That’s the title of a post at Poemas del río Wang that I discovered only because Studiolum opened up a back door to his wonderful site that allows you to see over a half-year’s worth of posts at once (as opposed to the normal view, which only shows one or two posts because of the large number of images each contains); I never realized how many posts slipped under my radar, and this one from a month ago it would have been a particular shame to miss, because Lev Nussimbaum/Essad Bey/Kurban Said is a favorite of mine (I’ve written about him here and here). I wrote a long enthusiastic comment only to have it eaten by some damn Blogspot glitch (reminding me of how glad I was to leave that motheaten venue back in 2003 when I got my own domain); too grumpy to try again, I licked my wounds and finally decided to post here instead.

As you would expect from río Wang, there are many gorgeous old images of Baku, as well as one of “Essad Bey in Caucasian mountain dress” and one of the cover of this issue of Azerbaijan International, entirely devoted to the silly business of trying to prove that the book was not written by its oddball Jewish/Muslim Azeri/German author but by a “real” Azerbaijani, “the national poet Yusif Vazir Çemenzeminli.” This is comparable to the desperate attempts to prove that the works of Shakespeare were actually by someone other than the commoner who wrote them, a mere actor who could not possibly have written great poetry and seen into the depths of the human soul (as we all know earls are able to do by virtue of their blue blood); class prejudice and nationalism are parallel forms of blindness, and I wanted to warn the good Studiolum against allowing his mind to be swayed (he writes that the issue “offers very convincing arguments for the authorship of Yusif Vazir”), but I’ll do so here rather than chez lui. At any rate, do go over there and enjoy the material on display, and perhaps bookmark that most useful back door link.


  1. For awhile I’ve been collecting books about imposters and forgers like Essad Bey: Baron Corvo, Trebitsch Lincoln, and Sir Edmund Backhouse. All of them were tremendously talented, all successfully assimilated themselves to some distant culture, all pretended to be someone they were not, all have been the subject of a full length biography (Corvo has two or three), and all but Lincoln have books in print today. Bey and Corvo at least have made permanent contributions to literature, and Backhouse donated an enormous number of books to the Bodleian library.

  2. marie-lucie says

    JE, I hope you write a book about them.

  3. It would be pretty secondary, no primary research. But maybe a long blog post.
    I just added a guy named David Plotkin who isn’t an imposter but whose works included a forged memoir by Nietzsche confessing to inc*st, a scurrilous novel commissioned by political opponents to smear Senator Burton K Wheeler, a pioneering lesbian porn novel, a cheesy war novel, a cheesy book about Asian politics, a cheesy dictionary of maxims, and a book of what looks to be Sandburg-esque poetry. He writes quite amusingly while faking his way through the facts.

  4. Apparently George Psalmanazar and his Formosan Alphabet haven’t figured here before. (Also not to be missed is “The reason why the Japan Language differs from that of the Chineſe and Formoſans” on the facing page.)

  5. JE: a pioneering lesbian porn novel
    That’s a neat characterization. It can be interpreted in several ways, depending on how the sentence is parsed. Here are some of those ways:
    1. a stylistically unremarkable novel in which Westward-Ho! lesbians engage in ho-hum sex
    2. a stylistically unremarkable novel in which suburban lesbians engage in unusual kinds of sex
    3. a stylistically ambitious novel in which suburban lesbians engage in ho-hum sex

  6. Martin Lehman says

    Before scoffing at the serious research that has been undertaking by the staff of Azerbaijan International magazine, over a period of six years, one should at least look seriously at it. You would be amazed at the in depth examination with incredible references after deep searching in the Azerbaijan National Archives, Institute of Manuscripts in Baku, visit to the Lichtenau Castle in Austria, etc. Documents in 10 languages were examined.
    So before attacking the work as spurious, and before attacking the incredible insightful and valuable Blog by rio Wang, be honest enough to know what you are talking about.
    Read first. Don’t expose your own prejudices before knowing the issues yourself.
    And, in fact, some of your own comments about Ali and Nino are referenced in the Azerbaijan International’s volume, a triple issue of Azerbaijan International, 364 pages available in English or Azerbaijan. Take a look at http://AZER.com.
    It’s Tom Reiss’s “research” that deserves much more scrutiny.
    Be careful who you are accusing of ethnic prejudice and nationalism before you examine carefully the issues.

  7. Martin Lehman says

    For more about the Essad Bey research. See Table of contents here.

  8. LH: Many thanks for linking the back door. Invaluable.

  9. before attacking the incredible insightful and valuable Blog by rio Wang
    And you’re urging others to “know what you are talking about”? I have never attacked Poemas del río Wang, one of my favorite sites, nor would I dream of doing so. But life is too short to investigate every enthusiastic claim that comes down the pike (and they’re all supported by piles of “serious research,” from cold fusion to perpetual motion to the Earl of Oxford); I’m afraid you’ll just have to live with my continuing skepticism. I realize how aggravating it must be for Azeris to think that the one novel about their country that any foreigners have read is by an ethnic outsider who left young and never came back, and I sympathize (a tiny bit) with their desire to believe in a more suitable author, but not being Azeri myself, I can’t be expected to share it. And frankly, the difficulty both Germans and Azeris seem to have in accepting a Nussimbaum as author of the book… well, as I said, my sympathy is limited.

  10. That said, I agree that Tom Reiss’s research leaves much to be desired, and I wish someone would do a better biography.

  11. dearieme says

    I offer, free, my latest sally at a non-Stratfordian. “He was uneducated” said he. “He went to a perfectly good Grammar School” said I. “He didn’t go to University” said he. “You mean Oxford or Cambridge” said I “but now that we suspect that he spent part of his ‘lost years’ at St Andrews, all is explained”. That had him stumped. Fiction, don’cha love it?

  12. Thank you for the share. There are still a lot of items not really clarified in the biography but it is a pretty good overiew. Would love to see someone out out something a bit better though.

  13. Why is it only the Earl of Oxford when there’s a duchy of Cambridge? As someone pointed out at the Guardian today, if they gave an OBE to an earl he would become an earlobe.

  14. As I have explained, the just-married Prince Wiliam is Duke of Cambridge and Earl of Strathearn, but he is not the Duke of Earl.
    They should have used the song somewhere in the wedding, though.
    I am willing to bet that sister-in-law Pippa frequently shows up in the tabloids. it’s as if she were designated for that role.

  15. I can see one headline coming up: “Pippa passes exams!”

  16. The Duke’s in his Earldom –
    All’s right with the world!

  17. Thank you very much, Language Hat, for your kind words and recommendation. I am sorry for the trouble with Blogger’s messaging service (now I have included an appropriate warning in the header of the comment window), but it seems that ultimately it was a felix culpa, for it has led to the birth of this post!
    In the following days I try to build out this “back door” so large that it would offer you an entrance to all the posts of the blog.
    What fascinated me in this issue of Azerbaijan International was not primarily its attempt to prove the authorship of Yusif Vazir – of which, being no specialist myself, I can only say that it brought up really convincing arguments while recognizing the hand of Essad Bey in the final, published version of 1937 -, but much more the thorough investigation of the sources behind the book and the identification of its scenes and motifs with spots and figures in early 20th-century Baku and Caucasus.
    This work, done by several dozens of contributors, has led to the accumulation of a very precious documentary and visual material, quite well presented in the volume: a proof for it are the illustrations of Río Wang’s post on Essad Bey, almost all taken from this issue of AI. In fact, I purchased this volume on the last day of my travel to Baku because, when hesitating in the bookshop between a number of monographs on the fascinating period of the city’s culture during the first oil boom, this one was definitely the most informative one in whatever language. I also intend to write more about its merits in a forthcoming post.
    I must also thank Martin Lehman for defending Río Wang, but I feel that absolutely no attack was done against it. Language Hat’s benevolent warning to me against becoming, for mere sympathy and on the basis of dubious arguments, an enthusiastic advocate of the cultural claims of a small nation is quite reasonable, as this danger always threatens the outsider when paddling on the troubled waters of those regions.

  18. I doubt that she’s taking any more exams; Pippa is, like dearieme, a graduate of the University of Edinburgh. Dearieme played number eight for the Edinburgh University Veterinary School rugby team. Like Pippa, he’s not even a vet.

  19. What about this?: “Shock! Pippa passes kidney stone!”.

  20. Hat! How cruel can you be?!
    But two months ago I was given three books by/about Essad Bey. And here I am, 6,000 miles distant from them and without means to add anything useful to this thread.

  21. All right, but you can’t expect me to know the name of every little thing by Robert Browning. Anyway, I find “the snail’s on the thorn” to be a disturbing and unlikely image – poor snail. “Pippa passes sister in race up aisle” would be another possibility, but I don’t want to get in trouble with the feminist police.

  22. I find “the snail’s on the thorn” to be a disturbing and unlikely image – poor snail
    I have always assumed “the thorn” to be synecdoche for some snail-encouraging habitat. Rose bush maybe, or do they only harbor slugs ? If Browning had meant “poor snail”, he would have written “the snail’s nailed on the thorn”.
    But maybe it’s a Christ image after all !? That would be a suitable examination topic for an EngLit course.

  23. Then he should have written “the snail’s on the rose bush”. “The snail’s nailed” is pretty good, though; well done.

  24. Yes, well done. Half the reason I keep this place going is so you lot can have a convenient venue to display the wit for which you are notorious, especially in Milan, where even the bishops kick up their heels.
    Hat! How cruel can you be?!
    Cet animal est très méchant: quand on l’attaque il se défend.

  25. Although, to be fair, I started by sort of attacking his magazine, though even vigorous disagreement should not be construed as personal attack. But that’s life in the fast lane!

  26. Ali and Nini is exoticizing Azeris and Islam, while being sympathetic and knowledgeable. That’s a pretty unsurprising perspective when you know the author’s biography, but isn’t what you’d expect from an azeri. The exoticising/romiticizing is the heart of t he novel, and is persistent and never crude. It’s not the thing an editor could insert, I think.

  27. An excellent point.

  28. Cet animal est très méchant: quand on l’attaque il se défend.

    I came across this recently, and I’ve been wondering how to translate it as a couplet. How about – “This animal’s a nasty thing: when it’s attacked, it gives a sting.”

    My concern about this, although I like it, is that “give a sting” is a little un-idiomatic, and doesn’t seem apt for mammals. (“…it’s like to sting” might be more elegant, but that would probably sound too antiquated.)

  29. @Lazar: I feel like changing “il se défend” to “it gives a sting” loses some of the sarcasm of the original.

  30. Yes, I’m afraid the original is essentially untranslatable into English (as clever light verse, which is its entire point).

  31. David Marjanović says

    From May 1, 2011:

    Also not to be missed is “The reason why the Japan Language differs from that of the Chineſe and Formoſans” on the facing page.

    Let me second that.

  32. This animal is very bad / When you aggress it bites your ‘nad.

  33. I love it: “the Japanneſe are continually changing and improving it every Day”!

  34. This animal’s reprehensible
    if you attack it defends itself

  35. marie-lucie says

    méchant is a lot stronger than “reprehensible”.

  36. Lars (the original one) says

    Like here: ce méchant Loup se jeta sur le petit Chaperon rouge, et la mangea. Which sentence I’ve been looking for an excuse to bring up, since I find it interesting that natural gender trumps grammatical in the pronoun but not the article. Do speakers of French accept with perfect equanimity masculine noun phrases as metonymy for female characters or vice versa?

  37. David Marjanović says

    I find it interesting that natural gender trumps grammatical in the pronoun but not the article

    The pronoun is farther away. That often has the same effect in German (making prescriptivists unhappy).

  38. lars: ce méchant Loup se jeta sur le petit Chaperon rouge, et la mangea.

    The entire phrase le petit Chaperon rouge is the name by which the little girl is known, with the article le agreeing with the noun chaperon ‘hood’*, but la refers to the girl, not to her name or any component of her name. You would not say in English, the wolf ate the little red riding hood, which would mean that the wolf ate the hood! Similarly the wolf did not eat le chaperon rouge, the red hood, in which case the Object pronoun would be the masculine le, but the girl referred to by the phrase, hence the feminine Object pronoun la. The distance of the Object pronoun from the noun phrase has nothing to do with the gender used.

    Compare with another legendary character, Barbe-Bleue, “Bluebeard”, named thus because il avail la barbe bleue, he had a blue beard. The noun barbe is feminine, and so is the whole phrase, but if referring to the character the gender of the noun or phrase used as a name does not matter, only the nature of the person: Barbe-Bleue était très méchant, il tuait ses femmes.

    *le chaperon is an old word for a short hooded cape, a medieval garment. Its only modern meaning is ‘chaperone’.

  39. Lars (the original one) says

    @m-l, thanks but that part I wasn’t really in doubt about — my real question is whether the mismatch between the noun phrase le petit Chaperon rouge and the girl it refers to causes any sort of cognitive dissonance for a native speaker of French? It does for me and my school French, but that’s not a reliable indicator.

    (I don’t think I’d read the tale in French until very recently, and I’m sure that before that I would unhesitatingly have used the feminine article if quoting the title since I did not imagine that a masculine noun could be used).

    Barbe-Bleue is used as a name, I think, so doesn’t have a problem with the article — but the question of grammatical gender mismatch remains, if you have that internalized (unlike me who has to rely on the articles).

    @David, the same question applies in German of course, but I get the impression that there are many more neuter diminutives used for people so maybe das Rotkäppchen is just business as usual. But what happens when diminutives are used as ‘real’ names in dialects where articles are mandatory? Is it das Hänsl or der Hänsl?

  40. my real question is whether the mismatch between the noun phrase le petit Chaperon rouge and the girl it refers to causes any sort of cognitive dissonance for a native speaker of French?

    It’s pretty clear from m-l’s comment that it doesn’t for her, and her explanation makes sense.

  41. @Lars: It’s das Hänschen / Hänsel*1) for me, which can be referred to both by neuter es or male er, depending whether the emphasis is on him being a child (das Kind) or a boy. The neuter pronoun would sound somewhat more literary / archaic to me.
    *1) Thinking about it, I might probably as well say der Hänsel, as the -el diminutive is not part of my idiolect (it’s Southern), and Hänsel is first and foremost the name of a (male) fairy tale character for me.

  42. Is it das Hänsl or der Hänsl?

    I find in Language Typology and Language Universals: An International Handbook, Volume 1 (Haspelmath et al., eds., 2001) an article called “Ökonomie” by Wolfgang Ullrich Wurzel, with this useful paragraph (broken into two for readability; apologies for any typos):

    Die Eigenschaften der verschiedenen Ebenen können prinzipiell unabhängig voneinander sein, sie können aber auch implikativ voneinander ableitbar sein. Ableitbare Eigenschaften müssen nicht explizit im Lexicon spezifiziert sein. Daher sind Lexeme mit partiell ableitbaren Eigenschaften für die Sprecher (unter sonst gleichen Bedingungen) ökonomischer als solche ohne ableitbare Eigenschaften. Entsprechend sind die Lexikoneinheiten in allen Sprachen hinsichtlich ihrer Eigenschaften in starkem Maße implikativ strukturiert. So ist typisch, daß bestimmte semantische Eigenschaften der Lexeme implizieren.

    Beispielsweise sind in Sprachen mit einem maskulinen und einem femininen Genus in der Regel Bezeichnungen für männliche höhere Lebewesen Maskulina und solche für höhere weibliche Lebewesen Feminina. Das wird durch bestimmte Sprachveränderungen, auch im Deutschen, bestätigt. Man vgl. die mit dem Morphem -el gebildeten Diminutiva zu Vornamen, die vom Neutrum zum Femininum bzw. zum Maskulinum übergegangen sind: älter das Gretel, das Hänsel > die Gretel, der Hänsel.

    The last bit says it all, even if you don’t read German (GT does a fairly good job, too).

    (Hans’s comment concurrent with mine shows that he is straddling the Sprachveränderung.)

  43. lars, LH: my real question is whether the mismatch between the noun phrase le petit Chaperon rouge and the girl it refers to causes any sort of cognitive dissonance for a native speaker of French?

    LH is right, it does not!

    Suppose a mother or grandmother had lovingly sewed a little red hooded cape for a little girl, she could refer to the girl when wearing it as “mon Petit Chaperon Rouge” and just afterward use “elle” or other feminine noun phrase.

    Perhaps there could be a cognitive dissonance if the girl’s nickname was not culturally well-known, for instance by a foreigner familiar with French but not with popularly known fairytales.

  44. David Marjanović says

    Der Wastl is a masculine diminutive.

    Other than that, diminutives are very rarely used as nicknames in dialects I’m used to.

  45. Lars (the original one) says

    Thanks for all the explanations!

    A somewhat comparable situation in Danish is the use of common/neuter deictic pronouns for ‘höhere Lebewesen’ — which only works for children below a certain age where the actual sex can plausibly be unknown. This means that babyen can be referred to as den, barnet as det, but personen only as han or hun. I think it’s slightly undefined whether you can use den/det of a child whose sex is known from context, but I’ve been thinking too much about this to trust my intuition.

  46. Lars: like Marie-Lucie, I see nothing remarkable about the (seeming) mismatch between masculine “chaperon” and the fact that the bearer of the name is female, and hence triggers feminine agreement.

    In some contexts, this mismatch triggers apparent mistakes/exceptions to gender agreement, even when non-human beings are involved: there exists a river in Quebec called “La diable”, which might seem to be a mistake (“Diable” is masculine, and a she-devil is a “diablesse”), except that here the feminine article “la” refers to the (understood) noun “rivière”: what was originally “La rivière du diable” was shortened to “La diable”.

    So, could “Le petit chaperon rouge” ever become “La petit chaperon rouge”, the latter understood as “la (petite fille au) chaperon rouge”? I do not think so, because, as Marie-Lucie pointed out above, “chaperon” is obsolete in the meaning “cape”: thus, “Le petit chaperon rouge” could best be described as a lexicalized phrase which refers to a well-defined fictional little girl.

    “Barbe-bleue”, on the other hand, which normally does not require the definite article, I could easily see being called either “La barbe-bleue” (with the article agreeing with the noun) or “Le barbe-bleue” (with the article agreeing with an understood “homme”). Testing native speakers on this point (which gender would they choose if they had to use a definite article with “barbe-bleue”) could prove an interesting experiment…

  47. @Etienne: every American with a little bit of knowledge of French stumbles over Le Reine Elizabeth (the hotel) when visiting Montreal.

  48. Gary: The queen is la reine Elizabeth, but the building is un hôtel named “Reine Elizabeth”, so if you omit the masculine noun hôtel you are left with the masculine article le followed by the name. But it is still odd to see this combination emblazoned at the top of the building.

    Etienne: I think that names which are plural present a similar case. Years ago I read a novel in which the love interest of the hero was referred to as les Yeux Noirs (I think this was translated from English “Dark Eyes”) rather than by a name. In English you would have to refer to this person as “she”, not “they”, and similarly in French, even if you had just used the (nick)name earlier in a sentence.

    With “Barbe-Bleue” I think I would use “la” for the nickname, but immediately switch to a masculine pronoun to refer to the man.

  49. Etienne, since you are here, can you send me an email? I still have the same address, but I think the one I have for you is obsolete. Merci!

  50. David Marjanović says

    There’s a beer brand called König which advertises itself as Das König der Biere in accordance with das Bier.

    Also, names of ships are mercilessly treated as feminine in German.

  51. Watching news about Norwegian election, run into ad by female politician who claims that she is “en statsminister for alle”.

    Poor Norwegians apparently find nothing strange in this phrase, because they abolished feminine indefinite article many years ago.

    Grammar male chauvinists

  52. SFR: female politician who claims that she is “en statsminister for all”.

    In French, many words for professions and such are single-gender, so for instance le médecin ‘the medical doctor’. There is nothing unusual in saying mon médecin (masc.) est une femme ‘my doctor is a woman’. In the modern period many such words have been provided with equivalents in the other gender, so nowadays you can say either le ministre or la ministre, but not all. (This actually started in Québec, with France following).

  53. And the political connotation is the reverse from English-speaking countries. Making gender distinction in (high-prestige, traditionally male) occupation names is progressive in France, while not making the distinction is progressive in Anglo-America.

  54. The English invariant article the descends from the Old English masculine nominative singular article se, with the /s/ replaced by the /θ/ of all the other OE article forms. The neuter singular nom/acc form þā is the ancestor of the case-invariant demonstrative that.

  55. Surely you mean þæt?

  56. Yes, I copied from the feminine rather than the neuter column. ~~ headdesk ~~

  57. Nobody’s perfect!

  58. Marie-Lucie: done.

  59. Lars (the original one) says

    @SFR, Bokmål (as a written standard) has only neuter and common gender for nouns, so there is no mismatch there. I don’t how what a speaker of a three gender variety would read it out, though.

  60. David Marjanović says

    Making gender distinction in (high-prestige, traditionally male) occupation names is progressive in France

    Also in German, to “make women visible” and avoid the impression that everyone is male either in reality or as a default.

  61. The first Swedish grammar I ever read called the common gender “nonneuter,” an interesting word.

  62. Stephen C. Carlson says

    The first Swedish grammar I ever read called the common gender “nonneuter,” an interesting word.
    Interesting. Mine calls it utrum. That’s like calling your genders “either” and “neither.” In a less technical setting, my SFI instructor simply called them ‘n’ (en) and ‘t’ (ett).

  63. Mine calls it utrum
    That’s a usage I’ve encountered before.

  64. Etienne, sorry, I still have not received your email, can you check or resend? my address is mllastname@hotmail.com. I want to ask a couple of questions. Merci!

  65. Lars: @SFR, Bokmål (as a written standard) has only neuter and common gender for nouns, so there is no mismatch there.

    No, it doesn’t. Bokmål allows any point along the continuum from two to three genders.

    2 genders:
    en mann – mannen (“man”)
    en jente – jenten (“girl”)
    et hus – huset (“house”)

    2,5 genders:
    en mann – mannen (“man”)
    en jente – jenta (“girl”)
    et hus – huset (“house”)

    3 genders:
    en mann – mannen (“man”)
    ei jente – jenta (“girl”)
    et hus – huset (“house”)

    The intermiediate points on the continuum are defined by the proportion of feminine nouns being written as feminines in either the indefinite (2,0-2,5) or definite (2,5-3,0) form.

    I don’t how what a speaker of a three gender variety would read it out, though.

    For me it depends on the context. If it’s poetry, I’d certainly follow the written text. If it’s prose, quite likely, and especially if the social status or aspirations of the narrator is conveyed by their grammar. if I’m reading a text meant to sound personal in my mouth, I’d change it.

    (Not staying for long. There’s more than a week of posts to wade through, and I’ll be busy elsewhere for a few more days.)

  66. Watching news about Norwegian election, run into ad by female politician who claims that she is “en statsminister for alle”.

    That’s not strange. A man can be ei bølle “a bully” or ei pingle “a sissy”, or in some dialects ei kjempe “a giant; a strong man; a champion”.

    Incidentally (but not important for the statement above), Prime Minister Erna Solberg is from Bergen and, as most Norwegians, she speaks her dialect also in her official capacity. The Bergen dialect is thoroughly two-gender. Apart from that it’s not far from standard Nynorsk.

    Poor Norwegians apparently find nothing strange in this phrase, because they abolished feminine indefinite article many years ago.

    We abolished the equation of semantic and grammatical gender many years ago.

  67. Marie-Lucie: done, again.

    On Scandinavian and Dutch gender: I find it interesting that the two genders of Standard Dutch, Danish and Swedish, and of some lects of Bokmål, are never designated by the terms animate/inanimate. Is this simply a matter of terminological history/preferences, or is there something about the gender system of these languages that makes them fundamentally different from the two-gender (animate/inanimate) systems of Algonquian languages or of Hittite?

  68. Stephen C. Carlson says

    Even within its two-gender system of common and neuter, Swedish has a masculine gendered weak form for adjectives. It shows up with the ending -e instead of -a, e.g.: den store mannen, ‘the large man’. I think it’s phasing out, however.

  69. Etienne, merci! It was my fault, not yours or the computer’s.

  70. David Marjanović says

    Is this simply a matter of terminological history/preferences

    I’m sure that, because of history, the terms animate/inanimate have never been considered for these languages; and indeed I’ve encountered the terms common/neuter for Hittite. However, in Hittite, “neuter” nouns have mostly inanimate referents, and “common” nouns mostly have animate ones, and AFAIK that largely holds for Algonquian as well; in the Germanic languages in question, the distribution is much more arbitrary, more like neuter vs. the other two in German.

  71. That was the common view, but it was exploded as long ago as Brosnan 1978 (Google Books), at least with respect to the (rare) u-stems and the (frequent) i-stems, where nouns with inanimate referents are equally likely to be of animate gender as of inanimate gender. Perhaps in consequence, more recent Hittitologists are using the names “common” and “neuter” instead of or as well as the traditional names, as a google for [hittite gender] will show.

    Per contra, the division in Russian of masculine nouns into two sub-genders, animate and inanimate, actually does reflect animacy, as the (linguistic) genders of English really do overwhelmingly reflect (socially assigned) gender or the lack of it. (In one of Delany’s novels, she has come to be applied only to known individual persons, he to generic or unidentified ones.)

  72. January First-of-May says

    the division in Russian of masculine nouns into two sub-genders, animate and inanimate

    By “masculine”, you mean “ending in a consonant”, right? Sure, all of those are masculine*, but many other masculine nouns don’t end that way.
    Besides, the distinction exists for the other nouns as well, it just doesn’t show up in the singular.

    (In fact, all six combinations exist, though neuter animate nouns are scarce; one example is чудовище, meaning “monster”.)

    *) some non-masculine loanwords end in a consonant, but they are undeclined

  73. It’s true that nouns of all genders may be animate or inanimate, but only in masculine nouns does this make a grammatical difference: in the singular, for animate nouns the accusative = the genitive, whereas for inanimate nouns the accusative = the nominative. This is not inherited from PIE, like the nominative = accusative in neuters (preserved in English only in the pronoun it), but an innovation in Slavic. Therefore we can speak of animate and inanimate sub-genders within the masculine gender, but not within the feminine or neuter gender, which make no such grammatical distinctions.

    (In Polish, things are inevitably more complicated: there are human, non-human animate, and inanimate sub-genders.)

  74. @Stephen C. Carlson: Standard Dutch also retains a sort of phantom feminine gender, such that some “common” nouns take non-biologically-determined feminine pronouns, although I’ve heard that in the vernacular they usually just default to hij.

  75. David Marjanović says

    Paul Brosman isn’t Pierce Brosnan… thanks for the link, though; I’ll read the rest later.

  76. January First-of-May says

    It’s true that nouns of all genders may be animate or inanimate, but only in masculine nouns does this make a grammatical difference: in the singular, for animate nouns the accusative = the genitive, whereas for inanimate nouns the accusative = the nominative.

    That’s what I was talking about 🙂

    To reiterate: this is true of all genders in the plural (and, in fact, all six possible combinations exist, though there apparently aren’t any animate plurales tantum), and only of a specific declension in the singular (which happens to include only masculine nouns, but not all of them).

  77. Lars (the original one) says

    @Stephen, the masculine weak form is indeed dying out — it remains as a sort of hyperpolite form in greeting phrases like Bäste kund (‘Dear Customer’), mostly in form letters directed at customers of unknown gender! (But calqued from old business correspondence where the (known) recipient was usually male, and the -a form would be used for female recipients).

    It was marked enough that Hemingway’s ‘The Old Man And The Sea’ bears the title Den Gamle och Havet in Swedish, no need to add the word man there — and lawyers will still insist that if they write avlidna = ‘the deceased’ they have actually told you that the deceased was a woman, because otherwise they’d have written avlidne. Nobody else knows what they’re talking about.

    An old strong masculine plural also survives in gode män, where god man is the designation for a legal guardian (regardless of actual gender). A quick Google search shows that this irregular form is then carried over to the weak flexion where I don’t think it belongs, like in de gode männen.

    On phantom genders, one of my colleagues would actually answer a question about the time (Vad är klockan) with Hon är två instead of the more usual Den är två.

  78. David Marjanović says

    I’ll read the rest later

    So… that book chapter only talks about the u- and i-stems, is meant to establish that the neuter was not dying out in Hittite (an idea I had never encountered), and starts from the assumption that the feminine gender was present in PIE and lost in Hittite, though it’s considerate enough to spell this out. It also says the meaning of quite a few of the counted words was unknown in 1978. So much research has happened since then that I can’t take the implied conclusions for granted.

  79. Stephen C. Carlson says

    @Lars, thanks for the information. It struck me as more literary than oral. I don’t think it was even taught in SFI.

  80. David Marjanović says

    Here it says on p. 20:

    PIE was traditionally reconstructed with three genders (masculine, feminine and neuter), but it appears probable that the earlier system, which may be preserved in the Anatolian languages, had only two genders (Matasović 2004): common (for all animates and a large number of inanimates) and neuter (for the remaining inanimates).

    So basically “common” and “inanimate”. 🙂

    Earlier in the paper (which is about something else, namely contacts between PIE and Caucasian language families) there are hints that the grammatically animate inanimates may have been those considered capable of action, like body parts.

  81. I think there may be a tendency towards a realignment of the genders in Norwegian, not as an animate/inanimate system but as abstract/concrete. At least I occasionally notice that people, very much including myself, may use den m./f. for concrete neuters and det n. for abstract concepts of any gender.

  82. David Marjanović says

    And wouldn’t you know, that’s one of the hypotheses on the origin of the IE feminine – as an abstract gender.

    (It hinges on the ideas that the *h₂ in the most easily reconstructible “woman” word, *|gʷenh₂|, was a suffix for abstract nouns and that the word was some kind of patriarchal euphemism derived from giving birth. I hope I can find that paper again, it’s somewhere on academia.edu.)

  83. David Eddyshaw says

    I know of another system of grammatical gender in decay where the abstract/concrete distinction has become salient.

    Kusaal has lost the elaborate system of grammatical gender based on noun classes typical of Gur (and other bits of Niger-Congo, like Bantu, of course.) It now has just a natural-gender opposition between persons and non-persons, but the morphology is still there. Adjectives preserve different class suffixes, but in free variation (as far as I can tell) instead of agreement with head nouns, apart from fossilised set expressions.

    However, some speakers still require agreement for a class which is very strongly associated in meaning with abstracts and liquids, as in “new beer”, where “new” still has to agree with “beer.” For them, there is a compulsory difference in the adjective agreement for the word for “thing”, depending whether the “thing” is concrete-and-countable, or not.

    I suppose what this amounts to is that grammatical gender distinctions can live on if they can be readily reanalysed as natural gender. Animate/inanimate and human/non-human are likely candidates, but abstract/concrete can happen too.

  84. This seems like a clearer case. In Norwegian it’s more like a vague tendency of misattribution that might turn out to be the first phase of something.

  85. Lars (the original one) says

    There is a similar case in West Jutlandic where the forms of the common gender are now used when something is countable and the neuter forms when it’s a mass — potentially with the same nouns. I haven’t found any specific examples of what happens with abstract nouns, but according to chapter 5.1 here they pattern with mass nouns.

    There is a similar phenomenon in Standard Danish where, regardless of noun gender, the neuter form of the indefinite determiner is used with nouns considered as a mass: noget kaffe ‘some coffee’ as opposed to kaffen ‘the coffee’. (In Danish this determiner cannot be used with singular count nouns and there is no gender difference in the plural, which is why it has to be contrasted with the definite singular form).

  86. Lars (and others): Newfoundland English, like the dialects of Southwestern England, likewise uses “he” as the pronoun replacing inanimate count nouns and “it” as the pronoun replacing inanimate mass nouns.

  87. David Eddyshaw says

    Count vs mass is maybe not so surprising as a newly minted “natural gender”, given that gender and number interact in a good many interesting ways anyway in many different languages. And that could easily spill over into concrete vs abstract: though all four combinations of count/mass and concrete/abstract are perfectly possible, there’s a pretty strong tendency for abstract to get assimilated to mass, or at least for countable abstracts to be less abstract in some sense: a question or a competition is not as abstract as “benevolence” or “zeugma.”

    (Having just written that, it occurs to me that this is probably a linguistic commonplace, but I never really thought about it in those terms before myself.)

    Apropos of that (more or less), I’ve read of ideas linking the Indo-European feminine singular with the neuter plural, supposing the latter to be originally a collective noun form co-opted as a plural (hence the Classical Greek singular verb agreement and all that); could that go together with the feminine-as-abstract-gender notion?

    David Marjanović will know …

  88. I always thought of the IE feminine as collective and the IE neuter as abstract, leaving the IE masculine as individual, at least in broad generalization.

  89. Lars (the original one) says

    I think that once upon a time I read that the noun classes in Niger-Congo languages were not originally aligned in neat singular/plural pairs, but had independent meanings; the sg/pl interpretation was generalised from cases like several sticks (one class) forming a bundle (another class).

    Could something similar have happened on the way from Pre-Pre-Proto-IE to PIE? A just-so story, ignoring Anatolian:

    Class 1 – Individuals / single things with agency >> masculine singular
    Class 2 – Groups with agency >> masculine plural
    Class 3 – Things without agency >> neuter singular
    Class 4 – Masses and collectives and abstracts >> neuter plural and feminine singular

    Class 1 and 2 forms a natural singular / plural pair, which pulls class 3 and 4 along but splits off the feminine gender from class 4 for nouns that want to stay singular, forming a feminine (nominative and) plural by analogy with class 2. And then we let more analogy smudge it all around for a thousand years or so to arrive at what we can see.

    Add hypotheses about class 3/4 not licensing ergative agreement if you wish, which might add motivation for the split of class 4.

  90. David Eddyshaw says

    noun classes in Niger-Congo languages were not originally aligned in neat singular/plural pairs, but had independent meanings

    While that’s perfectly possible, it’s very speculative, and would really need to be pushed back to a sort of Pre-Proto-Niger-Congo; indeed a good bit of the plausibility of the very existence of Niger-Congo as a real genetic thing rather than an Übersprachbund of some kind is that some of these pairings at least are shared everywhere the classes themselves can be recovered (for example Kusaal nid(a) “person”, plural nidib(a) matches the Bantu 1/2 pairing, as in Lingala moto “person” plural bato, while Kusaal tubur(e) “ear”, plural tuba(a) matches the Bantu 5/6 pair, as in the cognate Lingala word litoi “ear”, plural matoi. In Bantu, which has the system in its most developed glory, and Gur, which is what I mostly know about, regular pairings of particular class affixes are the norm rather than the exception, the more so as the system is better preserved in that particular language.

    In the extant Niger-Congo languages, moreover, there are certainly correlations between noun class and meaning, but in a lot of cases you need to squint pretty hard to see them, except with groups characteristically associated with people, liquids or abstractions, which tend to be pretty unmistakable. So attributing meanings to other individual affixes is very speculative indeed.

    Ideas about the origin of the pairings maybe arise from comparison of the core Niger-Congo languages like these Gur and Benue-Congo families with the less closely related groups like Atlantic and Khordofanian, but that’s just where speculation starts to outrun hard evidence as the correlations get harder to establish.

  91. Lars (the original one) says

    Clearly what I read was misrepresenting Niger-Congo or I misremembered what family it was. Athabaskan, maybe? The stick/bundle thing fits with what Wikipedia says about its classes. But I could be wrong about that too without making my just so story less so.

  92. David Eddyshaw says


    I wouldn’t be surprised if you had read something like that about Niger-Congo. Complicated gender systems cry out to be “explained”, and on first principles the N/C one might very well go back to some primordial system of once-unpaired classifiers which have learnt the joys of marriage, if not always of monogamy. But it’s like a lot of a deep-time speculation: perfectly plausible (and fun to think about), but probably destined to remain forever beyond the scope of either rigorous proof or thoroughgoing refutation.

    Also, there are people out there who know more about it than I do … there’s a lot of published stuff about semantics and the Bantu systems, especially.

  93. They cry “Proof! Proof!”, but there is no proof, at least not in science.

    “In science, ‘fact’ can only mean ‘confirmed to such a degree that it would be perverse to withhold provisional assent.’ ” —Stephen Jay Gould

  94. Right, but in these cases it is rare to have things confirmed to such a degree that it would be perverse to withhold provisional assent. That’s one reason I got out of historical linguistics (aside from the existential despair occasioned by grad school and teaching).

  95. That’s why demanding proof of linguistic hypotheses is Just Wrong, like demanding proof of historical ones. It’s asking for more than is humanly available.

  96. David Eddyshaw says

    There is no area of a posteriori knowledge where you could say it’s absolutely certain that we’re right (maybe not a priori either if you’re truly teh hardcorez.) That doesn’t mean that we can’t rationally say we’re a whole lot more certain in some areas than others.

    It’s a lot more secure that Gur and Benue-Congo are genetically related to each other than that the Mande languages also belong, for example. Rational believers in that relationship would at least agree with unbelievers that the evidence is shakier. Just because we don’t ever have absolutes in our sublunary knowledge doesn’t make it unreasonable to demand evidence.

    I agree that it’s harder in fields like historical linguistics than (say) physics to say where the evidence has become so cogent that disbelief argues for irrationality or unacknowledged or ulterior motives on the part of the unbeliever. But surely this is a difference of degree, not of kind?

    We might all enjoy a vigorous dispute about Altaic (say) here, but if someone posted that Welsh was unrelated to Urdu we would patiently explain to him that he was, in point of fact, Wrong.

  97. That’s why demanding proof of linguistic hypotheses is Just Wrong, like demanding proof of historical ones. It’s asking for more than is humanly available.

    What David Eddyshaw said. Once you start down the path of letting the rules of evidence slip because it’s just too frustrating, you wind up with Goropianism, Marrism, and all manner of nonsense.

  98. David E: Just so. I’ve just been reading the Essay on Human Understanding after David pointed out that he’s been quoting it. Sensible fellow, Hume, though like most philosophers too quick to say that something cannot be done just because he can’t see how to do it. We have learned a lot more of the “secret powers of matter” (like how it is that bread nourishes) than he thought possible.

    Hat: What Hume, the Davids, and I are saying is “Proportion the strength of your belief to the strength of the evidence.” There is no such thing as a fixed threshold above which the amount of evidence counts as proof, and below which the idea has to be rejected. I have only the weakest of belief in any sort of Nostratic, but it seems far more likely that the PIE and PU words for ‘name’ and ‘water’ are inherited from a common ancestor than that two separate stocks should borrow such words, and only such words, from one another at such a remote period. (Of course chance is more likely still, as with Japanese namae ‘name’).

  99. David Eddyshaw says


    I don’t think I’d quite agree with that summary of my position (flattering though it be to be included in such company. David M, anyway: Hume is a gleeful epistemological vandal.)

    I think that rather than a smooth rise in plausibility there’s very often a sort of catastrophe point in these things, beyond which it’s very difficult to accept that someone could in all honesty deny that the matter is effectively settled unless they have some private axe to grind, and I think it’s not doing violence to language to call that point by the hypocoristic name of “proof.” (For short.) Characteristically it’s where you get a whole lot of lines of evidence converging and reinforcing and illuminating each other in unexpected ways. This is just as possible in principle in historical linguistics as in physics, though the subject matter is more intractable and the triumphs perhaps less numerous. It’s easy to miss the successes, though, just because they are so uncontroversial: for every fanciful long-range hypothesis there are a dozen absolutely certain smaller genetic groups, to deny which simply means that you don’t know what you’re talking about … Indo-European, Semitic, Algonquian, Uto-Aztecan, Bantu … add your own examples.

  100. David Marjanović says

    Apropos of that (more or less), I’ve read of ideas linking the Indo-European feminine singular with the neuter plural, supposing the latter to be originally a collective noun form co-opted as a plural (hence the Classical Greek singular verb agreement and all that); could that go together with the feminine-as-abstract-gender notion?

    I’ll try to look things up on the weekend or soon after.

  101. JC: Again, I agree with David Eddyshaw. Your “weakest of belief” is exactly the sort of (booby) hatch through which wacky ideas can penetrate. I prefer to turn the coldest of shoulders to anything that can’t present a proper passport stamped with evidential visas.

  102. Stephen C. Carlson says
  103. David Eddyshaw says

    Fascinating paper – thanks. Very much apropos.

    Not persuaded by her actual argument (though that may well just reflect my lack of knowledge.) It seems to boil down to saying that the IE system cannot have arisen via development of the feminine from a collective because this implies an intermediate stage of the gender system which is typologically unattested; to get around the sonking great counterexample of Niger-Congo she cites Greenberg (who certainly counts as one of the people who knows more about it than I do, admittedly) speculating that the N/C system had an ultimate origin in classifiers. Even if this is true (and I’ve said what I think about it above) it doesn’t allow you to claim that real existing synchronic gender systems with no sex distinction but a distinction between mass and count are unattested. They aren’t. Her argument that PIE couldn’t have had such a system because it can only arise by grammaticalisation of classifiers depends crucially on the assumption that this is what has happened in N/C. It seems to me to border on circularity: I’ve no access to Greenberg’s paper, but I’d lay long odds that this will actually be part of his own argument for the origin of the N/C system, viz that a system like that must have arisen by grammaticalisation of classifiers. (As I said, this may very well be true, but I doubt whether it is really demonstrable.)

    (Anyway, PIE already has some pretty unusual typological features in its mainline incarnation. Paging Allan Bomhard …. Why not one more?)

    I’d forgotten about Kiowa. That’s a real poster child for interesting interaction between gender and number.

  104. David Eddyshaw says

    I might have to take the ultimate step of getting Alexandra Aikhenvald’s book on classifiers, which probably (based on the quality of her other work) answers all these questions as definitively as anyone ever will.

  105. David Eddyshaw says

    The stuff in the paper about the gender of deverbal nouns in German is interesting (p444).

    Vogel argues that in German deverbal action nouns are assigned different genders based on different degrees of individuation: neuter nouns indicate generic reference to a certain action (das Sitzen ‘the act of sitting’), feminine indicate specific instances of it (die Sitzung ‘the session’), while masculine most often tend to develop concrete meanings (der Sitz ‘the seat’).

    Kusaal has something rather similar (as I said, Kusaal has abandoned the agreement part of the noun class system, but the morphology is all still there.)
    Virtually all Kusaal verbs of the type capable of inflecting for aspect can make a deverbal abstract noun “gerund” simply by adding a noun class suffix to the verb stem, and the class is regularly predictable just from the length of the stem and its final segment. Similar formations with different class suffixes are regularly concrete instead of abstract:

    kugub(o) “resting on something” (regular formation with -bo)
    kuk(a) “chair” (with -ga)

    dugub(o) “cookery” (regular formation with -bo)
    duk(o) “cooking pot” (with -go)

    eenbug(o) “laying a foundation” (regular formation with -go)
    eenbir(e) “(physical) foundation” (with -re)

    It doesn’t work out quite as neatly as the German examples in the paper, though, because although words ending in -bo are never countable and nearly but not quite always abstract, those ending in -go are quite often concrete and/or countable. It seems to be the breaking of the regular class assignment that creates concrete derivatives rather than abstract, rather than the use of particular classes in itself.

  106. It seems to me to border on circularity

    Me too. Arguments from lack of evidence are never convincing anyway; you can say “that makes me suspicious,” but not “that proves it’s wrong.”

  107. I think it’s not doing violence to language to call that point by the hypocoristic name of “proof.”

    Well, I take your point; but you can still make a name for yourself by denying the Germanic origin of Yiddish, whereas the only name you get for your disproof of the Pythagorean Theorem is “crank”.

  108. JC: you can still make a name for yourself by denying the Germanic origin of Yiddish, whereas the only name you get for your disproof of the Pythagorean Theorem is “crank”.

    Where, among what kinds of people can you make a name for yourself with such “disproofs”, apart from that of “crank”? Perhaps there are fewer people familiar with Germanic and Yiddish than with the Pythagorean Theorem (which I suppose is taught in all high schools), but that is not the fault of either linguistics or mathematics.

  109. David Eddyshaw (and others):

    There is one basic difference between speculating on the origins of nominal gender in (Pre-)Proto-Indo-European and in (Pre-)Proto-Niger-Congo: the latter of the two language families remains grossly understudied. There remain a large number of languages which need to be fully and competently described, there remains a huge amount of work to be done on reconstructing various lower-level branches, and finally once that foundation will be in place it will be possible to forge a solid reconstruction of Proto-Niger-Congo. Similar work will also need to be done on other language families of Africa, and collaborative work between linguists working on African languages will be required to identify (especially early) areal influences and loanwords (within families and between different ones). Before all this work is done I think it is wildly premature to speculate on the genesis of the Niger-Congo noun class system.

    Considering the state of historical linguistics today, of course, it may well be that the scholarship required will simply never be done, and thus that we will never have as clear a picture of Niger-Congo as we have of Indo-European…

    Marie-Lucie: there is a linguist, who has published a good deal of serious scholarship, who has also quite seriously argued against Yiddish being Germanic in origin, and who remains quite respectable. If you write to me directly I’ll tell you more…

  110. David Eddyshaw says

    @Etienne: I am fully aware of this. I suggest you take the matter up with Greenberg.

  111. In Garifuna male speech, inanimate masculine nouns are given a feminine marker and feminine agreement to indicate plurality.

  112. David Eddyshaw says

    What happens in female speech?

  113. In female speech, only animate nouns are marked for number.

  114. Cranks: Gauss, Riemann, etc.

  115. David Eddyshaw says


    Eat your heart out, you unbelievers who deny the glorious possibilities of human language based on your myopic running-dog typological (bah!) preconceptions about what can be and what can not!

    (BTW the Semitic languages are obviously typologically impossible. I’ve always said so.)

  116. We’ve discussed Paul Wexler before, and yes, he isn’t considered a crank among Yiddishists: he holds a respected minority opinion. To clarify, nobody contends that the words of Yiddish aren’t Germanic: the dispute is whether that is a matter of common origin or massive relexification from some substrate language.

  117. David Marjanović says

    The stuff in the paper about the gender of deverbal nouns in German is interesting (p444).

    I still haven’t read the paper, but… this is a chicken-and-egg problem, isn’t it? Nominalized infinitives are automatically neuter, words in -ung are automatically feminine*, and bare roots used as nouns are… often masculine, though not always, I think. So do the meanings select the genders, which then select an appropriate method of derivation, or do the meanings simply select an appropriate method of derivation, which automatically comes with a gender that isn’t relevant to the issue at all?

    * Including such opaque examples as Zeitung “newspaper”. I happen to know that before newspapers existed, that word meant “report”, and one of the first newspapers in German was called Einkommende Zeitungen, “incoming reports/news”; but beyond the fact that it looks cognate with English tidings, I have no idea where the word comes from. I’m not aware of a verb from which it could be derived. Nowadays, “report” is Bericht m., from the verb berichten or perhaps the other way around.

  118. David Eddyshaw says


    I see what you mean, but Riemannian geometry is a generalisation which includes Euclidean geometry as a special case, rather than a refutation of Euclid. It’s the discovery (or construction, depending on your Platonism) of a new realm where Pythogoras’ Theorem need not be valid; it doesn’t invalidate it in the old realm. Same thing as with projective geometry, where parallel lines always meet …

    Either way, it’s part of a priori knowledge (if you believe in such a thing), where there is no place, in principle, for honest disagreement, but only for Correct and Mistaken. I understood John Cowan to be drawing a distinction between a priori and a posteriori when he opposed “Pythagoras’ Theorem” and the “Germanic Origin of Yiddish.” But he can (and will) speak for himself.

    I don’t think my poor monkey brain is actually physically capable of imagining a Riemannian linguistics. I’ll practice with Projective Linguistics.

  119. Zeitung is isolated because it is an etymologically nativized borrowing of Dutch tijding, which once also meant ‘newspaper’ but now retains only its earlier meaning of ‘news, message’ (there is a compound jobstijding ‘bad news’, lit. ‘Job’s tidings’).

    The OE verb tídan ‘happen, befall’ survives only in prefixed form as betide (impersonal), and apparently was lost altogether in German. I can’t figure out if it exists in Dutch.

    Berichten is a derived form of richten, so I should think that it came before Bericht.

  120. David Eddyshaw says

    @David Marjanović:

    Interesting. I was hoping you’d pick up on that.

    My supposedly parallel Kusaal examples are also potentially explicable by fairly humdrum mechanisms, thinking about it. The -bo suffix which is regular for two-mora stems is virtually always abstract. The only exceptions I’ve ever come across are “food”, “porridge” and “soap”, the first two of which at least are old deverbals from “eat” and “mash up” respectively, which have presumably acquired a concrete meaning secondarily. The -bo suffix occurs with longer stems in exactly one word, which is very evidently an analogical remodelling. In the fairly closely related language Nawdm, the cognate suffix -be is used with verb stems of any length, so it’s perfectly possible that in Kusaal (and its closest relatives) the use with gerunds from longer stems of other suffixes which are not invariably abstract in meaning is a secondary development, driven by some process that abolished all stems longer than two morae in the -bo class.

  121. David E: That was indeed the distinction I was drawing. However, I do not in fact believe in a priori knowledge: rather, 2 + 2 = 4 is an empirical generalization over our experiences with pairs: when we combine two pair of any two objects of which we have experience, we invariably have a foursome. To reject this is in principle always possible, just as with any other empirical generalization, but rejecting mathematical truths would cabbage our worldview so much that we simply don’t do it, which is what Quine called the “maxim of minimum mutilation”.

  122. JC: on the contrary, what came up in that blog entry is that Wexler’s opinion is in a minority of one, and that he is personally untrustworthy. Not a crank, maybe, but not respectable any more either.

  123. David Eddyshaw says

    nobody contends that the words of Yiddish aren’t Germanic: the dispute is whether that is a matter of common origin or massive relexification from some substrate language.

    There’s a somewhat analogous thing going on with Swahili, though in this case I don’t think any real mainstream linguist would subscribe to it. The idea is that Swahili originated as an Arabic creole, and has achieved its current very un-creole un-Arabic state by massive influence from the neighbouring Bantu languages.

    It’s perhaps unfair to call it a crank theory, particularly as, as Etienne rightly implies, proper historical linguistics is a comparatively fragile plant in Africa, so maintaining strange ideas about language history is more natural and forgivable than it would be in the context of Indo-European, say. Moreover, Swahili does in reality have some peculiar features (e.g. no tones, and lots of loanwords compared with most Bantu languages.) Nevertheless, the theory would surely have died the death long ago if it didn’t tap into some powerful wish-fulfilment ideas about the origins of the Swahili people themselves.

  124. Nobody contends that the words of Yiddish aren’t Germanic: the dispute is whether that is a matter of common origin or massive relexification from some substrate language.

    I have not studied Yiddish but have seen many examples of it in linguistic literature. I did study German in school many years ago, and the Yiddish samples I have seen look very German to me: not just lexical items, but pronouns, noun and verb endings, and other grammatically significant features. If Yiddish was only relexified from some unknown substrate language, one would expect at least some of these sorts of features to have survived in Yiddish and to be identifiable because of their strangeness. Historically speaking the obvious substrate would be Hebrew or Aramaic, but as far as I know the Hebrew component of Yiddish is lexical, not grammatical. What am I missing?

  125. David Eddyshaw says

    That’s really what I was driving at in saying that the Swahili-as-Arabic-creole thing was analogous. It doesn’t really stack up that that a language would adopt not only all the basic lexicon but also all or nearly all of the morphosyntax of a lexifier language. Real creoles aren’t like that.

    The nearest approach to something along those lines I know of is Lingala, which is basically what Esperanto would have been if Zamenhof had been from the Congo instead of Poland and thought that tones and noun classes were natural features of language instead of consonant clusters and cases (and drawn his lexicon from assorted Bantu languages instead of European, of course.) It arose in really just the same sort of way as a pidgin, but it has the whole Bantu paraphernalia of noun classes, just without any agreement, and the whole Bantu ultra-productive verb derivation system, making applicatives and causatives ad lib. Presumably because most of its parents were pretty similar Bantu languages, it ended up as Congolese Esperanto instead of as evidence for an imaginary bioprogram.

  126. What am I missing?

    That while Wexler’s theory is obviously wrong to anyone who understands historical linguistics, Wexler’s reputation doesn’t seem to suffer because he proclaims it (at first I wrote “holds it”, but nobody actually knows if he holds it or not). This however is a digression from my original point, and I’m not sure why I went off on it.

    David E is right to say that the proportion between strength of evidence and strength of belief is not linear: that eventually it becomes as hard to reject the “sphericity of the earth” hypothesis as it does to reject the “addition is commutative” hypothesis.

  127. David Eddyshaw says

    Real creoles aren’t like that.

    Actually, I suppose Copper Island Aleut and Michif are a bit like that, at least showing as proof-of-concept that morphosyntax can be borrowed wholesale. Still, Yiddish is nothing like either of those. Nor is Swahili.

    Thinking about it, I suppose you could maintain that it’s really all a matter of degree, or even of terminology. If a group of people once spoke A but now speak B, how similar does B have to be to A (and in what respects?) before you call B the “same” language as A, after massive influence from some B-like language, as opposed to a language unequivocally descended from Proto-B but with an A substratum?

  128. David Eddyshaw says

    Reminds me of the snark (unfortunately I forget whose) that the language war in Israel between Yiddish and Modern Hebrew was a struggle between two Slavonic languages.

  129. -Copper Island Aleut

    I have a suspicion that it arose as a result of an attempt by fully Russophone population to speak their half-forgotten ancestral language. This would tend to produce exactly the effects we see.
    Why this or that grammatical feature was borrowed from Russian? Because speakers have forgotten how to use native forms correctly.

  130. David Marjanović says

    Silvia Luraghi has a number of papers on PIE gender on Academia. Here’s one: https://www.academia.edu/2142822/_The_origin_of_the_Proto-Indo-European_gender_system_Typological_considerations_._In_Folia_Linguistica_45_2_2011_435_464

    I’ve finally started reading it, and I’m liking it so far; the paper I had in mind earlier is probably Luraghi’s from 2009.

    However, on p. 444–445:

    Regarding gender assignment to concrete entities in German, Zubin and Köpke (1986) observed that neuter gender is systematically assigned at the superordinate level of categorization, while masculine and feminine are assigned at the basic level and below. Thus, at the highest level of genericity, the word for ‘animal’, Tier, is neuter. At the next level, we find the words for ‘mammal’, ‘reptile’ and ‘insect’, also neuter (Saugtier, Reptil, Insekt), while at the basic level of representation we find masculine or feminine, e.g. der Hund ‘dog(masc)’ or die Katze ‘cat(fem)’. Assignment of an entity to a certain level may depend on different taxonomies, including perceptual and functional.

    …followed by more waffling for the rest of the paragraph.

    Yes, Tier is neuter. But it’s still neuter when it means “female red deer” in hunters’ jargon. Also, Pflanze is not – it’s feminine, unchanged from the Latin original planta.

    “Mammal” is neuter because it’s a compound ending in Tier; and the fact that it’s Säugetier, not from saugen “suck” but from säugen “suckle”, might tell us something about how carefully the argumentation was researched.

    That the Latin words reptile and insectum are neuter tells us nothing about German. In Linnaeus’s Latin they’re neuter because they’re adjectives agreeing with animal.

    “At the basic level of representation” we do indeed find plenty of masculine and feminine words. But neuters are by no means absent: Schwein “pig”, Schaf “sheep”, Rind “cattle” (much more commonly used than in English)…

  131. Säugetier is certainly standard, but there are a few hundred ghits for Saugtier in technical and semi-technical contexts as well:

    (context unknown) “Windhund ist fähig Geschwindigkeit bis zu 60km Stunde beim vollen Lauf zu erreichen und ist damit das 18. schnellste Saugtier der Welt und ein geborener Renner.”

    (from a patent application) “Verfahren zwecks der in-vitro-Diagnose der Neigung eines Saugtieres einen Krebsprozess in einem Zielorgan-oder Gewebe zu entwickeln, wobei das genannte Verfahren die Bestimmung des Abdruckprofiles der aus dem genannten Organ oder Gewebe stammenden Zellen beinhaltet.”

    (song lyric)
    Ach Mensch, du Saugtier
    Die Erde, die arme
    Ach Mensch, deine Sauggier

    (of polar bears) “Er ist ein Saugtier und ein Fleischfresser; er hat weisses Fell er ist gefarlich.”

  132. I looked at some of the technical / semitechnical contexts, and many look like bad translations from English or like something written by non-native speakers to me (many of them have language issues besides Saugtier). That includes the greyhound site, which has atrocious Geman overall. In other cases the word doesn’t actually mean “mammal” but “animal with suction cups” or “suckerfish”.
    The song lyrics are just artistic creativity; the compound clearly is formed to mean “sucking animal”, not “suckling animal”; it is wordplay alluding to Säugetier but meaning something different. I’ve found some more cases of such wordplay (e.g. Saugtier as nick- or usernames), and a few cases where it is used for irritating or abusive people – there it’s the same metaphor as in English “to suck s.o. dry”.
    So, in short, Saugtier doesn’t mean “mammal”, and where it’s used in that meaning it’s most probably a typo or a mistake by a non-native speaker.

  133. David Marjanović says

    er hat weisses Fell er ist gefarlich

    it has wite fur it is daunjerous

  134. David Marjanović says

    I forgot an interesting case of a neuter animal name in German. First, Ross, the cognate of the metathesized horse, is neuter. Second, the nowadays usual word Pferd is neuter; that must be copied from Ross, because Pferd is from a late Latin paraveredus, which was obviously masculine.

  135. David Marjanović says

    the paper I had in mind earlier is probably Luraghi’s from 2009

    Yes, this is the paper that postulated a transitional 3-gender system of human, abstract and concrete-inanimate.

    It’s not the paper that actually proposed reflexes of a verb root from which *|gʷenh₂| could be derived, however; it merely cites Brugmann’s idea from 1891 that such a verb might once have existed.

  136. David Marjanović says

    Long but interesting: this book chapter on intermediate stages of agreement as might have occurred during the development of gender.

Speak Your Mind