The Shakespeare Conference: SHK 24.0378 Thursday, 1 August 2013
Date: August 1, 2013 12:58:31 AM EDT
Subject: Two Sets of Controversies
Ward Elliott posting on SHAKSPER bet controversies 7 7 13
Two sets of controversies.
It’s summer, and both of SHAKSPER’s old-time bet controversies are back with us. The first controversy has to do with our £1,000 bet, which quietly passed its tenth birthday on June 9. It is still on offer and still seems controversial enough for people to gripe about, though never enough for anyone to take us up on it. If our bet were a kid, it would be alive and headed for the fifth grade. Joe Egert now has reservations about it. He thinks our evidence is circular, and he urges us to raise the ante.
The second controversy is the latest of three rounds of Michael Egan’s counterbets. These might seem more like feints than real bets, because every time he has almost gotten his demanded hearing underway, he has yanked away the football like Lucy and stormed out before a verdict could be reached. I don’t suppose that he, or Lucy herself for that matter, actually sets out consciously planning to play the Lucy role, so words like feint and strategy may not be quite the right ones. But it always seems to work out that way. My guess is that he starts out with an all-consuming desire to stir up some attention for his Woodstock ascription, demands a hearing for it in peremptory tones, then gets cold feet when others balk at his insistence on doing it his My Way, feels that he is about to be caught in someone else’s trap, and bails out in alarm. But it’s his own trap, not someone else’s; it has the same bad effect as an intentional Lucy bait-and-switch strategy; and it does drag out the controversy, in this case, longer than it deserves.
It also backfired badly in his last round. He stormed out, as always, but too late to forestall the verdict he had demanded. He got his verdict, lost, owes us £1,000, and won’t pay it. Last week he came back yet again, incredibly, denouncing Larry Weiss, the convenor of his demanded hearing panel, for taking an “unhealthy interest” in his case – and demanding that Mr. Weiss go another round with him, this one on his own home court. If the Egan counterbets were kids, and were still alive, his oldest would be eight this week. But all of them are dead, at Egan’s own hand. On the other hand, it looks like he has a new, fourth one on the way, but for Weiss, not us. Will it live any longer than the others?
Our original bet.
Our original bet comes up whenever the talk turns to the acceptability of stylometry, like ham and eggs, more often than not because I bring it up myself to reaffirm its original point: some debates, deadlocked on opinion, can be settled on the facts with a bet. Ten years ago, as now, many SHAKSPERians were uncomfortable with stylometry. They didn’t want to discuss it themselves, and they didn’t want us to bring it up either. Some strongly urged that stylometry be outlawed on SHAKSPER discussions as unacceptably deceptive, untrustworthy, and circular. My co-author, Rob Valenza, and I have many times defended our stylometry and resisted having it outlawed, wherever possible using simple, non-technical devices to help clarify technical points to a non-technical audience. Our bet was such a device. It was intended to clarify and unjam a deadlock of conflicting opinion with an appeal to facts. The gist of this long posting is that it worked, much better than weeks of fruitless head-butting over matters of opinion, belief, and acceptability, and that it still serves a useful function. For somewhat longer answers, read on, or skip to the summary of “lessons learned” at the bottom. For yet longer, more nuanced and documented answers, see our many previous postings in SHAKSPER’s archives, or our section on the original bet in our “Oxford by the Numbers,” 2004, 363-65, http://www.cmc.edu/pages/faculty/welliott/UTConference/Oxford_by_Numbers.pdf.
Let’s start by mentioning Joe Egert’s posting and come back to it at the end. The other week he deplored “the circularity that lies at the heart of early modern stylometric methodology,” and which certainly lay at the heart of our original bet. He thought our still-open thousand-pound bet offer was too niggardly, shouldn’t really count as a “victory” for us. He urged us to raise it to eighteen thousand and make it more enforceable, so, if he takes the bet, he won’t get stiffed. That may be worth some thought. In the old days it took us just a few seconds of reflection to raise it from $25 to $1,000, and two years to raise it again from $1,000 to £1,000. Does it need more?
Gatekeeping, proxy evidence, and transaction.
Nobody spoke or thought of our bet in quite those terms when we first offered it, but, in retrospect, the key problems we addressed were gatekeeping, proxy evidence, and transaction costs. Economists use the technical term transaction costs, and so do we, but jargon-spurners can think of them as time costs or hassle costs without losing too much fidelity. A surprising number of academic controversies, especially knotty, technical ones over authorship, are not settled on the merits, which most people don’t understand, but by proxies, which most people think are easier to understand – clever rhetoric, flaunting or weighing credentials, appeals to higher authority, haloes, anchors, votes, trials, bets, flipping coins, and so on. All of these can be proxies for breaking deadlocks and getting enough closure to keep a conversation alive. Few of Donald Foster’s American admirers had the foggiest idea whether his SHAXICON evidence in Elegy by W.S. actually proved what he claimed it did, but he was a good story-teller and, for a while, a thriving self-promoter in celebrity authorship cases. Everyone understood and admired when he correctly identified the author of Primary Colors. Many concluded that, if he got Primary Colors right, he must have gotten the Elegy right, too. They were right about Primary Colors, wrong about the Elegy. Later, when Foster conceded that he got the Elegy wrong, most people accepted that, too, again without ever trying to puzzle their way through his actual evidence or to study countercases from us (2001), Gilles Monsarrat (2002), or, especially, Brian Vickers ‘Counterfeiting’ Shakespeare (2002). Proxies can help you to infer the unseen from the seen, the complex from the simple, and the difficult from the easy.
Does anyone doubt that authorship itself is a kind of proxy writ large? Just the other week, according to Nielsen Bookscan, The Cuckoo’s Calling, by no-name “Robert Galbraith,” had sold only 500 copies in the U.S. Now, with the help of a bit of stylometry, it turns out that it’s actually by superstar J.K. Rowling. Would anyone want to bet that Rowling’s publisher guessed wrong in printing up another 300,000 copies? Or that Woodstock’s sales wouldn’t pick up considerably if it were shown to have been actually written by dead, white, male superstar William Shakespeare?
Proxies are usually quicker and easier than direct evidence, and often almost as useful for practical purposes. Most people don’t know stylometry, don’t want to learn it, and find it more convenient to defer to the say-so of stylometrists, or to other knowledgeable gatekeepers, or to the collective orthodoxies of their own group, than to pause to sort out knotty technical questions on the merits on their own. For most of us, most of the time, life goes farther and better with proxies than without them (Kahneman, Thinking, Fast and Slow, 2011, Ch. 7-9).
In our case, both proxies and collective orthodoxies were in play. In 2003 we were newcomers to SHAKSPER, outsiders who thought that our Shakespeare Clinic students, using unorthodox new-optics methods, had made some major advances, discoveries, and validations. We thought that we could contribute something worthwhile to SHAKSPER’s discussion of possible co-authorship in Titus Andronicus. But no. SHAKSPER had adamant, vocal, self-appointed gatekeepers who would have none of it, not because they had read and rejected our evidence, but because they didn’t trust statistics of any kind. To them it seemed a given that statistics are circular, tell you nothing you don’t already know, and have no place in a legitimate authorship discussion. They told us, firmly, pointedly, and repeatedly to leave our statistics at the door. I would guess that many, maybe most SHAKSPERians shared their distrust of quantitative evidence like ours, though perhaps not their adamant insistence that it must therefore be outlawed from SHAKSPER discussions.
We spent several fruitless sessions patiently responding to each of our gatekeepers’ points, but we soon realized that we were getting stonewalled. We had three choices: struggle on with the endless, uphill deadlock over admitting our evidence, fold our tents, or find a better proxy. We went with the proxy. When it occurred to us that the gatekeepers’ circularity theory was testable, it also occurred to us that a bet was a good way to test it. Whichever side was right, the bet could test it and settle it with objective, empirical evidence and with a minimum of fuss. Better that than spending hours and weeks venting people’s opposing philosophies about banning statistics but never settling the controversy. Everyone, said Daniel Patrick Moynihan, is entitled to his own opinion, but not to his own facts. Betting is better suited to events where there is an objective winner, like a horse race, than to events where winning is a matter of taste, like a beauty contest. With most horse races you can tell who won, without endless, inconclusive argument whether the judges were using acceptable procedures. With horse races, the transaction costs of telling the winner are much lower. Ours was a bet of the horse-race type.
If our gatekeepers really thought they were right, they would expect to win on the facts and take the bet. But they didn’t take the bet, and that tells us something.
The gist of our bet was this: that no one could come up with a single untested play, among hundreds available, that our Shakespeare-or-no tests would misclassify. If we were right that our new-optics results were non-circular and extrapolable, the patterns we found in tested plays would also be found in untested plays. We would win. If the gatekeepers were right, that our methods are circular and couldn’t be extrapolated, they would win. We originally conceived it as a $25 bet, but immediately realized that it had to be much higher to justify the testing costs. The transaction costs of editing and testing a new play are not zero, and not trivial. It makes little sense to incur a thousand pounds worth of preparation and testing cost to settle a $25 bet. We raised it at once to a thousand dollars, and eventually raised it again to a thousand pounds, both increases intended to discourage frivolous takers.
But we also made it clear that takers were welcome to pretest the plays themselves, with free software supplied by us, so that, if they wished, they could screen scores or hundreds of plays to find just one that our tests would misclassify, at zero financial risk to themselves. If our theory was right, and they didn’t hit the jackpot on the first couple of tries, and actually had to test scores of new plays to find one that would pass our Shakespeare tests, the takers, not us, would be the ones who would get stuck with the hefty editing and pretesting costs. If their theory was right, then the untested bulk of the haystack could just as well have been loaded with needles, which doubtless could easily be found with just a test or two, particularly if guided by the gatekeepers’ educated intuition, which they assumed to be far more reliable than our pages and pages of dull, dry, mechanical, circular statistics. That’s testable, too, if you are willing to find out the stats. Our gatekeepers loved sports imagery, as did we, and they used it both boldly and misleadingly, as a proxy to support their stats-are-circular theory. If they had read, Moneyball, which was first published only two weeks before our original bet, they might have been less wedded to their circularity theory.
We thought the circularity theory was wrong, confidently enough to bet a thousand pounds against it and on our own tests. It seemed like high stakes both then and now, and could have been taken as a sign that we were either very confident or very bold as bluffers, take your pick. In fact, our high confidence levels shouldn’t have mattered in the least, and probably didn’t (low confidence levels might have been more telling). They were, and they remain, a matter of our feelings and opinions, not of fact, and everyone knows that many views held with the deepest, most sincere and passionate of convictions have turned out to be false. What did matter, a hundred times more than ours, was the low confidence levels of our would-be gatekeepers themselves, which could easily be inferred from their behavior. If they really believed they were right, our bet offered them be a big, easy, risk-free reward, plus a gratifying, high-profile, clear-cut victory over the barbarians, which could give them enough credibility and clout, or us enough embarrassment, to keep us fenced out for many years. If they really believed our evidence was circular, they would have grabbed it.
But they didn’t, and haven’t, nor has anyone else in the ten years since—probably, we would guess, because they, too, silently sensed that the circularity argument was far too weak for any rational, loss-averse person to bet on. The original gatekeepers wisely folded their tents and have not tried to obstruct us from that day to this. For this they have our thanks. Our thoughts on Titus Andronicus were duly entered, and we have had years of free, fruitful, unobstructed access to SHAKSPER ever since. We are grateful to have gotten through the gates, with much help from our bet, and we remain grateful enough to revisit the bet episodically to see what can be learned from it.
Some Pros and cons of proxy proofs.
Have the gatekeepers’ years of reticence actually proved that we were right and they were wrong? No, it’s just a proxy, not a real proof. For a real hard-nosed standard-bearer, only testing all available plays could do that, and, even if that were done, well-entrenched skeptics could still protest that testing the known, available plays was not enough. Maybe some of the known lost plays, if they could be found, or the unknown ones, would prove their point. “Perfection or nothing,” is their unstated motto—as long as the other guy is the one expected to provide it. But the underwhelming response did show that the gatekeepers themselves didn’t and don’t believe enough in their own theory to bet on it, not even if they could pretest hundreds of plays ad libitum and risk no money. It was a proof by proxy, but a good enough one to free us from the unreasonable expectation that, to prove anything, we had to prove everything first. Our bet was, and it remains, a much better proxy than the gatekeepers’ own “everybody knows statistics are circular and can’t be extrapolated” argument, for two reasons: (1) In Twain’s words, “It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so;” and (2) the gatekeepers’ actions, if not their words, said clearly they didn’t believe in it enough themselves to risk money or time on it.
Did our proxy spare us endless wrangling and transaction costs as much as we had hoped? At the time, absolutely. The gatekeepers stopped obstructing, and we became full, free, happy participants in SHAKSPER authorship conversations, most of which would never have taken place in SHAKSPER’s absence.
The Egan Counterbets.
In short, our own bet still seems to me a success story. It paid respectful attention to the opinions of the gatekeepers, but did oblige them to decide, tacitly, whether they believed enough in them to bet on them. They wisely decided no, the deadlock was unjammed, and the gates were opened with much lower transaction costs than we would have gotten from more months of deadlock. It’s much harder to think of the Egan Counterbets as a success story. I think of them as tragicomedies of transaction costs, a series of battles over how much time and effort Egan could reasonably exact from the other side and the judges before he scampered off with the football in the final act. He seemed desperate to have a hearing for his “irrefutable” Woodstock case, with willing debate partners, judges, and a verdict, and he wanted us to provide all of these but the verdict in his favor.
We were willing to do so, with the help of volunteer mediator Larry Weiss, but only at a reasonable scale, not too grossly disproportionate to the amount at stake. We didn’t want to pay for Egan’s blue-ribbon panel, nor to oblige any panel, far less ourselves, to read everything he thought necessary to make his case, nor to let the debate drag on indefinitely. He wanted all of these at the outset, yielded on a few of them over the eight years of controversy, but was adamant to the end, even after agreeing to the contrary, that everyone had to read all, or, at the very end, two volumes, of his sprawling, $400, 4-volume Woodstock treatise before justice could be done. Whenever we balked at saddling the judges, and ourselves with £10,000 worth of transaction costs to settle his £1,000 bet, he would snatch away his football like Lucy, stomp out in a rage, denounce the process as a “charade and a show trial,” declare the case closed – and then, like Lucy, get ready to repeat the process a few years later. Our mistake was to suppose that his obsession with getting a public hearing for his case actually extended to getting it settled with a verdict. Obviously, it didn’t. Stir now, serve later, was his motto, and it produced much stirring, but little serving.
Though he published, posted, or sent out hundreds of pages of his data and analysis – or was it thousands, counting his webpage before it was discovered and he disavowed it and took it down? – he was quick to disavow all of it as “his case.” But for someone outwardly desperate to get a hearing and a favorable verdict, and whose only favorable reviewer described his sprawling, wandering, unindexed, Woodstock treatise as a “reader’s nightmare,” and despite repeated requests, he never came up with a compact case statement of his own till a year and a half after the third-round verdict had already been handed down. It’s as if he feared that his needles of evidence would lose their irrefutability if anyone separated them from his vast, protective haystack. If so, he was probably right, since his case was weak, and it was harder to hide that in 40 pages than it was in all four costly, sprawling volumes. Eventually, a year and a half after Round 3 was over, he did produce a 40-page brief, whose shortcomings were less like a needle in a haystack, and more like a banana on a pool table. By then, it was much too late to have had any effect on the Weiss panel’s verdict on Round 3. But perhaps, if he doesn’t disavow it and duck out yet again, and there is a Round 4, he can hope it will help him with that.
It’s important to note that all of Egan’s bets were his, not ours, and that we had no legal or ethical obligation to play Charlie Brown to his Lucy, far less to accept the heavy burden of unreasonable transaction costs he wanted us to shoulder. Our bet had to do with trying our tests on untested plays to see if they would work, as we expected, or not work, as the circular-theory folks expected. Egan’s bet had to do with applying his own evidence to a play already tested by us and roundly rejected as Shakespeare, as it was, and still is, rejected by orthodox Shakespeare scholars. Ours was on an objective question of fact, his on a matter of opinion. Ours was still a horserace, his a beauty contest. On the other hand, if he was right, it would be a glaring indicator that our tests weren’t 100% accurate, only 99%, and that there might be a grain of truth in the stats-are-circular theory. If so, addressing it directly and getting it decided, with Egan’s co-operation and Larry Weiss’s mediation, seemed less likely to make trouble than letting it sit and stew with the circularists.
So we negotiated with Egan in 2005, with help from Larry Weiss, to find a suitable panel, a debate prompt, and written safeguards against out-of-control transaction costs. Both parties were to submit short case statements and closing statements, with no limit on supporting documents, but no requirement that everything be read, and no more than a month of debate. Egan agreed at once to the panel, all of SHAKSPER’s membership, to the debate prompt, and to the unlimited right to submit documents. But not, most emphatically not, to the one-month limit, nor to letting the panel decide for themselves how much or little of his reader’s nightmare they needed to read before ruling. For him, it had to be the whole nightmare or nothing, with discussion for “as long as it takes.” We balked, and he stomped out, twice, denouncing us as “fakes and phoneys.” “The bet has been called off and the matter is now closed.”
That was the first two rounds, and of course the matter was not closed. He tried unsuccessfully to arrange a third round through Roger Stritmatter, of the Shakespeare Oxford Society. We declined. Then he tried again, but, having walked out twice, and called us less temperate names than most would use who are seeking to win someone’s co-operation, he was forced to sweeten the pot considerably to show some semblance of good faith. This time he insisted that he really, really wanted his case heard and decided, and would make us an offer we couldn’t refuse: Pick any panel we like, £1,000 for us if we win, nothing for him if we lose, time limits OK, case-statement word-limits OK, unlimited supporting submissions OK, but panelists were required to read only as much of both sides’ submissions as they deemed necessary to render a fair judgment. No obstruction or breach of agreement allowed, or Weiss could declare a forfeit. Both sides agreed to all of these. Weiss recruited a panel of himself and two other SHAKSPERians from the Golden Ears pool, Will Sharpe, and Dale Johnson. Then, in breach of his agreement, Egan changed his mind yet again and demanded that everyone on the panel read hundreds of pages of his reader’s nightmare, as his oversized case brief. Two of the panelists were willing to read all of this, one not. That was not enough for Egan, who once again snatched away the football, denounced the process as a “charade and a show trial,” and stomped out.
Another Aughh! situation for the Charlie Browns? Not quite. This time the submissions to the panel were complete in accordance with the rules to which both sides had agreed. The panel went ahead and decided the case on its merits, unanimously, in our favor. Larry Weiss’s opinion is available at http://shaksper.net/documents/doc_download/274-egan-v-elliott. Our SHAKSPER posting gives the gist of our case http://shaksper.net/archive/2011/303-september/28127-thomas-woodstock- and contains further leads to our brief and timeline, both available online.
A year and a half later, Egan finally did submit a 40-page brief as a rebuttal to the panel’s verdict. See below for the link: http://shaksper.net/archive/2013/337-february/29094-in-the-case-of-egan-vs-elliott-a-reply-to-larry-weiss-et-al
If your case is strong, a 40-page brief normally shows it much more clearly than a 2,100-page brief. If your case is weak, a 40-page brief shows that more clearly, too. Just like ripe ones, rotten bananas are more obvious on the pool table than they are buried in the haystack. Egan’s long-overdue brief makes it all too clear that his case is weak. The nub of his argument is that his own evidence, though weak by the ounce, becomes irrefutable by the ton. “No single datum clinches the matter but hundreds en masse seem unanswerable (his brief, 2013, 6).” That’s implausible enough on its face, but especially hard to swallow if there is also a bushel of strong, unanswered contrary evidence, as there is in Woodstock’s case. We took Shakespeare’s measure in 152 tests, old and new. In 62 of these, Woodstock fell outside Shakespeare’s range. That means we found 24 more Shakespeare rejections in Woodstock alone than we found in our entire corpus of 29 single-authored, core-Shakespeare plays combined. The odds that Shakespeare would produce that much discrepancy from his own norms by chance in just one play are lower than the odds of getting hit by lightning. He responded by declaring that our “assumption . . . that a writer as varied and complex as Shakespeare can be pinned down by numbers” is “unacceptable (his brief, 2013, 5).”
Does this sound familiar? It should. It’s another invocation of the circularity argument, just as rickety as the original ten years ago, and just as testable on the facts. That’s what our original bet was about, and why we have rise to its defense every few years whenever a reminder looks called for. If Egan is right, he could easily refute us by coming up with just one untested play that our tests misclassify. He hasn’t even tried. Once again, the proper question is not whether it’s acceptable, but whether it is so.
When he seemingly sought to settle his ascription via someone else’s opinion, as with his own dead counterbets, we were willing to go along with that, too, three times. In the end, after imposing a lot of transaction costs on the Charlie Browns, he got his judgment, which he lost and wouldn’t pay up. That should be enough baggage to end our years (if not Larry’s) of on-call servitude as Charlie Browns. Egan, having defaulted on the third round, has not tried to tempt us into a fourth round. Who would bet on the next sweetened pot if the last one was a fake? Our guess is that we’re off the hook at last. If so, I think we do have Egan’s own bets, plus years of patient negotiation and argument on our part, and patient mediation on Larry Weiss’s part, to thank for making each successive round a bit more demanding, and, in the end, for putting Egan’s Woodstock case out of its misery. It’s not very gracious of Egan, having badgered the good Samaritan, Weiss, into setting up the panel he sought so desperately, putting up with Egan’s moods, and struggling through his whole sprawling haystack of a case, to denounce him for his “unhealthy interest” in Egan’s own case. But there is a sense in which Egan could be right: that his overclaimed, underproved case has finally gotten aired, judged, and found wanting, and that interest in it is no as longer healthy or urgent as it might once have seemed. Reading Egan’s ungracious posting, my sympathies went out to Larry Weiss, the kindly judge who spent so many thankless hours and years trying to get Egan his desperately-sought hearing, patiently read through his reader’s nightmare, put up with his moods, and deserves his thanks, not his wrath, for taking any interest at all in his case, healthy or otherwise. No good deed goes unpunished.
Bottom line: Over eight years Mr. Egan has stormed into, and then tried to storm out of, three rounds of negotiation to get him his hearing, the last time after agreeing to terms which he breached. In the last round, he stormed out too late to halt the hearing. He got it, lost, hasn’t paid us the thousand pounds he agreed to, and now seems bent on starting the process yet again, as if nothing had happened, having too long refused to get down to cases, and never having tried to take account of the elephant in the room—the colossal stylometric discrepancy we found between Shakespeare and Woodstock. That’s a lot of transaction costs, many times more than the thousand pounds he hasn’t paid, on top of a mountain of his own and Woodstock’s credibility problems. It’s a lot of baggage to answer for next time he needs a Charlie Brown to help him stir his pot. I think that we’re very near to the last chapter on Mr. Egan’s bets, and that, in some sense, we have the bets themselves to thank for finally getting us there.
Upping the ante on our bet.
Now let’s go back to our own bet and Mr. Egert. Should we raise the ante as he urges? He carries very little baggage and, to his credit, is actually inquiring about our live, fact-seeking bet, not the dead Egan counterbets. To his further credit, he has shown more interest than anyone we know of in actually pre-testing some plays with our software. Our best guess is that it’s a wild goose chase, and we are glad that he, not us, is the one pursuing it, but it’s only a guess. If he, or anyone else, finds an untested play out there that passes our Shakespeare tests, we would very much like to know it. So would everyone else. At a minimum, it would be a little coup, a clear exception to our 100%-accurate track record of classifications to date. It could cut it to 99% or perhaps even less, if he comes up with more than one other-authored Shakespeare match.
Or it might just be a giant surprise of a coup, raising the question whether the exceptional play was not other-authored at all, but possibly the Lost Shakespeare of Egan’s and every other red-blooded, hyper-Stratfordian Ahab’s dreams (A hyper-Stratfordian is someone who thinks that there is a lot of Lost Shakespeare out there, awaiting “discovery” by seekers more keen or bold than their precursors). If confirmed, it could win for its discoverer the same kind of fame and glory that Egan craves, and that Donald Foster actually enjoyed in the 1990s before his Elegy bubble burst. Passing our tests could be a vital piece of the confirmation process. If that were so, £1,000 would be a trivial price to pay for turning up such a Holy-Grail moon rock from a pile of plays that generations of orthodox Ahabs, themselves desperately in search of Lost Shakespeare, had passed over as presumptive slag. But we haven’t heard much sign of success in his quest, so far; Mr. Egert hasn’t accepted our bet offer, only asked to have it raised; and that still seems to us a sign that he doesn’t give circularity any greater weight than the gatekeepers of old. It’s not much.
On the other hand, perhaps we are misinterpreting an important signal on his part. Why is he asking us to raise the ante to £18,000? Surely not to shame Mr. Egan into paying us on his bet; if Egan won’t pay us the £1,000 he owes us, why on earth would he, or should he, pay us another £17,000 which he doesn’t owe us? Mr. Egert’s apparent intent is to get us to commit to paying more ducats if we lose, and to make the commitment more enforceable. Perhaps he has already found his moon rock, wants a better price for it, and would like us to supply it. Or perhaps he has realized that finding the moon rock, if it exists, will cost more time and energy than £1,000, and that he really needs more incentive to justify the effort, and would like us to supply that.
What he hasn’t explained is why we should consider a higher-stakes bet preferable to the one that has served our purposes for a decade. Would it help us get past SHAKSPER’s gatekeepers? No, that already happened a decade ago. Would it discourage frivolous takers? No. If anything, the much higher stakes would encourage a different kind of frivolous taker, the clever rent-seeker, looking not just for the satisfaction of discrediting our methods or for the possible thrill of finding a moon rock, but for ways to profit from a higher payoff, sufficient to repay the cost of litigation, depositions, affidavits, stipulations, arbitrations, contracts, escrows, forum shopping, contingency fees, settlements, or other Due Process whatnots of modern high-stakes deals, any or all to be secured at $600 per billable hour. Would that cut transaction costs? When pigs fly. It’s much more likely to raise them.
Improving enforceability is a legitimate concern. What if Mr. Egert won and we stiffed him for the £1,000, or we won and he stiffed us, as Egan has? I could think of arrangements, such as escrows, which could make stiffing more difficult, and collection easier. I bet Larry Weiss would have something worthwhile to say on this subject, and I would be willing to discuss such arrangements with Mr. Egert, if they are important to him. But let’s face it: the point of our bet was not to make money on it, but to get the gatekeepers off our backs. It worked. The point of trying to accommodate Egan on his bet was not to make money from him, but to get him off our back. That didn’t work so well for a long time because he kept making Charlie Browns of us, and we let him. But he did eventually miss a step and picked up enough baggage that we don’t have to do it anymore. I think, our Charlie Brown era is over, with Egan at least, and I hope I’m right.
What about Mr. Egert’s own stated reasons for urging us to raise the stakes: our supposed need to impress him and others with the fervor of our belief in our own methods? Not enough. How passionately we endorse our own methods says very little about whether or not they are right. We didn’t offer our bet to impress people with the fervor of our faith in our own assumptions, but to test our critics’ faith in theirs. If their faith was too faint to risk £1,000, who would expect them to be bolder about risking £18,000? £1,000 has been enough to make our point for many years, even for Mr. Egert, who still hasn’t taken us up. He hasn’t yet given us a good reason to change it.
Can any lessons be drawn from the first ten years of our bet experiment? Let’s look at some, one by one:
Is it OK to use proxy evidence in place of direct proofs? In general, yes. I have spent the last 45 years teaching college kids, smart ones, and I try to reach them at whatever level will best do the job in the time available. If you care about getting through to someone, it’s better to use simple language and imagery that he understands, even if it’s proxy, than complicated language and imagery that he doesn’t understand. See Kahneman, Thinking, Fast and Slow, 2011, Ch. 7-9. Most of us use proxies and shortcuts most of the time, when we can’t take all night or all year to make the point properly, and there is good reason for it.
Is it OK to use proxies which presuppose that the earth is flat? Not if you can help it, not even if people understand it better. But there is a permissible arguendo exception, along these lines: “If people really believed that the world were flat, then they would accept our bet that they can’t find a single instance of someone falling off the edge.” The presupposition is false, but the proxy still works to show it so.
Is some statistical evidence circular or misleading? Yes, same as with other forms of evidence.
Is all of it circular? No.
Is banning all statistics the best protection against bad statistics? Not at all, especially where the alternative looks like a flat-earth proxy. With statistics, as with other things, it is best, where you can, to use the good, skip the bad.
Did our bet cut the transactions costs of getting through SHAKSPER’s gate? Unquestionably.
Has it ended the notion that all statistics are circular? No. Mr. Egert still seems to think they are. But he does fully understand that we’ve got a bet on it, and he hasn’t yet taken it.That can’t but dim his circularity argument a bit.
Did any of Egan’s bets cut anyone’s transaction costs? For a long time, no. In the end, yes. Trying to get them settled did waste a lot of time, more, in retrospect, than Egan or his evidence really merited. But my guess is that it did help get them settled in the end.
Have we heard the last chapter of the Egan bet? I hope so, but can’t guarantee it.
Could any of these questions be decided on the merits today without the proxy? I suspect so, but realize that that, too, is a question of transaction costs. How deep is SHAKSPER’s members’ distrust for statistics? Enough, ten years ago, to rule out the possibility that better use of statistics might explain how the Oakland Athletics, with a $40 million payroll, could compete on even terms with the Yankees, with their $125 million (Moneyball, 2003)? Enough, today, to dismiss Nate Silver’s calling all 50 states correctly in the 2012 elections by clever extrapolation of tiny samples (The Signal and the Noise, 2012)? Or to rule out the latest stylometric evidence that J.K. Rowling wrote The Cuckoo’s Calling? What about the Shakespeare Clinic’s own experience with the last four expansions of its archive of tested plays? Every time we added new texts, including the three dozen or so plays we have tested since we first offered the bet, we could see that the same Shakespeare patterns we found in the old rounds continued to be found in the new rounds. See our “Oxford by the Numbers,” 2004, 367.
If the circularity theory is correct, none of these should have happened. Neither Silver’s limited voter samples nor our limited play samples could be expected to tell us anything about 99.9% of voters who didn’t get interviewed or the 75% of plays which hadn’t gotten tested. Yet they did for Silver and they have for us. If facts matter, all of these extrapolations should put the “statistics-are-circular” theory into the same drawer as the flat-earth theory, and you shouldn’t have to bet on it to convince people that it is not so. If facts don’t matter, then the circular theory lives on in the mouths and souls of today’s gatekeepers, and it still takes a bet to demonstrate that it is dead in their own hearts and pockets. I suppose that’s better than nothing. But proxies do have their own problems, and I can’t help wondering whether stylometric facts are really as impenetrable as most people think. If you care about authorship, you should not rule out stylometry. Try a little the next time you find yourself feeding on proxy evidence, and see if its transaction costs are not lower than people think, and its benefits higher. You might find that you like it.
Does authorship matter? For sure.
What would have happened without the bets? Most people deal with challenges on one of two ways: sensitive and responsive or not-so-sensitive and not so responsive. There are good arguments for both. Under the sensitive scenario, we would have felt duty-bound to respond challenges seriously, even flat-earth ones, accept whatever transaction costs they entailed, and respond to every point seriously and politely. No one could do this with every flat-earth bright idea on offer; there are far too many of them, some dismissible out of hand, many not. But now and then one pops up in your back yard or blocking your path, and you have to decide between responding and ignoring. We responded for a while with our circularists, found ourselves getting nowhere, and used the bet to get unstalled. Absent our bet, we might well have tired of the hassle, given up on SHAKSPER, never had to bother with postings like this one, and never done our bet experiment, Egan’s bet experiment, or our Golden Ear experiment. SHAKSPER was indispensible for all of these conversations, and deserves a lot of people’s thanks, ours especially.
What would we have done without the Egan bet? Saved ourselves years of getting Lucied. Our experience with Egan commends the less-sensitive scenario. It is better suited for assertive, flat-earth challenges like Egan’s, which are not so serious, polite, or heedful of evidentiary niceties. Would we or anyone have paid any attention at all to Egan’s book, absent his direct challenge? I doubt it. Even with the direct challenge, we might have responded less seriously or less politely, or not at all, and done so earlier. It would have made sense in terms of transaction costs, but the not-so-sensitive approach hasn’t been our normal first resort. Normally, we have tried to wait till after someone’s third, fourth, or fifth rude, frivolous challenge before telling him off or shrugging him off (Is anyone surprised, by the way, that it has always been a him, not a her?).
I sometimes think that we spent more time than we should have in the responsive phase, with Egan and with other obstreperous gatekeepers. My wife has felt that way for a long time, Valenza too. He has joined in most of my defenses, but his own preferred response to rude challenges is to ignore them. Maybe there’s a lesson there. Or maybe there are some challenges which have to be addressed if you want to stay in the game. Either way, a well-conceived bet can be a better response than a shrugoff for unjamming deadlocks, cutting transaction costs, and leading the conversation in more fruitful directions, especially if you happen to be an outsider carrying the burden of proof. I think we did better with our own bet than we would have done without it, and it’s possible that, in the end, the same could be said of our taking Egan’s bet seriously enough to respond to it seriously and politely, again and again.
What would Egan have done without his bet? That is more his to say than ours. I suspect that he would have found some other ways to stir the pot for his Woodstock ascription, perhaps in his journal, which I gather is devoting a special issue to the topic. I also suspect that the outcome would have been essentially the same as with the bet, but years sooner: his book unread and no one of consequence persuaded. With the bet, we know for certain that at least four very patient people have read Egan’s evidence who would not have done so otherwise, and none were persuaded. Most people on SHAKSPER now know it, too. “Unhealthy interest,” as Egan charges? It seems odd, coming from him. But perhaps it was better than ignoring him.
Has Joe Egert found the Lost Shakespeare? I don’t know. He certainly hasn’t claimed so, but he is showing some of those tell-tale signs of finding a way to turn it to his greater advantage. If he has found it, I and many others would like to know. But not enough to offer him £18,000 just to make him try harder, far less as a kind of Mom tattoo to show our fervid devotion to our own methods—and least of all as a kind of prize to invite the other contentious hyperstratfordians of the world to our doorstep, demanding that we drop everything and attend to their moon-rock claims. Who, after eight years of Egan, would want that? Let’s face it. If our methods are wrong, raising our ante is not going to make them right. If things are slow in this summer, I don’t see a bigger offer from us livening them up much, least of all in a way that would minimize frivolous, contentious distractions for us. That said, I wish Joe well in finding the Lost Shakespeare, and I’m still glad that he is the Ahab on this one, not us.
Is this posting too short? Maybe so, for some. We’ve said a lot of it before. There’s plenty more out there, if you don’t mind reading it. Start with http://shaksper.net/archive/2011/303-september/28127-thomas-woodstock-
Is this posting too long? Maybe so, for others. It’s one of our longer ones. My apologies to SHAKSPERians for imposing on their patience with a fourteen-minute exploration of the place of bets in SHAKSPER arguments. There is one. Nothing is either good or bad, save alternatives make it so. The exploration is optional, and could be cut to four minutes by going straight to the “Lessons” section. It’s shorter and easier to understand in one big concentrated lump than it would be in ten little spread-out lumps. It is much, much shorter than what we and SHAKSPER were up against in our wrangle with the gatekeepers, pre-bet. And it is infinitesimal compared to what the Weiss panel, and we, had to take on to get the case of Lucy v. Charlie Brown finally heard, decided, and over with. I hope SHAKSPERians will learn something from it. We have.
Do transaction costs matter? Much more than most people realize.
Can bets lower them? Our fact-bet, yes. Egan’s opinion-bets, in the end, eventually, though it took six years, and many more transaction costs than it deserved. Egan’s repeated demands, first to get a verdict on Woodstock, and later to repudiate it, are reminders that bets can carry their own seeds of contention. His bets have led to new frontiers of confrontation and contention. On the other hand, I suspect that they have also helped ring down the final curtain on Egan’s Woodstock drama. I am grateful to Larry Weiss, and to his panel, of himself, Will Sharpe, and Dale Johnson, for taking on thousands of murky pages of transaction costs and for ruling, correctly, that Mr. Egan failed to make his case or live up to his agreements. I am sorry, though I can’t really say I am surprised, to see Mr. Egan berating them for taking an “unhealthy interest” in his cause, after all his years of pleading and prodding to get someone to read his book, all of it, and pass judgment on it. They did both, and they did it well. All SHAKSPERians, including Hardy, who has made debates like ours possible, are in their debt for making such debates other than interminable.
Ward E.Y. Elliott
Burnet C. Wohlford Professor of American Political Institutions
[Editor’s Note: There is an excellent essay by Mac Jackson discussing among other things the Ward/Valenza methods and findings in Shakespeare Beyond Doubt. –Hardy]