Volume III, Issue 2 -- May, 2002
Reactions to Richard Posner's
Another Study in Decline
page 1 | page 3
III. The Question of Decline:Richard Posner's Public Intellectuals has a subtitle: A Study of Decline. Curiously, then, decline is not a matter of concern to a number of the reviewers -- ten largely or wholly ignore it, never really questioning what exactly Posner means by it.
Posner's subtitle is, in fact, a problematic one. Public Intellectuals is not primarily a comparative study, nor a historical one. It does not trace a path of decline; instead it posits decline without much of a frame of reference. True, PIs from times gone by are introduced, the world of yore brought up -- but there is no systematic comparison of PIs across the ages.
Gertrude Himmelfarb's point is well taken:The subtitle, "A Study of Decline," points to a more serious difficulty, for we are never told what these public intellectuals are in decline from. When and where were public intellectuals not in decline ? And who might those admirable people be ?Carol Polsgrove, in The American Prospect similarly notes: "But the 'Decline' in the subtitle is in fact hypothetical -- pure tease. Posner makes no attempt to trace a decline; this book is not a history."
Thomas Nagel is more succinct: "I don't think the book supports the subtitle, A study of decline".
At the very least, Posner's choice for this prominent description of his book warrants mention in the reviews and commentaries -- if only to dismiss it. Even a toss-away acknowledgement -- such as Kenan Malik, writing in the New Statesman "Public Intellectuals is not so much an analysis of decline as an expression of it" -- is at least something. But unaccountably, many reviewers simply entirely ignored the issue.
- Return to top of the page -
IV. The ListsPerhaps the reviewers and commentators didn't worry much about definitions of terms such as "public intellectual", or about the book's subtitle because they wanted to get to the juicy part of the book: the lists. It is these that seem to have captured the popular (and much of the so-called critical) imagination. No doubt: these are seen as the highlight (or lowlight or, in Adam Liptak's words, "the goofy centerpiece") of the book.
The lists are among the twelve tables Posner provides (they are numbered 5.1 through 5.10 -- but there is a 5.8a to go along with 5.8, and a 5.9a to go along with 5.9). They are presented at the end of chapter five, smack in the middle of the book. The center of the book -- and the heart of it, most seem to believe. Certainly, they have been the center of almost all the attention.
Some of the tables involve regression analysis and one even provides circulation data for selected magazines (Table 5.6), but only three really caught the public and critical eye:
Almost everyone harped on these lists -- though in shows of remarkable reviewing bravado Peter McKenna, reviewing the book in the Christian Science Monitor, and Kenan Malik (in the New Statesman) ignore the tables completely, and Robert Boynton, in The Washington Post almost does as well.
- "Table 5.1: Public intellectuals: media mentions, Web hits, and scholarly citations (1995-2000)" -- this is the infamous table of 546 PIs selected by Posner
- "Table 5.3: Top 100 public intellectuals by media mentions" -- also for the span 1995-2000, this is the infamous list topped by Henry Kissinger
- "Table 5.4: Top 100 public intellectuals by scholarly citations (1995-2000)"
Table 5.1 -- the initial selection of names of PIs Posner made to work with -- is perhaps the most controversial one. Here Posner winnowed down the list of all potential PIs to a manageable 546. He explains his criteria -- though it is not a clear-cut process. There was no simple test he used to determine whether or not a person qualified as PI, though there was at least some method to his madness. His definition of PIs certainly plays a role (though, as we have seen, many reviewers didn't concern themselves greatly with what exactly his definition was). He also tries to explain why certain people made the list and others didn't -- though his many explanations may have sown more confusion than cleared things up.
Posner uses Table 5.1 as the basis for his "statistical study of public intellectuals" (p.7). He relies on these names in almost all the additional tables (where most of the statistical work is done). In Table 5.2 he offers "Public-intellectual summary statistics" based on 5.1 (looking at the demographics of the group, and their fields). Later tables also use the names from 5.1 -- for example: the one hundred PIs from 5.1 with the most media mentions are listed in 5.3 and the one hundred with the most scholarly citations are listed in 5.4. Other tables then use these lists (and also compare them to the larger list in Table 5.1) as Posner does his regression analysis. (Only 5.6 -- offering circulation data for the principal print venues for PI expression -- is independent of 5.1.)
In considering Table 5.1 it is important to bear in mind that it is Posner's list, using his definition and criteria. One does not have to agree with his definition of "PIs", but that is the definition the list must be judged by. In addition, one must remember what Posner himself says about his choices:
Many reviewers don't bother much with Posner's criteria and accept the list without much comment. Despite Posner providing reasons why certain people are not included, these reasons are not always repeated in the reviews (see for example Adam Liptak in The New York Observer).
- He admits: "The list in Table 5.1 is not exhaustive" (p.169)
- He writes of "the nonrandom and nonrepresentative character of the sample of public intellectuals in Table 5.1" (p. 170)
- He writes: "I acknowledge the arbitrariness of many of my decisions on whom to classify as a public intellectual." (p. 172)
Several of the reviewers also willfully (or carelessly) ignore Posner's statements about the list. Gertrude Himmelfarb, for example, writes that Table 5.1 is "a presumably comprehensive roster", despite Posner protestations that it is not.
Christopher Caldwell writes: "His choices as to who qualifies are also dubious" and notes: "The closer one looks at the list, the more arbitrary it appears." Caldwell lists quite a few people he is surprised to find missing -- including numerous foreign PIs -- but doesn't look closely at Posner's criteria. Posner's criteria for inclusion and his definition(s) of PIs are not always particularly clear, but do offer an at least adequate justification for most of his choices (including, for example, the absence of many foreign PIs).
Some accept Posner's acknowledgment of arbitrariness, and yet insist that isn't good enough. Eric Alterman, in The Nation, writes: "Admitting that the construction of any such list is necessarily a subjective enterprise, Posner has nonetheless proven himself to be a profoundly deficient craftsman." Alterman gives several names of PIs he believes should be "not merely on a list of 546 but on any competent grouping one-tenth its size". (He gives seven names, which would yield more than 12 percent of a list one-tenth the size of Posner's -- one wonders who he would leave out in filling the remaining places.)
Alan Wolfe insists: "'I acknowledge the arbitrariness of many of my decisions on whom to classify as a public intellectual,' Posner writes. But his decisions are not merely arbitrary; they are nonsensical." And Carol Polsgrove writes: "Posner built this list out of one compiled by another author in 1970, adding names, apparently, as they occurred to him. It is impossible to find any basis for Posner's inclusion of certain names on the list and not others."
Some of the criticism of the list is valid, and some of the names left off striking (as are a few that do make the list). Posner himself notes that after completing this study he expanded his list "from 546 to 607, by adding names that occurred to or were suggested to me" (p.193). (The complete list can be found at Posner's faculty page.) There is no doubt that there were notable omissions (and dubious inclusions), even accepting Posner's definition of PIs and all the criteria he sets forth. But is it a fatal weakness, as several of the reviewers claim ?
It is important to remember the purpose of Table 5.1. It is meant to be a sample of PIs, not a list of all PIs. Posner acknowledges that it is neither complete nor random. For Posner's purposes the list seems perfectly adequate -- and it is Posner's purposes that must be the test.
Posner claims little more than, for example, that 5.1 serves as an adequate sample for the more specific Table 5.3 (top 100 PIs by media mention), leading to a list of one hundred that:(...) probably includes most public intellectuals who enjoy prominence today in the United States, and is at least representative of them.Admittedly, this isn't quite as exciting as having a definitive list of the 'Top 100 PIs' (or top 546) (not just by media mentions or scholarly citations, etc.). But making such a list was definitely not Posner's purpose: indeed, throughout the book he makes a point of emphasizing how shoddy the work of many of these "top" (or rather top-ranked, according to certain criteria) PIs is.
One can also criticize what Posner then actually does do with his representative samples (see below), but the basis for them should also be clearly understood -- yet essentially none of the reviewers make Posner's intention regarding 5.1 clear.
Table 5.3 has turned out to be the most misconstrued one. Based on the 546 PIs Posner selected for 5.1, Table 5.3 lists the "Top 100 public intellectuals by media mention" for the period July/August 1995 to July/August 2000. (Posner also explains in some detail what he means by "media mentions", and how he obtained these results (relying on three Lexis/Nexis databases).)
It is this list which prompted headlines such as: Kissinger rated No 1 brain and How I unwittingly made Kissinger the No1 intellectual in America (both The Guardian, 21/1/2002 and 23/1/2002). Kissinger was the top-ranked of Posner's PIs on Table 5.3 -- but it is something of a stretch to label him the top PI based simply on that.
Duncan Campbell's article is typical of popular press reaction. He calls it: "A list of the world's supposed top 100 public intellectuals", not mentioning any of the other lists (neither Posner's original Table 5.1 of 546 PIs, nor 5.4 (listing the PIs in scholarly citations-order)). At least he does vaguely suggest what "top" here means in describing how the list was put together (though ignoring that it is based solely on Posner original list of 546), writing it was done "using the internet to count the number of media mentions of anyone who expressed themselves on matters of general public concern between 1995 and 2000".
William Grimes' article in The New York Times (19 January 2002) reproduces Table 5.3 (along with five photographs of listed PIs). The heading and subheading above the table there read:A Name-Dropping Hall of Fame for Great MindsThis was presumably the popular account of the list that received the greatest attention and largest audience. Unfortunately, the description of the table is incorrect: this table (5.3) did not use scholarly citations in "measuring" the PIs. 5.3 only looks at media mentions (and 5.4 only looks at scholarly citations). A list of the top 100 according to the criteria claimed by The New York Times would have led to a completely different ranking.
The top 100 public intellectuals measured by scholarly citations and media mentions, according to Richard A. Posner, judge, scholar and author.
In the article itself Grimes is more accurate, writing of "the already notorious Table 5.3, a list of the top 100 public intellectuals, ranked by media mentions." Still, he fails to note that the top-100 list is based on the 546 PIs from 5.1, writing only that Posner "fed a large assortment of names into his computer". (He does, however, later discuss some of the "oddities" of 5.1 as well.)
Christopher Hitchens' reaction also muddles facts. He refers to the chart as being of "the chosen 100", making no mention of the only chosen ones -- the original 546. Hitchens does then mention that: "The ranking was done by his own computation of media mentions !" -- something apparently very surprising in a list titled: "Top 100 PIs by media mentions" !
Hitchens also comes to a conclusion that most seem to have drawn from 5.3: "That's right: the more references to you in media sources, the more your stock as an intellectual goes up." Posner's "top" is perceived as meaning "best", but that is not correct.
Posner goes to great pains to emphasize that that is not what he means at all. "Top" on these lists is not a value judgment, it is merely an indication of -- in the case of 5.3, media mentions, in the case of 5.4, scholarly citations.
(Matters are a bit complicated by the stock-comparison, as, in a sense, one's "stock" as a PI might indeed rise with mere media mentions (rather than by producing high quality stuff).)
Reviewers might have been expected to be more attentive to the details in discussing 5.1, but many weren't.
Christopher Caldwell boils it all down to:Posner assembles his profile by feeding a list of intellectuals' names into Google and Nexis, and tallying up their mentions in media, websites, and scholarly literature between 1995 and 2000. The result is a league-table of the "top" thinkers of our timeCaldwell gives an example from this "league-table" which makes it clear that he is referring to 5.3 -- but his description (of what went into that list, and what it means) is completely misleading.
Fred Inglis, in The Independent, offers the most impressive example of misleading readers with his ridiculous spin on the tables:He counts internet hits and media mentions for an "arbitrary" but very long list (546) of names. The names themselves will have you clutching weakly at the table for support (Newt Gingrich), and the winner will have you reaching for the rotten tomatoes (Henry Kissinger).The above is a misstatement beyond the pale: the only mention of Newt Gingrich in the book is on page 171, where Posner specifically explains why he excluded him from his list of 546 (i.e. no matter how weakly you clutch the table, you won't find Newt there). And to call Henry Kissinger the "winner" is a complete misrepresentation of all the information that Posner has presented (not just 5.3, a list which Kissinger does top). Certainly, Internet hits did not count towards determining prominence (in fact, Kissinger doesn't rank particularly high in the category (for which Posner does not make a separate list)). Inglis also forgets to mention that Posner also counted scholarly citations for his list of 546, and then for 5.4.
David Brooks, the The New York Times Book Review describes 5.3 as: "a chart of the top 100 public intellectuals for the years 1995-2000" -- only parenthetically noting (in the next paragraph) that "the rankings are based on media mentions, by the way". Without describing how Posner arrived at it (he ignores Table 5.1 completely), he nevertheless feels comfortable in summing up Henry Kissinger's status as making him "National Champion Public Intellectual".
Others also oversimplify or incorrectly describe 5.3: For example, Gary Rosen, in the Wall Street Journal, unhelpfully notes:Especially unhelpful is his (already much-talked-about) "top 100" list. Basing his pecking order on nothing more discriminating than the number of times an individual's name turns up on Lexis/Nexis (...)
Not all the reviewers made a hash of 5.3: many were more judicious in their description of this table and these rankings. Overall, however, they did a very poor job of conveying to readers what exactly Posner's list was, and what significance (or lack thereof) could be attributed to these rankings.
The table got blown out of all proportions -- and few made much effort to clear up any of the confusion. If the public wanted to accept it as a "top-100 list" of PIs, most apparently decided they weren't going to stop the fun.
- Return to top of the page -
V. Improving the Performance of the MarketIn Public Intellectuals, Posner posits PI-market failures. He provides a variety of evidence -- statistical and anecdotal -- to support his claims and arguments, and then in a brief (eleven page) conclusion suggests some possible remedies. Posner does not go in great depth, and only offers a few suggestions. He also discusses only "cheap methods", believing the PI-market does not warrant "costly methods of correction" (such as government regulation).
The book discusses what Posner believes to be wrong with the PI-market, and closes with suggestions as to what might be done to help improve it. The suggestions would appear to be an integral and important part of the book -- yet many reviewers saw fit to completely ignore them and make no mention of them whatsoever in their reviews.
Caleb Crain, writing in The Nation, might be right that: "even if Posner's suggestions were adopted, they would change nothing". At least Crain mentions the existence of Posner's suggestions (and even discusses some of what he calls these "extremely modest proposals for reform"). A few others also expound on them -- notably Alan Wolfe in The New Republic and Carol Polsgrove in The American Prospect. Too many others -- inexcusably, surely -- didn't even bother to inform readers of the existence of any such proposals (or their substance).
Posner makes several proposals, and some of the reviewers also only selected some or one from among these to comment on. The most popular one was, as The Economist put it, that PIs should "post their columns on university websites (where colleagues will supposedly spot and correct their errors)" -- a proposal that Thomas Nagel in the TLS believes "shows, I am afraid, that his own grasp of reality is sadly defective". (Neither of these two reviews addressed any of the other market-improving suggestions Posner made.)
One market-improving norm that Posner hopes will emerge in the PI-arena is:a norm against magazines' commissioning or accepting book reviews written by persons criticized in the book to be reviewed, at least without full disclosure in the review that the reviewer was criticized in the book.Fully five of the reviewers considered here have such potential conflict-of-issue problems. The five -- Himmelfarb, Jacoby, Wolfe, Brooks, and Nagel -- all made Posner's 5.1 list (of 546 PIs), and Brooks and Wolfe both made the 5.3 list (100 top PIs according to media mentions) -- ranking 85th and 98th, respectively. Each of the five also finds at least one additional mention in the book proper (though Brooks' is nothing more than a citation in a footnote). Wolfe and Nagel are cited as examples of PIs who have erred in their ways, while Himmelfarb is presented as a prime example of the ills Posner writes of, the errors of her ways discussed at considerable length.
All five offer some sort of disclosure -- probably close enough to the norm Posner hopes for:
Overall, this is one area in which the reviewers seem to have done an adequate job: noting the issue, and addressing it to the extent necessary (by at least making readers (vaguely) aware of possible conflicts of interest). However, the reviewers hardly consider the significance of Posner's mentions of them, and whether this might not have an effect on their reviews, leaving such concerns entirely to the reader. (Himmelfarb comes closest to considering the implications of her work being savaged in the book.)
- Russell Jacoby offers a parenthetical: "(Disclosure alert: Posner credits me, in my The Last Intellectuals, with coining the term, and I marginally figure in his book.)" -- but does not mention that Posner counts him as one of his 546 PIs
- David Brooks begins his review: "I'm not sure I'm qualified to review this book. Starting on Page 209, there is a chart of the top 100 public intellectuals for the years 1995-2000. I'm all the way down at No. 85" -- as noted above, this is a misleading description of 5.3 (making a bit more of Brooks than Posner does)
- Alan Wolfe begins his review: "Any method for ranking the top one hundred public intellectuals in America cannot be all bad if it includes my name among them. Opening Richard A. Posner's book to the chart that most of his readers will immediately consult, I find myself barely making it at number ninety-eight" -- again a misleading description of 5.3 (though he does clear that up a bit later in his review). He also notes (when discussing Posner review-norm recommendation), parenthetically: "(Posner makes a number of references to me in his book, some of them supportive, some of them critical, so I am not sure whether, under his proposal, I would be required to disclose what I just disclosed.)"
- Thomas Nagel writes about Posner's definition of PIs: "To my distress, it includes me. I would reject the label, but apparently writing book reviews for the general reader is enough to qualify, not to mention an isolated venture into political polemic -- a protest in the TLS against the impeachment of President Clinton, for which Posner rebukes me sternly. "
- And, in the most comprehensive disclosure statement, Gertrude Himmelfarb begins her review: "Declaration of interest: I am one of the many public intellectuals criticized in Richard A. Posner's latest book, Public Intellectuals: A Study of Decline. Indeed, I apparently have the distinction of having inspired it." She notes also that among the other PIs disparaged by Posner are her husband and son (impressively, mom, dad, and son William all are among the 546 PIs of Table 5.1 -- though only William makes 5.3 (and none make 5.4)). She also acknowledges about her review that: "This is a critical essay as well as a statement of full disclosure", and directly addresses Posner's comments about her work.
One should note that this is a norm that is widely (if not completely) adhered to in any case -- at least as far as making readers aware of potential conflicts of interest between reviewers and the reviewed. (Himmelfarb also notes: "If I shared Posner's utilitarian cast of mind, I might suspect that his disparagement of so many public intellectuals in this book (...) was intended to diminish the supply of critical reviewers.")
Despite Posner's hopes, publications will, however, certainly continue to commission reviews from those who are criticized in the books to be reviewed. For one thing, it makes for fun, antagonistic reviews: Himmelfarb's, for example, is among the most entertaining responses to Posner's book (and also one of the most comprehensive -- though it is certainly also not without its faults). But Posner does have a point: something is lost in the antagonism, and the focus of the reviewer tends to be more on justification than assessment. The reviewers who have conflict of interest issues re. Posner's book (specifically Himmelfarb, Wolfe, and Nagel) do a good enough job, by and large, in their reviews, but the personal colouring does not seem to serve either them or readers who want to learn about Posner's book ideally.
Overall, the selective and limited reactions to Posner's recommendations are disappointing. Those that did comment were generally very dismissive -- which seems fair enough. Unfortunately, many commentators ignored them completely, not even making readers aware that Posner did offer such recommendations.
- Return to top of the page -
VI. ErrataMany commentators found fault with Posner's list of PIs, his methodology, his reasoning, his conclusions -- even his hypotheses. Curiously, then, few found fault where he was most clearly vulnerable: actual mistakes.
Most prominent are the misspellings of certain PIs' names in Table 5.1. William Grimes, in his article in The New York Times points out two: André Malraux and Ivan Illich (given by Posner as "Andre Malreaux" and "Ivan Ilych"). Carol Polsgrove notes poet Allen Ginsberg's name is misspelt ("Ginsburg") -- and that Garry Wills' name, while correctly spelled on Tables 5.1 and 5.3, is misspelt in the text proper (as it is in the index), as "Gary".
Carlin Romano's jokey commentary in the Chronicle of Higher Education notes a fill of misspellings:Andre Malraux as "Andre Malreaux," Zbigniew Brzezinski as "Zbigniew Brzezinsky," Lewis Lapham as "Lews Lapham," Walter Lippmann as "Walter Lippman," Garry Wills as "Gary Wills"But these were the only commentators surveyed that pointed out the incorrectly spelled names -- and note that only one of the three was a reviewer.
(To the list of mistakes we can add: Frantz Fanon (given as "Franz Fanon"). Also: other Andrés -- Maurois and Gide -- are among several foreigners presented accentless -- though Céline and others get theirs.)
Spelling mistakes -- and printer's slips -- will happen, and reviewers usually don't harp on them. In this case, however, the mistakes are relevant. If the names were misspelt consistently (i.e. during all the web searches, and searches for media and scholarly citations) then they can dramatically skew the results. Seven of the names are clearly spelt wrong -- Malraux, Illich, Ginsberg, Brzezinski, Lapham, Lippmann, and Fanon -- (Garry Wills name is only misspelt in the text, not on Table 5.1 and 5.3). In addition, there are numerous names with a missing accent (Gide, Maurois, Ortega y Gasset, and others). This obviously affects Posner's findings. Consider:
In our review we pointed out:In table 5.1 "Andre Malreaux" is listed as having received 26 Web hits (great research assistants Posner has -- the fact that "Andre Maurois" got 40 times as many hits didn't tip someone off that maybe there was something wrong here ?). On 7 February 2002 we ran the name-variations through Google, getting:These numbers are the raw data (as Posner explains, he used a bit more refined technique to arrive at his totals-- cf. below and his p. 188), but they still are devastating. (Posner only got 26 hits for "Andre Malreaux"; presumably duplicates, mirror sites, and other irrelevant sites were not counted.)
- for "Andre Malreaux": 73 hits
- for "Andre Malraux": 5930 hits
- for "André Malraux": 25,100 hits (a top-tier result compared to the Web hit totals others in table 5.1 achieved)
Posner explains that:Each name was first searched for as it appears in Table 5.1 (.....) The first thirty hits were examined to determine the percentage of correct ones. The percentage was then applied to the total number of hits. (p.188)Which makes one wonder what he and his research assistants were looking at when they looked at the first thirty hits for "Andre Malreaux". None of those could be "correct" ..... (It is remarkable that no one involved in this endeavour noticed that "Andre Maurois" -- just a few names down from Malraux (or rather "Malreaux") -- got 1,045 web-hits, while "Malreaux" only got 26, a disparity that must shake anyone who has even the vaguest idea of who these two guys were into thinking that something went very, very wrong here.)
Even if many of the hits we got when we tried the correct name would not count under Posner's system, it is clear that, as far as André Malraux goes, Posner's web-hit tally is off by a factor of over 10² (and possibly close to 10³). Presumably the media and scholarly citation counts are wrong by a similar order of magnitude. It is also noteworthy that merely adding the correct accent made a huge difference in the number of hits received (as noted, Posner skips a considerable number of accents).
(There can be no question -- at least in Malraux's case -- that this is merely a printer's slip: the wrong spelling was clearly used to obtain these ridiculous results. Remarkably, two of the PIs -- Brzezinski and Ginsberg -- make it onto Table 5.3 despite their names being misspelt; possibly these are only typographical slips and the correct spellings were used in obtaining the data.)
Presumably, at least some of the data for the other misspelt names is also off (though probably not quite as much, as the misspellings are far more common ones). Surely this calls some of what Posner is doing here (and the care with which he undertook the study) into serious question. But only one reviewer even mentioned the mistakes.
Posner isn't taken to task about other slips either. Gertrude Himmelfarb objects to Posner's ideological categorization of some of the periodicals he lists, and expresses doubt about the circulation totals given for The Public Interest, leading her to conclude: "It does not inspire confidence in Posner's other tables that this one, whose facts and figures are readily available, is so faulty." (Note that she doesn't make mention of the spelling mistakes in Table 5.1, which would certainly bolster her contention.) But most of the other nit-picking is about questions of interpretation, rather than facts.
With his careless spelling mistakes Posner handed his critics a golden opportunity to attack his text. Yet most of them seem to have completely overlooked them. This may be forgivable in those who ignored 5.1, but anyone who harped on that list, Posner's choices of PIs, and what he did with his data should have noticed at least some of these slips. (Alan Wolfe even mentions that Allen Ginsberg made 5.1 -- but makes no mention of the misspelling ! And in William Grimes' article in The New York Times Table 5.3 is reproduced in its entirety -- with the names of Brzezinski and Ginsberg correctly spelled.)
page 1 | page 3
- Return to top of the page -
Current Issue | Archive | about the crQuarterly | the complete review
to e-mail us:
© 2002 the complete review Quarterly
© 2002 the complete review