explaining the long tale

The long tail has been extensively discussed.  But what about the long tale?  What is the nature and significance of the long tale?

Consider two very long tales. The longest tale printed with a Latin or Cyrillic alphabet is Madeleine de Scudéry’s Artamène, ou le Grand Cyrus.  This type of work is known among specialists as a “roman de longue haleine” (long-winded novel). First published in Paris, 1649-1653, Artamène consists of ten volumes encompassing 7,443 pages and about 2.1 million words. A second long tale is Samuel Richardson’s Clarissa, or, the History of a Young Lady, published in London in 1748. Its first edition has seven volumes with a total of 2564 pages and about a million words.[1]

Reading a long tale takes a long time. At current, typical prose reading speed, Artamène would take about 140 hours to read. But in the seventeenth century, books were often read aloud. Reading aloud takes roughly 70% more time than silent reading.[2] Moreover, if reading occurred by candlelight, the need to maintain and trim the candle plausibly might increase reading time by 5%. So reading Artamène could easily have required 250 hours of reading time. Reading Clarissa could easily have required 100 hours of reading time.

While they were long tales, Artamène and Clarissa were also best-sellers of their times. One scholar declared of Artamène:

from 1649 to 1654, from one end of France to the other, at the court and in the most aristocratic circles, as well as among the more cultivated bourgeoisie, at Paris and in the provinces, in all ranks of a society the most polite in the world, one read them not merely with pleasure, one seized upon, one devoured bit by bit as they appeared, every one of those ten great volumes.[3]

In the course of printing, the printer increased the print run for currently printing volumes and printed additional copies of earlier volumes. While printing of the first edition finished in 1653, by 1655 the printer had already produced a complete fourth edition and a printer in England had already printed an English translation of the full, ten-volume work.[4] From 1654 to 1660, Scudéry produced another ten-volume work Clélie, Histoire Romaine. That action testifies to the success of Artamène. Clélie turned out also to be highly popular.[5]

The success of Clarissa can be described more quantitatively. Richardson probably printed 3000 sets of the seven-volume, first edition of Clarissa in 1748.  He printed additional editions of Clarissa in 1749, 1751, and 1759.  These later editions probably amounted in total to about 3000 sets.[6] Through 1769, a total of eleven editions of Clarissa were printed in London and Dublin.  For comparison, few editions of British novels between 1750 and 1770 had print runs greater than 1,000, and most probably were printed in 500-800 copies.[7]

Just as did Scudéry, Richardson quickly followed one long tale with another.  About five years after writing and publishing Clarissa, Richardson wrote and published a new work, The History of Sir Charles Grandison (1753). Its first edition has seven volumes comprising 2459 text pages and about 907,000 words.[8] Three editions comprising 6,500 sets were printed within a year.[9] Both the size and print runs of Grandison suggest the prior success of Clarissa.

The commercial success of Clarissa measures reasonable well against that of Richardson’s path-breaking best-seller, Pamela, or Virtue Rewarded (1740). Pamela probably sold 20,000 two-volume sets within fourteen months after it was first published and had fourteen editions through 1769.[10]  Total volumes sold of Pamela through 1769 probably did not exceed by more than 50% those sold of Clarissa. Moreover, in 1766, the copyright of Pamela sold for £288, while the copyright for Clarissa sold for £600.[11] Both Pamela and Clarissa were best-sellers in America. Pamela, published in the U.S. in 1744, sold more than 10,000 copies through 1749. Clarissa, published in the U.S. in 1786, sold more than 25,000 copies through 1789.[12]

Long tales published in the twentieth century differ significantly from Artamène and Clarissa. Marcel Proust’s À la recherche du temps perdu has nine volumes totaling about 3,200 pages and 1.5 million words.[13] Despite considerable advances in writing and printing technology, Proust’s work was published over a fifteen-year period (1913-1927), while Scudéry’s Artamène, about 50% longer, was published over only a five-year period (1649-1653). Moreover, Artamène was a best-seller, while À la recherche du temps perdu was far from a best-seller. The first volume of Proust’s work had an initial print run of 1,750 copies, and perhaps 4,100 copies were printed between 1913 and 1918.[14] A best-seller in the U.S. about this time would sell 900,000 copies to a population about twice the size of France’s.[15] Other long tales of the twentieth century attracted even less popular attention than Proust’s work.

The closest the past century has come to producing a best-selling long tale is J.K. Rowling’s Harry Potter series.  The seven Harry Potter books, published from 1997 to 2007, have a total of 4175 pages and about a million words.[16]  The final book in the series broke sales records by selling 2.7 million copies in the U.K. and 8.3 million copies in the U.S. in its first 24-hours on sale.  The Harry Potter series as a whole differs from a long tale in that its volumes were marketed as single works and not widely sold as a set.  Moreover, the Harry Potter series was published over a period more than twice as long as that for Artamène and Clarissa. Rowling shows no signs of adopting the form of the Harry Potter series as a template for another work.  Instead, Rowling plans to take time off and then write an encyclopedia of Harry Potter characters and places. The Harry Potter series has not re-established the long tale as a generic type of work.

Vertical integration favors the production of the long tale. The longer the tale, the greater the cost and the risk in printing it.  Richardson was not only an author; he was also a master printer who printed his own works.  Thus he did not have to pay another printer for the cost and risk of printing a long tale.  The more imperfect the market for printing and risk-bearing, the greater the advantage to being able to assume both these functions within an author-printer enterprise.  Richardson produced Clarissa and Grandison with the advantage of vertical integration at a time when transaction costs associated with the nascent novel-printing business were relatively high.

Social influence favors the success of the long tale. Recent research indicates that greater social influence favors greater concentration of demand among the most highly popular works.[17]  Salons and coffee houses were important social institutions in seventeenth-century France and eighteenth-century England. Scudéry herself conducted at her Paris home an important salon known as Samedi:

the main purpose of the salon was for amusement. Among the activities were excursions, elegant dinners, and surprise visits to friends staying in the country. The glory of a certain pastry shop in rue Saint-Honoré that Mlle de Scudéry and her friends loved to frequent has come down to us and we also know of Mme Aragonais’ dolls, which the ladies of the Samedi dressed in the current mode. Other diversions were the experiments done by Claude Perrault, architect and anatomist, to observe the chameleon’s ability to change to change color according to its environment. … Poems were exchanged, of course, as were certain gallantries…[18]

The vibrant salon world of seventeenth-century France created extensive, powerful channels for social influence. Social influence arising from these salons, and from Scudéry’s position as a leading salonnière, are probably an important part of the explanation for the long tail.

starting to read Richardson's Clarissa

The communication industry has changed greatly since the time of Artamène and Clarissa. The average duration of online videos watched in the U.S. in March, 2008, was only 2.8 minutes per video.  That’s much, much less than the 250 hours it probably took to read the best-seller Artamène in seventeenth-century France. Less vertical integration on the supply side and less social influence on the demand side may be an important part of the explanation for this huge difference.

*  *  *  *  *


[1] Artamène is available online. While the authorship of the work is not obvious, most scholars believe that Madeleine de Scudéry wrote it. The online source states that the first edition had 13,095 pages, while the online (1656) edition has 7443 pages. If that’s correct, the first edition must have had either a very large typeface or widely spaced lines. Wikipedia lists the word count as 2.1 million. I’ve verified the plausibility of this figure with page sampling from the online edition. Clarissa is also available online. My page count is first-edition text pages, as documented in Sale (1969) pp. 45-8. The word count is from the online edition; see long-tale data.
[2] Calculation based on a typical reading speed of 250 words per minute, and a typical speed for spoken text of 140 words per minute.
[3] Cousin (1886) v. 1, p. 2.
[4] Newman (2003) p. 1.
[5] Aronson (1978) pp. 54, 82.
[6] Keymer (1994) pp. 392-3. The later figure is based on scaling Rivington’s revenue figures.
[7] Raven (1987) pp. 15, 40.
[8] Page count for first edition, based on Sale (1969) pp. 70-4. Word count scaled from words in the online volume 4. See long-tale data.
[9] Eaves and Kimpel (1971) pp. 384, 401.
[10] Keymer and Sabor (2005) p. 20; Raven (1987) p. 15.
[11] Eaves and Kimpel (1971) p. 490.
[12] Mott (1966) p. 304. Grandison, also published in the U.S. in 1786, was a “better seller” (not quite a best-seller) from 1786-1789.
[13] The page count and word count are from Wikipedia, here and here.
[14] Tadié (2000) p. 595. According to a history of Éditions Gallimard, which became Proust’s publisher, the company sold more than four million copies of À la recherche du temps perdu (in French, apparently worldwide) in seventy years through its copyright expiration in 1987. A significant share of these copies may have been purchased due to course assignments.
[15] Mott (1966) App. A.
[16] From WikiAnswers here and here. Included in long-tale data.
[17] See Salganik, Dodds, and Watts (2006).
[18] Aronson (1978) p. 39.


Aronson, Nicole. 1978. Mademoiselle de Scudéry. Boston: Twayne Publishers.

Cousin, Victor. 1886. La société française au XVIIe siècle d’après Le Grand Cyrus de Mlle de Scudéry. Paris: Perrin & Cie.

Eaves, Thomas Cary Duncan, and Ben D. Kimpel. 1971. Samuel Richardson: a biography. Oxford: Clarendon.

Keymer, Tom. 1994. “Clarissa’s Death, Clarissia’s Sale, and the Text of the Second Edition.” Review of English Studies, New Series, v. xlv, n. 179, pp. 389-96.

Keymer, Tom, and Peter Sabor. 2005. Pamela in the marketplace: literary controversy and print culture in eighteenth-century Britain and Ireland. Cambridge: Cambridge University Press.

Mott, Frank Luther. 1966. Golden multitudes: the story of best sellers in the United States.

Newman, Karen. 2003. “Volume Editor’s Introduction,” in Scudéry, Madeleine de, and Karen Newman. 2003. The story of Sapho. Chicago: University of Chicago Press.

Raven, James. 1987. British fiction, 1750-1770: a chronological check-list of prose fiction printed in Britain and Ireland. Newark: University of Delaware Press.

Sale, William Merritt. 1969 [1936]. A Bibliographic Record of His Literary Career with Historical Notes. Archon Books.

Salganik, Matthew J., Peter Sheridan Dodds, and Duncan J. Watts. 2006. “Experimental study of inequality and unpredictability in an artificial cultural market,” Science, 311, 854-856 (2006).

Tadié, Jean-Yves. 2000. Marcel Proust. New York: Viking.

5 thoughts on “explaining the long tale”

  1. Here’s my question, though. To what extent is the length of popular works related to the total numbers (or volume) of material available? That is, in 17th century France, if I had access to X hours of total material, don’t I probably have access to something like 1-bjillion-X hours of material today. So… does that mean that a commitment to spend 250 hours with one work is less of a big deal because there’s less choice? Or more of a big deal, because it takes up a larger percentage of the total available options?

    I’ve read very long series, including Harry, and have enjoyed them immensely. Dorothy Dunnett’s historical fiction comes to mind; House of Niccolo is (approximately) really bloody long. But I also enjoy the 2-second video of the startled hamster.

    I don’t know about this whole “Google is making us stoopid” and the death of deep reading thing. I compare what I generally do on the Web to what I do with my TV and am glad for the change.

    In the absence of hard data, I default to “more choices is better.”

  2. a commitment to spend 250 hours with one work is less of a big deal because there’s less choice? Or more of a big deal, because it takes up a larger percentage of the total available options? Good questions. In generally, I don’t think persons choose over the space of all available goods. Real constraints of time, personal experiences, and mental capacity constraint any specific person’s choice set. At the same time, a greater over-all number of items probably leads, through complicated social processes, to more satisfying choice sets (but not necessarily larger ones) for individuals. The issue, then, is not the change in the over-all number of items, but the change in the relative value of different items in the individual’s choice set. It’s difficult to relate that to the change in the over-all number of items.

  3. Hello Douglas
    Greatly enjoyed this piece which I found thanks to an alert on De Scudery in whom I took an interest some time ago.I hadn’t thought of a comparison with Harry Potter, but that’s the joy of seeing other people’s connections. Will be mentioning you on my blog some time soon.
    Best wishes
    AliB (UK writer and blogger)

  4. For reasons that seemed good at the time, I once took undergraduate courses in single a semester that required me to read 32 novels. Some of them were the longish tales and I read “Tom Jones”, “Barchester Towers” and others at single marathon sittings. It was an interesting experience but one I wouldn’t care to repeat.

    By the time I was ready to claim The Novel as a graduate specialty for doctoral written exams, I saw that I had read 300 of the 350 listed in the back of my Thrall and Hibbard and read the other fifty (by Norris, Dreiser, Smollett, Dorothy Richardson et al) for my exams. Harvey Curtis Webster gave me an individual seminar on the novels of Hardy and I was in his C. P. Snow seminar (all of Strangers and Brothers) and met Lord Charles, who is an obvious victim of the anti-long tale trend. And I had an “itch” in those days to read the complete works of everybody I liked one of.

    But I had to retake the exams (except Hardy as Major Figure). For one thing, “everybody” was talking about William H. Gass’ “Fiction and the Figures of Life” and there was a major question for which reference to Gass was the correct answer, though none of my professors had mentioned it to me or put it on a bibliography.

    That seemed unfair but there was a point there it took me some years to grasp. Reading really long novels is not a skill for which there is a market, either in the scholarly or commercial world. I should have been aware of what was hot in secondary sources.

    There is not, of course, any job market for video-watchers, either. But I am likelier to watch a short music video (and sometimes enjoy them) than undergraduates of my acquaintance are to read C. P. Snow. For that matter, I am likelier to read Henning Mankell than C. P. Snow and watch “guilty pleasure” television and movies.

    I find if I want to read, say, Drabble or Byatt, I need to clear some time and make sure I have some fresh earplugs. All of my neighbors have one or two dogs and there are children and motorcyclists, though this is a cul de sac street quieter than I have ever lived on, man or boy.

    I read “Tom Jones” those years ago while blocking out dormitory noise with The Well-Tempered Clavier, twice. Those were those days. These are these days.

    Terrance G. “Terry” Shults, Ph.D.
    Head of Technical Services
    J. Conrad Dunagan Library
    University of Texas of the Permian Basin
    4901 E. University
    Odessa, Texas 79762

  5. This is interesting — it seems the whole post, though, would benefit from a more in-depth analysis of the sites of reading, both then and now. What time of day did people read? In what kind of light? in what positions/on what kind of furniture? for how long? for what purpose? The only time I have to read long novels is in the subway or an airport — i.e., in locations and positions not available to the 17C reader.

Leave a Reply

Your email address will not be published. Required fields are marked *

Current month ye@r day *