Madame Butterfly: The Search Continues

Jan van
Rij’s Madame Butterfly: Japonisme, Puccini, and the Search for the Real
is the latest ó but by no means the last ó in a series of passionate
educated guesses attempting to track down the “real” Madame Butterfly. The
problem has been trying to read the minds and interpret the words of a number of people
involved in the creative process leading up to the first performances of the opera in Italy in
1904. Of particular interest to historical sleuths are comments made by John Luther Long, the
Philadelphia lawyer and writer, in the preface to his 1903 version of Madame
(the story that inspired the opera), and those of Jennie Correll, his sister
and source for the original 1897 Butterfly account, in a talk to the Tokyo Pan Pacific Club
and a subsequent magazine article in 1931. The remarks by both Long and Correll are extremely
vague and subsequently have led to a wide variety of interpretations. In Long’s
response concerning the models for Cho-Cho-San, Pinkerton, and even the original story, he
states the following: “And is she a fancy, or does she live? Both. And where is
Pinkerton? At least not in the United States navy – if the savage letters I receive from his
fellows are true. Concerning the genesis of the story I know nothing.”[1] Jennie Correll, who lived in Nagasaki with her Methodist missionary husband from
1892 to 1897, stated in her September 1931 magazine article that the model for Cho-Cho-San was
a certain Nagasaki tea-house girl named Cho-San. The young girl, along with her baby, had been
abandoned by her lover, who had promised to return by ship but never did.[2]

While these vaguely-worded assertions may have inspired van Rij and others to search for
historical models of characters in both the book Madame Butterfly and the opera
Madama Butterfly, the investigative trail has, indeed, been a foggy one. Van Rij
begins what his publicity release refers to as his “operatic detective
story” by examining the role that two relationships between European men and
Japanese women from Nagasaki might have had on the book and the opera.

The German physician Philip Franz von Siebold worked for the Dutch trading post at Dejima
from 1823 to 1830 before being expelled from the country. During his stay in Nagasaki, Siebold
lived with a Japanese woman named Otaki-san and their daughter Oine-san. Siebold returned to
Nagasaki in 1859 after the foreign settlement in the city opened and met briefly with his
Japanese “wife” and daughter.[3] Van Rij refers to
Siebold’s story as ”a prototype of sorts as it contained many of the
ingredients of the later novel and play that would be the precursors to Puccini’s
opera” (p. 24).

While the links between Siebold and Otaki-san to the Butterfly accounts are, to this writer,
tenuous at best, the second relationship mentioned by van Rij – that of Pierre Loti and his
young Japanese “wife” Okiku-san – clearly did have a strong influence on
both Long and Puccini. Pierre Loti (the pen name of the French naval lieutenant Julien Viaud)
stayed in Nagasaki during the summer of 1885 and two years later published an account of his
adventures in the city entitled Madame Chrysantheme.[4] I agree with
van Rij that “the Butterfly theme…entered Western literature for the first time
through Loti’s book” (p. 33) and that ”much of the material of
Long’s novel comes straight from Loti” (p. 112). Indeed, the contention
that Madame Butterfly was based on Loti’s account has been argued by
historians for years.[5]

Van Rij next examines the impact of Japonisme on the development of Puccini’s
opera. The curiosity involving “things Japanese” for Europeans and
Americans in the last half of the nineteenth century is undeniable and this fascination led to
a boom in Western artistic works concerning the customs and lifestyles of these exotic people
from the Orient. Van Rij investigates a number of possible influences of Japonisme on Puccini
or his librettists ó especially Camille Saint-Saens’s La Princesse Jaune,
Pietro Mascagni’s Iris, Judith Gautier’s La Marchande de
, Sidney Jones’s The Geisha, Gilbert and
Sullivan’s The Mikado, and Andre Messager’s Madame
. He also notes the possible inspiration of a Japanese touring group and
Claude Debussy. Van Rij concludes this chapter of his book somewhat uncertain of specific
influences, but absolutely convinced that Japonisme in some form played a key role in the
opera’s creative evolution. “Whatever Puccini’s sources were, it
is clear that Madama Butterfly was a product of Japonisme” (p. 56).

Chapter Three concerns the development of John Luther Long’s book*
Madame Butterfly and the subsequent play developed from the novel by David Belasco.
Van Rij investigates what Long did with the tale of “Cho-San,” as told to
him by his sister Jennie Correll from her days in Nagasaki, and how he fused this story with
Loti’s Madame Chrysantheme to create his best-selling account. Van Rij
concludes that “Obviously, nearly all of the elements of Long’s story have
been borrowed from Loti” (p. 66). Long is given credit for little more than
“chang[ing] some of the persons he borrowed from Loti into new characters
and…introduc[ing] a plot” (p. 67). One important change concerned the character
Cho-Cho-San and for this he relied heavily on his sister’s account. Van Rij also
asserts that Long added a moral message to his account that was lacking in the more
diary-based descriptive work of Loti (p. 74).

Long’s novel soon came to the attention of the American playwright David Belasco.
Belasco developed the book into a one-act play that debuted in New York in March 1900. Belasco
changed the characters little but added two major components to the plot: a more dramatic
vigil by Cho-Cho-San and her ultimate suicide. In June, Belasco took the play to London, where
it was viewed with great delight by Puccini, who was in the city to supervise production of
Tosca and to search for new ideas for his next opera.

In Chapter Four, van Rij discusses the making of the opera Madama Butterfly by
Puccini and his librettists. He especially concentrates on material borrowed from the novels
of Loti and Long and the play of Belasco. He concludes the chapter with an analysis of the
failure of the opera’s disastrous premiere at Milan and the changes made before its
successful second performance three months later at Brescia.

Chapter Five, in which van Rij speculates as to the real-life models for the characters in
Madama Butterfly and Madame Butterfly, is the most controversial section
of the book. Here he argues, not altogether convincingly according to this reviewer, that:
Cho-Cho-San was modeled after a Nagasaki woman named Kaga Maki; her son, Trouble, was modeled
after Kuraba Tomisaburo, a boy of mixed Scottish-Japanese parentage; and the father of
Tomisaburo was more than likely not Thomas Glover, the famous merchant who (with his Japanese
wife, Tsuru) raised the boy as his own, but Glover’s brother, Alfred.

At this point, van Rij enters a contentious debate to which he alludes in passing, but does
not do justice to in all its convolutions. The main players in this recent international
debate concerning models for characters in the Butterfly book and opera are Brian
Burke-Gaffney, a Canadian scholar and translator who lives and works in Nagasaki; Noda Kazuko,
a descendent of Tsuru Glover’s presently living in Tokyo, Arthur Groos, a professor
of German Studies at Cornell University in the United States, and van Rij, a former lawyer and
diplomat now residing in France.[6]

In 1991, Groos argued that, based upon Jennie Correll’s talks of 1931 and
subsequent research on his part, the model for Pinkerton was most likely an American naval
officer named William B. Franklin, who was in Nagasaki in 1892 aboard the U.S.S.
Marion.[7] Groos also speculated as to possible models for other
characters in the opera, but as for Cho-Cho-San he states that “we know almost
nothing about the original Butterfly” [8] and “we may in
fact never know any more about the historical O-Cho than what Jennie Correll tells us. To be
sure, the real madame Butterfly remains mute.”[9]

In the meantime, Noda set out to prove that her great- great-grandmother,Tsuru Glover, was
the model for Cho-Cho-San. Her efforts grew out of a crusade to rebut the claims by
Burke-Gaffney in 1989 that, according to family records, Kaga Maki, not Tsuru, was the mother
of Kuraba Tomisaburo.[10] Noda has gone around the world pressing her claim
that Tsuru was not only the mother of Tomisaburo, but that she was – as Noda’s
father had speculated in a 1972 book – the model for Cho-Cho-San as well.[11] The latter assertion is based on a collection of circumstantial evidence, including
the fact that Tsuru wore a butterfly crest on her kimono.

Last spring, Burke-Gaffney countered with his book on Butterfly, claiming that the search for
historical models beyond the general ones found in Loti is useless because the
Glover-Butterfly connection was fabricated by Americans in Nagasaki during the occupation of
the city after World War Two. To prove his point, he claims – among other things – that prior
to the war there was no mention of this connection in either local guidebooks or local
histories. He demonstrates how the wife of an American Occupation official living in the
Glover house deliberately began the rumor and how Nagasaki officials later picked up on the
story to publicize the large tourist attraction at Glover Garden in the former foreign

Van Rij has now added to this controversy with his book. He discounts Groos’s
assertion about the Pinkerton model, by saying that Groos relied too much on the historical
authenticity of Long’s book and Jennie orrell’s 1931 accounts (pp. 112-115
and note #1, pp. 175-177). Instead, van Rij argues that “B.F. Pinkerton…is
obviously modeled on Pierre Loti” (p. 69).

Van Rij rejects Noda Kazuko’s and Noda Heinosuke’s view that Tsuru
Gloveris Tomisaburo’s mother and accepts Burke-Gaffney’s claim that Kaga
Maki is, indeed, the mother (pp. 120-121 and note #8, p.178). He then goes even further by
naming her as the Butterfly model ó a rather unconvincing claim made earlier by a biographer
of Thomas Glover.[13] His identification of Alfred Glover as the most likely
candidate for Tomisaburo’s father (p. 134) is his own contribution as far as I can

I remain unconvinced by van Rij’s claims regarding historical models for
characters, because I do not believe that they exist for any of the main characters in
Madame Butterfly or Madama Butterfly. More convincing is van
Rij’s contention that the accounts by Long and Correll should not be viewed as
repositories of historical truths. I would also question the historical importance of some of
van Rij’s third-party sources: especially the one attributed to Miura Tamaki (the
Japanese soprano famous for her role as the operatic Butterfly) told to her by Long that
Tomisaburo was Butterfly’s son by an English merchant (p. 118 and note #7, pp.
177-178), and a second story Tomisaburo supposedly told another person in 1931 or 1932 that
was passed on years later by a third party to Noda Heinosuke. Since van Rij rejects other
aspects of the Noda accounts as unreliable, it seems surprising that he would place his
historical faith in this story. Finally, van Rij’s claim that Alfred Glover was
Tomisaburo’s father is pure speculation and frankly it is unlikely to receive much
support from scholars in the field.

Van Rij’s final chapter briefly discusses the fact the Madama Butterfly
has never been well-received by Japanese audiences. And the reasons for this should not be
difficult to understand. The heroine, Cho-Cho-San, is hardly depicted in a complimentary
manner and the opera is filled with historical and linguistic inaccuracies. As van Rij
rightfully notes, ”The only attraction of the Japanese element of the story is the
flattering fact of a foreign novel with a plot set in Japan and turned into a popular opera by
a famous European composer(p. 140).

This has not, however, kept generations of Westerners – including van Rij – from falling in
love with the opera. Madama Butterfly remains one of the most often performed operas
in the world and undoubtedly speculation will continue to abound as to the historical models
behind its main characters. Rather than accept the vagaries of undetermined influences from
figures such as Pierre Loti, Jennie Correll and John Luther Long (and, as van Rij has shown,
the genre of Japonisme) the search for the ”real” Cho-Cho-San
will go on – and on and on!


1. John Luther Long, Madame Butterfly (New York: Grosset and Dunlap,
1903), p. xiii.

2. Mrs. Irvin H. Correl [sic], “Madame Butterfly: Her Long Secret
Revealed,” The Japan Magazine, 21 (1931), pp. 341-345, as found in Van Rij,
p. 60 and Groos, p. 130.

3. During the almost thirty year interval, Siebold married a European woman
with whom he had five children.

4. Pierre Loti, Madame Chrysantheme (Paris: Calmann-Levy, 1887).

5. For example, see Jean-Pierre Lehmann, The Image of Japan (London:
George Allen and Unwin, 1978), p. 91. Lehmann also argues that “O-Kiku san became
Cho Cho san, from Madame Chrysantheme to Madame Butterfly” (p. 92), but van Rij
would say that it is more complicated than that.

6. In addition to van Rij’s book under review here, the most
representative works of the other three are: Arthur Groos, “Madame Butterfly: The
Story,” in Cambridge Opera Journal, Vol. 3, No. 2 (July 1991), p. 125-158;
Noda Kazuko, Madame Butterfly and Madame Tsuru Glover (Tokyo: DISCO Networking,
1998); and Brian Burke-Gaffney, Cho cho fujin o sagashite [Searching for Madame
Butterfly] (Tokyo: Kurieitsu Kamogawa, 2000).

7. Groos, pp. 141-148.

8. Ibid., p. 156.

9. Ibid., p. 158.

10. Burke-Gaffney, Hana to shimo [Blossoms and Frost] (Nagasaki:
Nagasaki Bunkensha, 1989), p.98.

11. Noda Heinosuke, Guraba fujin [Mrs. Glover] (Nagasaki: Shinnami
Shobo, 1994 [orig. 1972]), pp. 51-58.

12. Burke-Gaffney, Cho cho fujin o sagashite, especially pp. 54-70.

13. The claim was made by Alexander McKay in his book Scottish Samurai:
T.B. Glover, 1838-1911
(Edinburgh: Canongate Press, 1993).

Reviewed by Lane Earns (,
Department of History, University of Wisconsin Oshkosh. This review first appeared in H-Net and is copyright 2001 by H-Net, all rights reserved. It is reprinted with the
permission of the author.

* [OT: As a point of clarification, “Madame Butterfly”
first appeared in the form of a novella in the January 1898 issue of The Century
Illustrated Monthly Magazine
. In 1903, Grosset and Dunlap issued “Madame
Butterfly” in novel form merely by “stretching out the story by printing
only a small square of text on each page and by adding illustrations.” Brian
Burke-Gaffney, Starcrossed ó A Biography of Madame Butterfly (Norwalk: Signature
Books, 2004), 76.]

Addendum. Since Rij’s book appeared, two other studies have been
published: (1) Brian Burke-Gaffney, Starcrossed ó A Biography of Madame Butterfly
(Norwalk: Signature Books, 2004) and (2) Jonathan Wisenthal et al., A Vision of the Orient
ó Texts, Intertexts, and Contexts of Madame Butterfly
(Toronto: University of Toronto
Press, 2006).


image_description=Jan van Rij. Madame Butterfly: Japonisme, Puccini, and the Search for the Real Cho-Cho-San
product_title=Jan van Rij. Madame Butterfly: Japonisme, Puccini, and the Search for the Real Cho-Cho-San
product_by=Stone Bridge Press (Berkeley, CA), 2001. 192 pp.
Photos, map, chart, afterword, chronology, notes, bibliography, and index of personal names.
product_id=ISBN 1-880656-52-3