard to litigations, is counterproductive.36 Although accepting an unfavourable settlement (the risk-averse Spartan tendency) or rejecting a favourable settlement (the risk-prone Athenian tendency) on a given case may save the trouble of, say, a law suit, an organization that systemically deviates from rational­ity in either way stands to lose much in the long run. Yet for much of the war, it did seem as if the Athenians were faring better than the Spartans. It seems that despite the fact that they overestimated their chances for gains when adopting many of these ‘unusual’ ideas, they still tended, on the whole, to make more accurate estimations. Why, readers ought to ask, should that be? Not, if we are to follow the logic of Thucydides’ claims about the continuity of human nature, because Athenians differ from their opponents in their fundamental nature. Nor does Thucydides’ narrative support the idea that (as the Athenians were prone to think) Athenians were inherently cleverer than other people, or (as optimistic democrats might think) that democracy inherently yields more rational judgments when managing risk.37 Thucydides suggests, rather, that at the time to which the Corinthians referred (in the years before the outbreak of the War), Athens had superior leadership. 6

Pericles and Leadership

Leadership is a prominent topic in Thucydides’ narrative. Thucydides did not expect Humans to change their fundamental nature by transforming themselves into Econs as a result of reasoned arguments. But he did, as we have already seen, believe that some individuals are better than others at employing 36 Kahneman, Thinking, Fast and Slow, pp. 320-21. 37 Though democracy has distinct advantages, and disadvantages, over its contemporary alternatives in the eyes of Thucydides. See Ober, ‘Thucydides on Athens’ democratic advantage’; J. Ober, Political Dissent in Democratic Athens: Intellectual Critics of Popular Rule (Princeton, N.J.: Princeton University Press, 1998), chapter 2.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

Thucydides as a Prospect Theorist

229

Econ-like reasoning in times of need. That ability is, for Thucydides, a necessary but insufficient condition for superior leadership. When leaders do not take the human thing adequately into account, as Nicias and the Athenian envoys did not, their societies risk leaning too hard on Orwin’s Athenian thesis, with potentially catastrophic results. But if a leader is able to reason as an Econ and is capable of anticipating and addressing the foreseeable weighting errors of others, and if his position is strong enough to enable him to direct state policy over a period of time, the result may be outstanding state performance. This is exactly what Thucydides leads his readers to believe was the case in the lead-in to the Peloponnesian War, when, Thucydides tells us (2.65.9), Athens was nominally a democracy but Athenian policy was directed by Pericles, Athens’ ‘first man’.38 Although Thucydides is not blind to Pericles’ policy failures, Pericles is presented as a near-ideal leader for Athens in the run-up to the Peloponnesian War and in the early years of the war.39 Pericles’ excellence as a leader lies in part in his Econ-like capacity to judge situations dispassionately, in the light of all relevant information, and to give the right weight to probabilities in designing policy. Central to the presentation of Pericles’ virtuous leadership is Pericles’ conviction that policy based on reason can be used to overcome the vicissitudes of chance.40 But that is only part of the story. Because Athens remained procedurally a democracy, and all legislation had to be enacted by a citizen assembly, Pericles must also excel in his capacity to recognize the likelihood and the direction of probability-weighting errors on the part of his fellow Athenian citizens. Finally, he must be able to at least partially correct for those anticipated errors through affecting the relevant emotions, and thus the behaviours and attitudes, of the citizenry. The tool by which he accomplished 38

J. Ober, The Athenian Revolution: Essays on Ancient Greek Democracy and Political Theory (Princeton, N.J.: Princeton University Press, 1996) chapter 6. Ober argued that this is not an accurate portrayal of Athenian democratic government, but that is beside the point of this paper, which is to understand Thucydides’ thought, not Athenian politics as such. 39 Orwin, The Humanity of Thucydides, p. 28. 40 Edmunds, Chance and Intelligence, pp. 7-22, 205-214; Desmond, ‘Lessons of Fear’, pp. 371-3. We cannot discuss Edmunds’ comprehensive study on chance and planning in this paper, though it is important to note that Edmunds concludes that Thucydides, though close to Pericles, ultimately thought that chance can be overcome by rational planning only with the hindsight of a historian. Whether or not it is possible for a statesman or historian to completely eliminate the influence of chance, the main point remains: Thucydides thought it worthwhile for statesmen and leaders to invest in strategic planning in order to overcome uncertainties and Pericles is Thucydides’ prime example of the benefits of this effort.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

230

Ober and Perry

this was rhetoric. Rhetorically sophisticated public speeches are, of course, an outstanding feature of Thucydides’ text, and Thucydides was a careful student of rhetoric and rhetorical effects. Pericles is singled out by Thucydides for his distinctively productive (when compared to later, inferior Athenian leaders) use of rhetoric. As Thucydides famously noted (2.65.9), Pericles used rhetoric to correct for potentially dangerous swings in public emotion. His speeches encouraged the Athenians when they were tending to be overly cautious (i.e. too likely to accept unfavourable settlements) and spoke to restrain them when they were over-enthusiastic (i.e. too likely to reject favourable settlements). All of this points to a leader who makes policy based on expectations (and thus proper weighting of probabilities) and who understands the tendency of the citizenry to fall into prospective errors. Thucydides’ Pericles, we may say, employed his rhetorical skills to keep Athens within a tighter band of formal rationality than they had resources to keep themselves: he encouraged them to judge prospects more as if they were expectations, in light of relevant information and statistical probability. This is, we think, a plausible context for understanding what Thucydides emphasizes as Pericles’ central skill as a leader: his foresight, pronoia. To attribute foresight to Pericles (or other leaders, notably Themistocles) is not to imply that Pericles could somehow mysteriously predict the future. Rather it is to underline that he was very good at information management, at calculating probabilities, and at managing prospective judgments, unbiased by hope or fear. Pericles’ possession of pronoia does not mean that his plans always go well. While the risks Pericles urges the Athenians to take are reasonable, they inevitably entail a chance of loss or unrealized gain. Inevitably some of Pericles’ rational gambles will not pan out, and we should expect some noteworthy failures. But we should also expect that, over time, rationally weighting probabilities will work in Athens’ favour, and thus that under Pericles’ leadership Athens will grow in wealth, power, and security. In the decades before the Peloponnesian War, both of these predictions prove true (as we see from Thucydides’ narrative of the so-called Pentacontaetia): there are setbacks (e.g. the Egyptian campaign of the 450s, the Boeotian campaign of the early 440s: 1.109-114). Yet in the period of Pericles’ leadership Athens became exceptionally prosperous (2.65.5, 2.65.13).41

41

Although Pericles is not prominent in the Pentacontaetia narrative, Thucydides’ astute reader is, we suppose, expected to read the assessment of Pericles at 2.65 back into the period of Athenian growth that corresponded with his preeminence.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

Thucydides as a Prospect Theorist

231

Thucydides famously castigates Pericles’ successors as inferior leaders, who ‘ended by committing even the conduct of state affairs to the whim of the multitude. This, as might have been expected in a great and sovereign state, produced a host of blunders, among them the Sicilian expedition . . .’ (2.65.10-11). Thucydides’ analysis of why the Sicilian expedition failed is complex, but this passage brings us full circle – to a condition in which foreseeable prospective errors are augmented, rather than corrected by the speeches of would-be leaders (like Alcibiades and Nicias). This will in turn lead to bad results that may still shock, but will no longer surprise, Thucydides’ attentive readers. Reading Thucydides as a prospect theorist interested in leadership has the potential to enlighten more passages than we have space to discuss here. Consider for example, the ‘sunk cost fallacy’ – as it is expressed in the tendency of leaders to become over-invested in their favourite projects. The Sicilian expedition turns from a very bad situation into a total disaster, because Nicias, the last Athenian commander standing in Sicily, becomes grossly over-invested in the expedition he had once opposed, and refuses to cut his losses after the circumvallation strategy for capturing Syracuse fails. Again, human nature is consistent over time: Nicias’ behaviour closely models familiar modern cases of CEO’s whose over-investment in failing projects creates serious principalagent problems in respect to the organization they lead.42 7 Conclusions We have focused in this paper on Thucydides’ precocious grasp and clear presentation of certain persistent features of human psychology and behaviour, features that have only recently been adequately explained by contemporary social scientists. As noted above, Thucydides has been embraced by some theorists of International Relations as a Realist avant la lettre. He is certainly that, when read on one level. But, unlike some strong Realists, Thucydides was well aware of the limits of ordinary human rationality. Thus, he was also an astute critic of the kind of thinking that takes ‘Realism as rational choice’ to be the whole truth about human nature. Thucydides was intensely concerned with humanity ‘in the round’. He does not make the error of thinking that Humans are, or ever can be, entirely rational. But neither does he suppose, with tragic Realists, that our irrationality dooms

42 Kahneman, Thinking, Fast and Slow, pp. 345-46.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232

232

Ober and Perry

us to the certainty of tragedy.43 Rather, Thucydides suggests that the right kind of leader can help to control the more extreme effects of hope and fear on risk assessment. Moreover, leadership of this beneficent kind need not be merely an accident of nature. Thucydides wrote in the hope that his work would be ‘judged useful by those inquirers who desire an exact knowledge of the past as an aid to the understanding of the future, which in the course of human things must resemble if it does not reflect it’. (1.22.4). We take this to mean that at least part of Thucydides’ goal in writing was to produce future statesmen with the skills necessary to improve their communities. One of those skills is, we have argued, an understanding of what we now call prospect theory and the rhetorical tools capable of affecting public emotions and thereby limiting the impact of foreseeable prospective errors. Attending, with Orwin, to the ‘humanity of Thucydides’, means that we must resist falling into the hero-worshiping error of imagining Thucydides to be an infallible, god-like guide to human behaviour in conditions of extremity. We must attend not only to Thucydides’ insights but also to his limitations. If Thucydides has one core lesson to impart it might be summed up as, ‘Observe accurately! Criticize sharply! Test theory against evidence!’ In that spirit, we need not accept Thucydides’ analysis of the operations of Greek political institutions, or his explicit or implicit arguments for why (or even whether) they failed. We have discussed, elsewhere and in some detail, some areas in which Thucydides’ argument about the inherent flaws of Athens’ democracy in the absence of Periclean leadership may be regarded as having gone wrong.44 But, by the same token, we remain convinced that the more carefully we read Thucydides, as a historian and as a social theorist, the more impressed we will to be at his analytic rigour, his prescience, and his humanity.45 43

On tragic realists, see R.N. Lebow, The Tragic Vision of Politics: Ethics, Interests and Orders (Cambridge and New York: Cambridge University Press, 2003). 44 Ober, The Athenian Revolution, chapter 6; Ober, Political Dissent, chapter 2; Ober, ‘Thucydides Theoretikos/Thucydides Histor’; Ober, ‘Thucydides on Athens’ democratic advantage’. 45 This paper originated in a panel at the American Political Science Association, August 30, 2013, centred on an appreciation and re-evaluation of Orwin, The Humanity of Thucydides. Our thanks to Clifford Orwin, the other participants in that panel, the members of Stanford’s Ethics and Politics, Ancient and Modern workshop and Tomer D. Ullman for helpful comments.

polis, The Journal for Ancient Greek Political Thought 31 (2014) 206-232