Artemis a Model for Widows?

Like Edith Wharton (previous post), Machado de Assis has what looks very like a mythological blunder in his very first short story (first collected, in his case), “Miss Dollar”. The very handsome and affordable new translation of the Collected Stories translated by Margaret Jull Costa and Robin Patterson (Liveright, 2018) includes these remarks about a beautiful widow very reluctant to remarry (13-14):

“She wants to remain faithful unto the grave, an Artemis for our own age.”

Unconvinced by this reference to Artemis, Andrade smiled at his friend’s remark, . . .

* * * * *

This quote had the effect of silencing Andrade, who believed about as much in constancy as he did in Artemises, . . .

Of course, the classical model for inconsolable widows was not the virgin goddess Artemis but Artemisia, specifically Artemisia II of Caria, who built one of the Seven Wonders of the World, the Mausoleum, as a tomb for her husband (and brother) Mausolus. Except for helpful stress-accents, the two names are the same in Portuguese: Ártemis and Artemísia. (Being ignorant of Portuguese, I would not have been able to check this easily until the last few years, since dictionaries usually omit proper names. Now I can just look up Artemis and Artemisia – the queen, not the generic name of mugwort, wormwood, and sagebrush, which comes up first – on Wikipedia, and click on Languages / Português in the left margin to see how the articles are titled in Portuguese.)

Though I may be wrong, the apparent blunder does not seem to be the translators’. Machado de Assis’ Obra Completa is out of copyright (he died 110 years ago last Saturday) and on-line at a Brazilian government URL ( The text of “Miss Dollar” there (click on ‘Conto’, then ‘Contos Fluminenses’) reads ‘Artemisa’ or ‘Artemisas’ in all three places. Could this be an earlier spelling of Artemis or Artemisia? That would be awfully confusing: it looks more like a conflation of the two, falling between two stools, as it were. Did Machado de Assis himself, or his copy editors, proofreaders, or typographers drop an I and an accent to turn Artemísia into Artemisa? That is a question only an expert on Brazilian Portuguese and the works of Machado de Assis can answer. However, he must have meant the woman whose English name is Artemisia, not Artemis, so the translation is definitely wrong.

Perhaps I should add that I’ve been very impressed by the quality of the stories I’ve read (four so far) and the translation reads very well, though I’m obviously in no position to judge its accuracy except on this one tiny point. I would have liked footnotes for some of the geographical and literary allusions, but there was hardly room for them: the book is already xxv + 931 pages. Perhaps someone could put together a companion website, with contemporary maps, explanations of what the named streets and neighborhoods imply socially and economically, identification of now-forgotten (at least outside Brazil) authors, and so on. (If such a site already exists in Portuguese, a translation would be very much appreciated by readers of the new Collected Stories.) I would be glad to help with the frequent classical references and allusions.

This entry was posted in Nachleben, Orbilius and tagged . Bookmark the permalink.

3 Responses to Artemis a Model for Widows?

  1. Raphael Soares says:

    I forgot to mention that the accent in crescent diphthong was introduced in the 1945 reform, and abolished in the 1998 reform, and this is why you have “Arthemisia” and “Arthemisa” pré-1945, “Artemísia” after 1945 in Brazil, and “Artemisia” in Portugal (and Brazil post-2009)… The acent in “Artemísia” is a brazilian feature of the XXth century, that is no longer accepted…

  2. Raphael Soares says:

    Hi, my name is Raphael Soares, and I came to this blog for Claudian, and I got a pleasure to see Machado de Assis in this place… As a Brazilian that had my own life changed by the influence of Machado’s work (I pursued my Master’s degree in Letters [Brazilian equivalent to Linguistics and Literature degree] after reading one of his masterful novels), I will try to explain and illustrate the point (including the last commentary), as an scholar in XIXth century literature. As a native brazilian portuguese speaker, some points looks obvious to me, but because is not necessarily obvious to everyone else I’ll try to explain every thing that I can thing that needs explanation… Please, don’t understand this as doubting your inteligence, is just that I want to give the better possible answer istead of a dogmatic “is this, period!”… Also, I need to say that I never studied this text extensively, partly because, as Brazilian, we tend to read the “early” work of Machado as a prelude to the “great” work by him, which blurs our senses in a certain way. I will divide this in two parts, I.Orthography and Orthoepy; and II.Interpretation and Sources.

    I. Orthography and Orthoepy
    Different orthographies for greek names aren’t uncommon. In english, for instance: “Clytenmnestra”, “Klytemnestra”, “Clytaemnestra” and even Browning’s “Klutaimnestra” are all valid spellings for the same name, even when the pronunciation is the same. But the case with “Artemis” and “Artemisia” has two distinctively portuguese features of double orthography, one orthographic and one orthoepic.
    The orthographic problem is more simple. Machado wrote “Arthemisa”. The edition that you consulted don’t have the “th”, that was removed in the 1945 reform. You can see the first Book edition here (, and I have a copy of the first publication (in a newspaper) and it is the same. This is just a minor issue, and don’t matter to the interpretation.
    On the other hand, the Orthoepic side, Portuguese don’t have “helpful stress-accents” and “inconsistent/sloppy accenting”. Portuguese is a language with very strict acentuation practices (there are inconsistences over time, because changes of rules or letters, like “portuguez” vs “português”), so when you see different acents is because the word not only is pronounced (aways) differently as they feel different words. The “classic” example are the words “sábia” (wise[f.]), “sabia” (knows[he]) and “sabiá” (Turdus fumigatus) they will never be confused with each other and are very distinct. In greek words that entered our vocabulary, there is a standard tendence to make them paroxytone, according to the “latin” practice (after the losing of long and short distinction), and this is why we have “catéter”, “Psique” and “Artemis” (all paroxytone, or feminine); at the same time, in late XIXth and early XXth there was a tendence to “correct” and adequate to Greek (and, sometimes, French) pronunciation, and we have “cateter” and “Psiquê” (both oxytone, or masculine) and “Ártemis” (proparoxytone). This is a difference in pronunciation, but it changes our perception of the language, because the words are, from a perspective of a speaker, different (unninformed people think that “Psique” and “Psiquê” are different entities), and sometimes blatantly wrong (there’s a long and awful grammatical debate about if “catéter” is correct or if “cateter” is a valid imposition). But Art[h]emis and Árt[h]emis are perfectly valid and acceptable ways, even if they predicate a different pronunciation. Machado’s spelling predicates a paroxytone, so I’ll use “art[h]” now on, representing also his “th” spelling, that doesn’t change pronunciation.
    The case with Artemisia is similar. The “ia” and “a” in portuguese are very weakly distinct, so is a pretty common change. In today’s portuguese, “Artemisa” is not a valid portuguese spelling of Artemisia (or even Artemis), but in the past it was. So “Art[h]emisa” is a valid spelling for “Art[h]emisia”, but not from “Art[h]emis”. On the other hand, “Artemisa” is a valid Spanish word for the Goddess, and the “-isa” ending sounds like a “ritualistic” feminine like the english “-ess” (priestess, poetess, prophetess), so, Machado made a blunder?

    II. Interpretation
    The short answer is no, because “Art[h]emisa” is a valid spelling of “Art[h]emisia” but not from “Art[h]emis”. The translators probably were way more proficient in spanish than in portuguese (which is awfully common), and this is the reason why I tend to recommend Grossman’s version (with all his blunders) over more modern translations. But, there’s a lingering chance that Machado de Assis got this from spanish, but is also not the case. The form “Art[h]emisa” is common to mean “Artemisia” in Portuguese, and I found in one of his books a reference to Bernardino da Silva’s “Defensa da monarchia Lusitana” (in defence of the portuguese monarchy):
    “como o [sepulcro] que fez Art[h]emisa a seu marido, e irmão Mauseolo Rei de Caria, conforme contra Strabo liuo 14″ [as the one [sepulchre] that Artemisa build for his husband, and brother Mauseolo, King of Caria]. As a heated monarchist, Machado probably read everything related to a defense on monarchic principles in his own library so, I think it closes the case. The translators made the error…

    I don’t blame them, partly because I aways assumed that he was talking of Arthemis here and never made a second tought on this. Again, I never read the “Contos Fluminenses” the way that the text probably deserves, simply because we are so blind over the other parts of his work that we don’t study the “lesser” (if is the case that these works deserves that title) works…

    I hope it helps…

  3. Alfred M. Kriman says:

    Not a complete solution to the puzzle, but of apparent relevance:

    (1) The last major reform of Portuguese orthography was only in 1945 or thereabouts, moving it towards the less historical style used in Spanish. This late change is the reason the English word commando (from Portuguese via southern Africa) is spelled with two m’s.

    (2) Just past the title of the Portuguese Wikipedia page for Ártemis, one reads that Ártemis and Artemísia are regarded as alternative names for the goddess. That seems to be an editor of whoever wrote it into the article. Googling yields the following hit counts:
    (Don’t strain your eyes: if two search strings seem nearly identical, the second one is missing an accent.)
    “a deusa Ártemis” 23100
    “a deusa Artemis” 17300
    “a deusa Artemis” OR “a deusa Ártemis”, inconsistent with the above 21200
    (using OU is much worse)
    “a deusa Artemisa” 1570
    “a deusa Artemísia” 141
    “a deusa Artemisia” 7
    And the Artemisias of Caria are not entirely forgotten:
    “Artemísia” “rainha” “Cária” 2490
    “Artemisia” “rainha” “Cária” 258
    “Artemísia” “Cária” 4490
    “Artemisia” “Cária” 1250
    “Artemísia II de Cária” 1850
    “Artemisia II de Cária” 9370
    “Artemísa II de Cária” 0
    “Artemisa II de Cária” 447

    I infer tentatively that Artemisa is a rare (~5%) variant of or error for both Ártemis and Artemisia, and that Artemisia is very rarely confused with Ártemis. The Miss Dollar spelling looks like a simple error.

    Interestingly, that Portuguese Wikipedia page for Ártemis contains an apparent error in footnote 1. The note is about Ártemis representing the moon, but it gives her name in the Miss Dollar form of Artemisa without explanation.

    I cannot understand the apparently inconsistent accenting of Artemisia and Artemísia, but Google and writers can both be a bit sloppy about accents. There shouldn’t be much contamination from Spanish, since Caria clearly has no use for an accent on the first a in Spanish, and adding an unnecessary accent is an unlikely error. (If you know, you leave it out because that’s correct; if you don’t know, you leave it out because you don’t care.)

    FWIW — possibly something — in Spanish the dominant form of the name of the goddess is Artemisa (by maybe a factor of 2 or 3).

Leave a Reply to Alfred M. Kriman Cancel reply

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

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>