Malchukov on Grammatical Case.

I hadn’t been familiar with Serious Science, which according to its About page is “an independent, non-profit, non-governmental project whose main mission is to spread knowledge and share the greatest advances in various realms of academia with our readers.” One of the pages on the site is called Grammatical Case: A Deceptively Simple Concept by linguist Andrej Malchukov; the “deceptively simple” of the title presumably means “you might think it’s simple, but it’s not”:

A textbook definition of (morphological) case defines it as marking of a noun (or noun phrase) for its syntactic and/or semantic role. Thus, the use of accusative case on the noun identifies it as a (direct) object of the verb, and instrumental case on the noun (as found for example, in Russian) most commonly marks instruments. While the definition as such is fairly clear, it is more controversial to what extent this definition case be extended to other similar patterns. If, say, agreement on modifiers (adjectives agree in cases with nouns in Russian) is the same phenomenon, as the case on nouns? Or does, the vocative case, as found in Classical Greek and some Slavic languages, represent case, even if it does mark dependency, but is rather means of address? And, indeed, does English have case, if it is only visible in pronouns?

More problems will be found if we extend the discussion beyond European languages. Does Japanese have cases? This is not obvious, as the morphological status of these forms is unclear; traditionally they are called postverbal particles. Moreover, some of them such as the subject (“nominative”) markers ga and wa additionally express discourse meanings (focus vs. topic, that is new vs. given information). Another controversial example comes from Even (a Tungusic language I did fieldwork on), which has a special case called “designative’’. Its use is illustrated by the following example: Bej hin turki-ga-s emun [man your sledge-DES-your brought] ‘The man brought the sledge for you’. The designative case is similar to accusative in that it marks the direct object, but, in addition, reinterprets the possessor as a beneficiary (the meaning of the example above is ‘brought a sledge for you’, not ‘brought your sledge’). Is it OK for a case to mark different roles of different arguments? Does it still qualify as case, or maybe, is a marker of prospective (future) possession? There will be different answers depending on the definitions of case.

There are sections on Case Systems, Short History of (research on) Case, Functions of Case, and Evolution of Case Systems. I’m glad to see this kind of information presented in an accessible manner online.

Comments

  1. David Eddyshaw says:

    There are quite a number of interesting articles; they’re aimed at interested non-linguists, and written by real experts. I liked the ones on Eskimo by Michael Fortescue and on Australian languages by Nicholas Evans; the latter has a bit on Kayardild, which would be a strong contender for Weirdest Language EVAH (I am the proud owner of a copy of Evans’ grammar in the Mouton Grammar Library series.) I’d missed the fact that the egregious Pinker in the 1990’s made one of his sweeping statements about possibilities in human language which is directly refuted by Kayardild.

    (The assertion was ‘‘No language uses noun affixes to express tense.’’ Kayardild does exactly that. Using unequivocal case suffixes, no less, not affixes just meaning “a former …” or “a future …”)

  2. Ha! Take that, egregious Pinker!

  3. “(The assertion was ‘‘No language uses noun affixes to express tense.’’ Kayardild does exactly that. Using unequivocal case suffixes, no less, not affixes just meaning “a former …” or “a future …”)”

    Good catch. This is not unusual in American languages. I think South Sierra Miwok did this and so probably did other related languages. And of course if you insist that verbal and nominal roots are inherently distinct, you have Salishan languages that mark nouns for tense (with exactly the same affixes as they use for verbs…)

  4. David Eddyshaw says:

    Bininj Gun-wok, yet another Australian language of which Nicholas Evans has written a suberb grammar, also inflects nouns for tense, using a subset of the verbal tense suffixes, e.g.

    Gorrogo al-wanjdjuk bininj-ni.
    before II-emu person-P
    “Long ago, emu was a person.”

    Kayardild goes one better: it really does use noun case endings to mark tense, and not content with that, also uses verb tense endings to mark case. (Really and truly …)

    http://serious-science.org/australian-languages-7356

    All this in a language with a robust distinction between nouns and verbs, moreover (it’s not at all like Salish.)

  5. PlasticPaddy says:

    The Evenki are clearly an optimistic people. Otherwise they would have created an anti-designative case expressed instead and more cumbersomely with the preposition “on” in sentences such as “She hid it on me, the $%!#&” and “He broke the $%!#& yoke on me, after $%!#& swearing that he only needed to borrow it for a few minutes.”

  6. Judith Tonhauser studied tense marking on nouns, centered around Guaraní.

  7. David Eddyshaw says:

    @Plastic:

    Kolyma Yukaghir (as we’re in Siberia now) has a “predicative” case, used for nominal predicates (duh), but also in subject or object roles where the noun is focussed.

    A lot of languages can tense-mark nouns used as predicates, but it’s less common to be able to tense-mark nouns used as verb arguments. There are plenty of languages that can, though: another one is the Amazonian language Jarawara, described by another of those Australian grammar-writing sexy machines, RMW Dixon:

    ee kaa hemejo-ba fonai mata ne-ba ee-ke
    1inc POSS medicine-FUT FUNAI send AUX-FUTm 1inc-DECf
    “FUNAI will send the medicine intended for us.”

  8. John Cowan says:

    And of course we have the prefix former and the suffix -to-be, which nicely mark noun tenses in English. Lojban has noun tenses because its nouns are verbs (except foreign proper names), but more interesting to my mind is that it has both temporal and spatial tenses. Thus “John at-here sings to Mary” is a spatial present tense, formally equivalent to “John at-this-time [i.e. now] sings to Mary.” Of course, there are only two directions in time, pastward and futureward, whereas there are a dozen or more in space, which can be compounded: “John leftward [and then] upward [of me] sings to Mary”.

    Further, there are particles indicating distance from the speaker: the remote past is formally paralleled by the remote north, and one can even express distance without direction: “a short way from here” is parallel to “a short time from the present, in the past or future according to context”. Even the event contours like perfective, inchoative, and so on have spatial equivalents: “That is the south-facing side of a rock north of me” can be expressed using the spatially-marked initiative: the rock “begins” at that point and “ends” further north. If you wanted to speak of the region closer to you than the rock, you could use the inchoative, and if the rock stretches further than expected, then its extension, considered as a location, could be expressed by the superfective.

  9. Le Cracquere says:

    I badly want English to innovate grammatical cases to replace what it’s lost. Very badly.

  10. January First-of-May says:

    The designative case is similar to accusative in that it marks the direct object, but, in addition, reinterprets the possessor as a beneficiary (the meaning of the example above is ‘brought a sledge for you’, not ‘brought your sledge’).

    …It might be interesting to compare how different languages distinguish those situations. Russian, for example, would use something to the effect of “brought you.DAT sledge.ACC”.

    (I vaguely recall having seen a discussion in a popular linguistics book about how this construction is also used in “the barber trimmed my beard”, and a literal possessive translation would be understood as either an error or perhaps implying that the barber wasn’t supposed to do that.)

  11. It might be interesting to compare how different languages distinguish those situations.

    One way:

    The applicative voice (abbreviated apl or appl) is a grammatical voice that promotes an oblique argument of a verb to the core object argument. It is generally considered a valency-increasing morpheme. The Applicative is often found in agglutinative languages, such as the Bantu languages[1] and Austronesian languages.[2] Other examples include Nuxalk, Ubykh, and Ainu.

    https://en.wikipedia.org/wiki/Applicative_voice

    (And now I have a bad case of One Way or Another earworm.)

  12. @David Eddyshaw, it (Evans) is fun to read for a Russian.

    Consider “my brother’s wife”. Here you take “my brother” and put the genitive case on each word of it – zhena mojego brata. If, in a language like Russian,I then say “I give the book to my brother’s wife”, I put “wife” in the dative, but I leave “brother” in the genitive: zhene mojego brata.

    He is code-switching! He introduced an example without warning what langauge it would be in. But again too much Chomsky. WHAT he is doing there? There is search for universals in typology (Greenberg: “All phenomena which occur with signigicantly more than chance frequency in languages in general are of potential psycholinguistic interest”) and it is interesting.

    P.S.The myth of language universals: Language diversity and its importance for cognitive science, (Evans and Levinson:

    Keywords: Chomsky coevolution constituency culture dependency evolutionary theory Greenberg linguistic diversity linguistic typology recursion universal grammar
    1. Introduction

    According to Chomsky, a visiting Martian scientist would surely conclude that aside from their mutually unintelligible vocabularies, Earthlings speak a single language.
    — Steven Pinker (1994, p. 232)

    :-/

    P.P.S. By too much Chomsky (he is mentioned only once) I mean: hypotheses about lingusitic rarities and universalia, they are fun. Counter-examples are fun too. Both things are good. But then add “Chomsky” and …:-//

  13. David Eddyshaw says:

    Evans only invokes Chomsky as an Awful Warning.

  14. Basic Verbs of Possession
    A contrastive and typological study
    Åke Viberg

    Verbs of possession such as HAVE and GIVE have been extensively studied both typologically and from a cognitive linguistic perspective. The present study presents an analysis of possession verbs as a semantic field with a focus on the most basic verbs. It combines a corpus-based contrastive analysis with a sketch of a general lexical typology of possession verbs. The contrastive part consists of an analysis primarily of the Swedish verbs ge ‘give’, få ‘get’ and ta ‘take’ and their correspondents in some genetically and/or areally related languages. Data are taken from two translation corpora, the large English Swedish Parallel Corpus (ESPC) and the Multilingual Pilot Corpus (MPC) consisting of extracts from Swedish novels and their published translations into English, German, French and Finnish. The study of ta is concerned in particular with the relation between the many concrete uses of the verb, which are based on the interpretation of taking as a goal-directed action sequence. The account of Swedish ge ‘give’ and få ‘get’ are brief summaries of earlier studies concerned with patterns of polysemy and grammaticalization. In particular the verb få ‘get’ has a complex and relatively language-specific such pattern including modal, aspectual and causative grammatical meanings. The meanings GIVE, TAKE, GET and HAVE are all realized as verbs with very high frequency in the Germanic languages. This appears to be a rather language-specific characteristic. The typological part presents a tentative typology and gives a brief overview of some of the ways in which the corresponding meanings are realized in languages that are not included in the corpus.

    https://journals.openedition.org/cognitextes/308?lang=en

  15. PlasticPaddy says:

    @juha
    The Swedish and English verbs for take are independent, if at all, only in a typological sense, because take replaced an earlier English verb under the influence of Scandinavian. It can even for me not be excluded that English borrowed the typology as well. The earlier verb seems to have emphasized not the seizing aspect of taking but the cooperative “accepting” aspect, as in taking advice or taking part, so does not appear to be related to a hand motion.

  16. David Eddyshaw says:

    Western Oti-Volta is actually boringly Germanic in this respect, though e.g. Kusaal mɔr is probably derived historically from a verb meaning “hold” (cf Nawdm mada “hold”, which corresponds exactly etymologically.)

    One difference is that the commonest “take” form is not derived from anything to do with hands, but from eating:

    O di pu’a.
    3sg eat woman:SG.
    “He’s got married.”

    O di na’am.
    3sg eat chieftaincy.
    “He acquired the chieftaincy.”

    Ba di nyan.
    3pl eat shame.
    “They have been disgraced.”

    BTW Sango is definitely a creole. Viberg seems unclear about this.

  17. David Marjanović says:

    anti-designative case expressed instead and more cumbersomely with the preposition “on”

    Dative. Emphatic dative.

    …though Basque has a special benefactive for that, perhaps unsurprisingly.

    And of course we have the prefix former and the suffix -to-be, which nicely mark noun tenses in English.

    That’s much more restricted. In the Jarawara sentence above, the future marker -ba is both on “will” and on “medicine”. The Nature or Science article where I read about Kayardild lo these onescore years ago said that in a sentence like “the man speared the fish”, the verb “spear” is marked for past tense, and so is the noun “fish”.

    this construction is also used in “the barber trimmed my beard”, and a literal possessive translation would be understood as either an error or perhaps implying that the barber wasn’t supposed to do that.

    Middle and Modern English are actually very unusual in IE for not doing that.

    applicative voice

    German be- is applicative – but it’s derivational, not inflectional.

  18. David Eddyshaw says:

    Bantu languages are extremely fond of applicatives, and often form them with such total freedom that it’s a bit moot whether you should call them derivational or flexional. So (from Ashton’s Swahili grammar)

    Nikupikie chakula?
    “Shall I cook some food for you?” (pika “cook”)

    Watoto walituimbia nyimbo.
    “The children sang songs to us.” (imba “sing”)

    Even Lingala, the Esperanto of the Congo, maintains highly productive applicative formations:

    Tíndélá ngáí mokandá.
    “Send me a letter.” (tínda “send”)

    Lingala doesn’t mess with all that incorporation of object pronouns into the verb like Swahili, but still makes the verb applicative before the “indirect object” ngáí “me.”

  19. David Eddyshaw says:

    (In other words, Bantu languages use verb applicatives rather than nominal datives.)

    Many West African languages use serial verb constructions instead of “datives”: thus in Kusaal (although strictly speaking this is a kind of clause subordination in Kusaal, rather than a serial verb construction stricto sensu):

    M kuos bʋŋ la n tisi ba.
    1sg sell donkey:SG ART CAT give 3pl
    “I’ve sold the donkey to them.” (there is no question of having given it away.)

    In Kusaal, though, you can put an indirect object after any verb at all, transitive or intransitive, the only formal marking being that it must precede all other objects, so you could equally say

    M kuosi ba bʋŋ la.
    “I’ve sold them the donkey.”

  20. Kayardild

    Kayardild Morphology, Phonology and Morphosyntax
    A Dissertation
    Presented to the Faculty of the Graduate School
    of
    Yale University
    in Candidacy for the Degree of
    Doctor of Philosophy
    by
    Erich Ross Round
    2009

    https://ling.yale.edu/sites/default/files/files/Round_Kayardild_morphology_phonology_and_morph.pdf

  21. I’ve read Malchukov’s nice, but extremely brief grammar of Even written in English. He wrote ten times bigger “The Syntax of Even language” in Russian based on his PhD dissertation. It’s table of contents (in Russian):

    https://www.dissercat.com/content/sintaksis-evenskogo-yazyka-strukturnye-semanticheskie-kommunikativnye-aspekty

    Couldn’t find anything on Even designative case there.

  22. David Marjanović says:

    “Shall I cook some food for you?” (pika “cook”)

    German has jmd. bekochen “cook for someone”, though it doesn’t seem to be in anyone’s active vocabulary, and I’ve only read it in disparaging contexts.

    “The children sang songs to us.” (imba “sing”)

    besingen : singen :: canere : cantare… though “sing someone’s praises” isn’t really in anyone’s active vocabulary either.

    “Send me a letter.” (tínda “send”)

    Beschicken exists, so *beschicke mich mit einem Brief wouldn’t sound wholly impossible and could easily be an inside joke, but actually it refers to filling an oven in metallurgy…

    1sg sell donkey:SG ART CAT give 3pl

    Reminds me of Chinese: “for” = “give”, “in” = “be in/se trouver à“, “with” = “accompany”…

    A Dissertation

    Thanks, downloaded. 🙂

  23. Lars Mathiesen says:

    Vreden, gudinde, besyng, som greb peleiden Achilleus! — and yes, it’s archaic. A newer translation has

    Syng os gudinde om vreden der greb Peleiden Achilleus
    vreden den fæle som voldte Achaierne tusinde kvaler,
    sendte behjertede sjæle af talrige helte til Hades
    og lod dem selv blive slængt som æde for hurtige hunde,
    grådige gribbe og ravne så Zeus´s vilje blev fuldbragt
    helt fra den første stund da striden begyndte imellem
    folkenes drot Agamemnon og gudernes lige Achilleus.

  24. on Australian languages by Nicholas Evans; the latter has a bit on Kayardild

    Nice four-minute talk by Evans about how he came to be a linguist and study Kayardild.

  25. Ba di nyan.
    3pl eat shame.
    “They have been disgraced.”

    “Eat humble pie” sounds very Kusaal.

  26. Nicholas Evans: “Yilahdjarrkkarrewaniyan: Following the leg into the Dalabon meaning-world”
    https://www.youtube.com/watch?v=ccWIgDtpRZ8

    (Does he really say Eljin, as in the Elgin Marbles, to pick up on the gyre/gynecology?)

  27. David Eddyshaw says:

    @Hat:

    Thanks for that!
    (Evans has an engaging mad-scientist vibe in the video.)

  28. PlasticPaddy says:

    @sfr
    Try the Oxford Handbook of Case (edited by Andrej L. Malchukov and Andrew Spencer), ch. 36, p. 549-561.

  29. @DM: As some of your examples already illustrate, many verbs formed with be- have been lexicalised with specific meanings, so one needs to know which ones can be used freely as applicatives.

  30. PlasticPaddy says:

    @hans, dm
    Some of these be- forms seem to be ways of “verbing” nouns, e.g., beeinflüssen, benachrichtigen. And what about Begabung? Is the corresponding verb meaning obsolete?

  31. Nicholas Evans: “Yilahdjarrkkarrewaniyan: Following the leg into the Dalabon meaning-world”

    Thanks for that; I only listened to the first half-hour or so, but it was full of good stuff. For those who are curious about the title (the discussion of it starts at 17:45 if you want to know more), it breaks down as yila-h-djarrk-karrû-wa-niyan, where the main morphemes are -wa- ‘follow’ and -karrû- ‘calf (of leg); leg; “leg” of song cycle; song; culture, ancestral law of group; voice’ and the yila- is an exclusive ‘we’ (doesn’t include the addressee). I liked the word shibbolethonyms for language names based on distinctive words (e.g., Dalabon = dala ‘mouth, language’ + bon ‘go,’ where other nearby languages use different words for ‘go’).

    (Does he really say Eljin, as in the Elgin Marbles, to pick up on the gyre/gynecology?)

    Didn’t get that far, but if he does, there’s nothing wrong with that — lots of people do. The fact that Elgin in Scotland is /ˈɛlɡɪn/ and therefore Lord Elgin said it that way doesn’t determine the pronunciations of derived terms (e.g., Elgin /ˈɛldʒɪn/, Illinois), any more than gerrymandering /ˈdʒɛrimændərɪŋ/ has to start with /g-/ just because Elbridge Gerry’s name does.

  32. David Marjanović says:

    beeinflüssen

    beeinflussen with u, no ex-causative umlaut. But sure, to influence someone is to apply influence to them, and to inform someone is to apply a message or news to them.

    And what about Begabung? Is the corresponding verb meaning obsolete?

    In the present tense it may never have existed; how often do you talent someone? But begabt “talented” is current.

  33. In Swedish the verb “begåva” exists, and it means both “donate” and “be talented”. I guess like the English verb “gift”? You can be gifted something, and you can be gifted (intelligent/talented). “Begåvning” means talent. The meaning “donation” is obsolete, and sounds weird today, even though it’s easy to see the connection.

    I’ve been looking for examples, and in Swedish you can be gifted health, a voice, a sense of humour, business ideas, teachers, and so on. A city can be gifted a new shopping centre and I even found a sarcastic comment about being gifted a headache. I’m not sure what the corresponding German expression is.

  34. @languagehat: The Elgin National Watch Company was named after Elgin, Illinois, and the town of Elgin, South Carolina is named after the Elgin National Watch Company. After nearly a century in Illinois, the company moved their factory to Blaney (previously Jeffers), South Carolina in 1963, and the town renamed itself in honor of the big new employer. The move was part of an effort by the company to save money, in the face of much stiffer competition from resurgent European watchmakers. Elgin is an exurb of Columbia, South Carolina, but at the time it was far enough away from the urban core that the land there was incredibly cheap (just as the land the company had bought around Elgin, Illinois had been in the 1860s).* However, in spite of the low costs of the land and labor in the area, the Elgin factory only lasted five years, and the company went out of business in 1968.

    Although the authentic Lord Elgin pronounced his name with a /g/, the Elgin National Watch Company sold some of their watches under the brand names “Lord Elgin” and “Lady Elgin” (the latter, obviously, for their higher-end ladies’ timepieces), and naturally, the company used the /dʒ/ pronunciation in these brand names.

    * Although the city limits of Columbia and its inner ring suburbs have not moved appreciably in the intervening decades, there had been a tremendous amount of urban sprawl. There is an uninterrupted urban environment running northeast along Route 1 and Interstate 20 past Pontiac, South Carolina (which used to be exurban) and almost to Elgin (which is the next town after Pontiac along in that direction).

  35. besingen : singen :: canere : cantare

    The two Latin verbs are basically synonymous, as far as I can tell. You can canere / cantare carmina, and Horace has nauta absentem cantat amicam.

    English has the “applicative” be- in a handful of verbs too — bewail, bedaub, besprinkle… (Le Guin: “And then comes the final test, the infallible touchstone of the seventh-rate: Ichor. It oozes out of severed tentacles, it beslimes tessellated pavements, bespatters bejeweled courtiers, and bores the bejesus out of everybody.”)

  36. @Paddy: What be- does is create transitive verbs. It can be used to create verbs from nouns, it can make intransitive verbs transitive (leben – > beleben), and it can be used to change the valency of transitive verbs, demoting the direct object of the unprefixed verb and promoting an indirect object or a prepositional complement to direct object. In some cases, the simple verb doesn’t exist any more; some of these cases look like nouns having been verbed, but I can’t remember any examples right now.
    (There is also a very small number of intransitive verbs with the prefix, like bestehen “exist”.)

  37. Although the authentic Lord Elgin pronounced his name with a /g/, the Elgin National Watch Company sold some of their watches under the brand names “Lord Elgin” and “Lady Elgin” (the latter, obviously, for their higher-end ladies’ timepieces), and naturally, the company used the /dʒ/ pronunciation in these brand names.

    There you go!

  38. A couple of my high school friends and I went through a period of making up new English verbs starting with be-: befucker, befungus, and many others that weren’t as funny.

  39. John Cowan says:

    any more than gerrymandering /ˈdʒɛrimændərɪŋ/ has to start with /g-/ just because Elbridge Gerry’s name does.

    (very quietly but with enormous strength) But it does move start with /g/.

  40. beverb looks useful

  41. Middle and Modern English are actually very unusual in IE for not doing that.

    Not IE but SAE: Celtic, Armenian, and Indo-Iranian don’t have external dative possessors, but European Uralic languages do.

    The two Latin verbs are basically synonymous, as far as I can tell.

    There’s a general trend for a Latin non-first-conjugation verb in -ere, -ēre, -īre to be given a derivational ending that makes the result a first-conjugation verb in āre, and then to be bleached back to the original meaning. It’s still the case in the Romance languages today that (barring a very strong analogy) neologistic verbs end in -ar(e).

  42. Elgin/Eljin

    I misremembered that part. It was Nicolas Thieberger’s lecture:

    Nicolas Thieberger: “Before us, the deluge: What is the path to better records of more languages?”

    https://www.youtube.com/watch?v=t1RV0-fZ4u4

    Back to Nick Evans. I did notice that he said deCOrous, which is not quite usual but not unheard-of.
    And very interesting stuff indeed!

  43. I say DEKorous, because I think of ‘decorate’.

    deKOrous would be patterned on ‘decorum’.

  44. after us, however, the delu/g/e!

    or possibly even the beluga.

  45. Lars Mathiesen says:

    external dative possessors — I don’t think we have any of those around here either. There might have been some in ON, though, it sounds familiar. (Possibly 800 years without accusative/dative distinctions did them in). Spanish goes all in, of course, much more than for instance French, but then it still has the distinction in pronouns.

  46. I just noticed that Malchukov spells his first name as Andrej.

    I wonder how Americans pronounce it.

  47. David Eddyshaw says:

    Andreg.

  48. David Eddyshaw says:

    external dative possessors

    Western Oti-Volta, too, is like English and not (say) French in this regard; thus in Kusaal

    Fʋ piesidi fʋ nu’us.
    “You wash your hands.”

    Not *Fʋ piesidif nu’us. *”You wash you the hands.”

  49. “I just noticed that Malchukov spells his first name as Andrej.

    I wonder how Americans pronounce it.”

    speaking from experience I can tell you that Americans inevitably say the hard j, and that’s so even in cases like Andrej and Sergej where the name has an obvious English equivalent.

  50. I suspected so.

    Didn’t expect /sɜːdʒɛdʒ/ though.

  51. David Eddyshaw says:

    I know a Russian Sergey here who has long since made his peace with being called Sir Guy.

  52. David Marjanović says:

    There is beschenken “give unspecified gifts to someone”, from schenken “give as a gift”.

    And then there’s sich nach … begeben “go to”.

    The two Latin verbs are basically synonymous, as far as I can tell.

    Ah, I’ve been lied to. I’m not surprised, though, because morphologically cantare is the frequentative of canere, not something detransitivized.

    bestehen “exist”

    Mostly aus … bestehen “consist of”.

  53. beschenken – yes, that’s one of the clearly applicative cases (also a derivation where the indirect object of the unprefixed verb is promoted to direct object).
    sich begeben also, in an even more archaic register, means “to happen”. One of those cases with a lexicalised meaning that is hard to deduct from the simplex.
    bestehen without “aus” frequently in formulaic expressions like es besteht die Gefahr / die Hoffnung / die Möglichkeit etc.

  54. Lars Mathiesen says:

    Men det begav sig i de dage, … They changed it now.

  55. David Marjanović says:

    Ah, yes. Es begab sich in jenen Tagen…

    frequently in formulaic expressions

    Also with further prefixes, like weiterbestehen “continue to exist”.

  56. Just after reading these comments I went out for a walk and saw that some scholar and fan of the dative had graffitied on a nearby wall „lutsch mir“ ( suck me DAT). I instinctively supplied a missing „den Schwanz „ but then got to wondering if one could say „belutsch mich“ in the applicative sense, even if just jocularly.

  57. If that was in Berlin, lutsch mir could be a direct translation of “suck me”, because in Berlinerisch mir indicates both direct and indirect object.
    And yes, belutsch mich is possible as a jocular utterance.

  58. H. Rider Haggard used “bepraise” twice in Eric Brighteyes:

    Now, it is in my mind to send Koll the Half-witted, my thrall, whom Asmund gave to me, to Ospakar as though by chance. He is a great talker and very clever, for in his half-wits is more cunning than in the brains of most; and he shall so bepraise Gudruda’s beauty that Ospakar will come hither to ask her in marriage; and in this fashion, if things go well, thou shalt be rid of thy rival, and I of one who looks scornfully upon me.

    Although this occurrence is obviously in dialogue (by the witch Groa), Haggard uses the same word in the narration to describe what Koll actually does when he reaches Ospakar’s steading in the North of Iceland.

    Haggard also provided an interesting introduction to Eric Brighteyes, in which he praises the Norse Sagas as an underappreciated literary canon among the British (second only to the Illiad and Odyssey among epic literature, in his view). He wrote Eric Brighteyes as something halfway between the Icelandic saga style and a modern novel. It has less of the repetitions and digressions common in the originally oral epics, but it also maintains something of all archaic styling and diction—hence bepraise.

    The word bepraise did (according to Google Ngrams) enjoy a relative burst of popularity in the early to middle nineteenth century. But even at its most popular, bepraise was thousands of times less common than praise (whereas it is closer to a million times less common today). By the time Eric Brighteyes was published in 1890, the word’s frequency had fallen steeply from its earlier peak.

    The novel itself was dedicated to the dowager empress of Germany, Victoria, eldest child of Queen Victoria, because Haggard had heard that the late Frederick III, during his slow death from cancer, had enjoyed reading Haggard’s adventure novels.* I am not sure whether the Norse subject matter of the book played any role in this, although the dedication certainly might have been tied into the confused pan-Germanic sentiment common in England at the time.

    * Influenced by his wife, Frederick was a noted Anglophile and liberal, although when the treatment of his chronic laryngeal problems was turned over to the English physician Morell Mackenzie, Mackenzie’s incompetence may have cost the Kaiser his life. The growth on Frederick’s vocal chords was correctly identified as cancerous before Mackenzie was even called in, but thanks to a combination of his bad judgement and inept biopsy procedures, Mackenzie continued to deny the diagnosis for about a year. Although it is not clear whether Frederick could have survived with aggressive treatment in 1887, by the time the cancer diagnosis was finally accepted in spring 1888, the disease had definitely metastasized to the lungs, and long-term survival was impossible. Aside from her general preference for an English physician for her husband, Princess Vicky was apparently offended that some of the German doctors who saw her husband wanted to have a frank discussion with her of the additional complications posed by her husband’s tertiary syphilis. (Supposedly, Frederick caught syphilis while partying with the Egyptian khedive Isma’il Pasha in Suez in 1869, but that account has been questioned. In fact, there is a lot of confusion and outright misinformation about Frederick’s syphilis. For example, it has sometimes even been claimed that it was neurosyphilis, not cancer, that killed Frederick, which is absurd.)

  59. David Marjanović says:

    “bepraise”

    Funnily enough, German just has preisen there.

  60. J.W. Brewer says:

    I like the surviving Old Norse texts just fine, but Haggard rating them higher than the corpus of Old Irish literature perhaps reveals an unfortunate corollary to “confused pan-Germanic sentiment.”

  61. Stu Clayton says:

    There’s bepreisen, but that means “to [assign a] price [to]”.

  62. He wrote Eric Brighteyes as something halfway between the Icelandic saga style and a modern novel. It has less of the repetitions and digressions common in the originally oral epics, but it also maintains something of all archaic styling and diction—hence bepraise.

    But the Icelandic sagas are not archaic at all; if you know enough Norse to read them, they are absolutely straightforward, in normal words as spoken by the Plain Man of Iceland. (Except, of course, when they’re quoting poetry.) It’s just the irresistible pull of Historical Novelese.

  63. David Eddyshaw says:

    I like the surviving Old Norse texts just fine, but Haggard rating them higher than the corpus of Old Irish* literature perhaps reveals an unfortunate corollary to “confused pan-Germanic sentiment.”

    I must say that I agree with Haggard, despite having no pan-Germanic sentiment in the game at all, not even a confused one. As far as prose goes, anyhow; I’d give the prize to Irish for poetry.

    But then, it’s not a competition.

    *”Old” Irish technically excludes most of the stuff of literary interest, but I’d persist in the same view even if you include stuff only surviving in Middle Irish garb.

  64. Yes, I agree with Haggard as well. That’s not pan-Germanic sentiment, it’s just recognition of rare greatness.

  65. J.W. Brewer says:

    Perhaps I do not take the Norse stuff seriously enough because at one point in time I was able to kind of sort of read them (ON having seemed an easy-peasy language when I took a class in it a mere three and a half decades ago this coming fall) whereas my attempt the following semester to dip my toe into the grammar and lexicon of Old or perhaps Old-to-Middling Irish was a humiliating failure. So the Ulster Cycle I really know only in translation.

    (My copy of E.V. Gordon’s _An Introduction to Old Norse_ currently resides on the bookshelf that is at eye-level to my right when I am seated in front of this computer – pulling it down and opening it to a portion of Egils Saga I get the pleasing sense that it would not take that long to get myself back in the position of making sense of it, but for the distractions of children and employment etc from which I was free when I was one-and-twenty.)

  66. My copy of Gordon is not at eye-level, but it’s to my right and within easy reach!

  67. David Eddyshaw says:

    No dispute about the relative difficulty of Old Norse versus Old Irish, though. But then, Old Irish is hard to beat at that game. It gives you something of the same feeling as Navajo: that the tribal elders must have all sat down together to devise how best to make their language unlearnable for foreigners. (“Well, Patrick: we should start by dropping every other vowel*. And then make sure never to level out the resulting forms, except to make them more irregular. Also, we need more consonants. You can never have too many different consonants. We should add a few dozen that none of the neighbours can pronounce. That’ll teach them.”)

    *The Kartvelian language Svan does much the same thing. Presumably the tribal elders’ feeling was that being as complicated as Georgian morphologically was not really enough.

  68. J.W. Brewer says:

    Re copies of Gordon close to hand, cue Dr. Evil saying to Austin Powers “We’re not so different, you and I.” (Which, as between myself and hat, is the super-villain in the scene is left as an exercise for the reader.)

  69. David Eddyshaw says:

    It all comes down to cats.

  70. Old Irish literature is crazier though. And I once attended a seminar dedicated to Irish (mostly Old) words for crazy.

    If you need an example of how understanding texts produced within another cultrue can be difficult, it is perfect.

    Cf., in Stokes’s translation:
    White as the snow of one night were the two hands, soft and even, and red as foxglove were the two clear-beautiful cheeks. Dark as the back of a stag-beetle the two eyebrows. Like a shower of pearls were the teeth in her head. Blue as a hyacinth were the eyes. Red as rowan-berries the lips. Very high, smooth and soft-white the shoulders. Clear-white and lengthy the fingers. Long were the hands. White as the foam of a wave was the flank, slender, long, tender, smooth, soft as wool. Polished and warm, sleek and white were the two thighs. Round and small, hard and white the two knees. Short and white and rulestraight the two shins. Justly straight and beautiful the two heels. If a measure were put on the feet it would hardly have found them unequal, unless the flesh of the coverings should grow upon them.

    And this is in translation. (Togail Bruidne Dá Derga)

  71. @J.W. Brewer: You mean one of you is a sock [puppet]?

    Seriously, a villain saying, “We’re not so different, you and I,” is such a cliche that I am surprised you took Austin Powers as an example. The Austin Powers scene is a parody of the existing cliche, playing on the fact that both characters were played by Mike Myers. (Incidentally, I remember being impressed that, in spite of the first Austin Powers movie getting a huge amount of pre-release advertising and publicity, the fact that Mike Myers also played Dr. Evil was kept relatively under wraps until the film actually came out.)

  72. J.W. Brewer says:

    @Brett: I prefer to think of all of those earlier cliched usages in other films as simply the prelude to Austin Powers. A praeparatio evangelica, as the theologians would say.

  73. external dative possessors — I don’t think we have any of those around here either. There might have been some in ON, though, it sounds familiar.

    Indeed, and in OE too: þa sticode him mon þa eagan ut lit. ‘then someone gouged him the eyes out’ is a much-loved example from the OE translation of Orosius. But what you have instead is locative external possessors, which are a “Northern” feature.

    In ON the DEP could be expressed either by a plain dative, or by dative + á, cognate with English on. The languages separately created upp á and upon, but while the latter survived unchanged, the former became , giving us sentences like Norw. Leveren matte de fjerne på ham lit. ‘liver.DEF must they remove on him’ or ‘they must remove his liver’ in DEPless English. I don’t know if this works in Danish; Swedish definitely has LEPs, with DEPs only in dialect (possibly from Low Saxon?).

    LEPs then spread to Finnish and Estonia, where they were expressed with the adessive case, and from there to the Russian preposition u, which now competes with the original DEP construction. There was also a westward spread to Old Irish.

    DEPs cannot be PIE, because Anatolian and Celtic lack them completely; their absence in English is probably yet another British Sprachbund effect. They are rare to nonexistent in Skt (partly because of dat/gen merger in pronouns, which are the most common kind of possessors). and in early Gmc; there are none in Runic Norse, almost certainly none in Gothic (the syntax is artificial, so we don’t really know). Homeric Greek has them, but they disappear in or before Classical times, only to return in Modern Greek. They are not found in Turkic or any Caucasian family as far as is known (people don’t usually write about the constructions a language does not have).

    Per contra, Hungarian alone among Uralic languages has DEPs, probably picked up from Slavic (which has them), and so do Basque and Maltese. Lastly, modern Hebrew (but no earlier varieties) has them, which could be from almost any Central or Eastern European language or all of them together.

  74. David Eddyshaw says:

    DEPs cannot be PIE, because Anatolian and Celtic lack them completely

    While your conclusion seems very likely, this is surely a non sequitur? The same argument would show that PIE had no voiced aspirate stops, after all.

    If it’s a rareish feature cross linguistically (which seems likely)*, frequent yet independent spontaneous loss of such a peculiar feature wouldn’t be a far-fetched hypothesis.

    *Proper polysynthetic languages just say “they eye-outgouged him”, of course. Why just incorporate the pronoun à la française when you could incorporate the whole noun phrase?

  75. And I once attended a seminar dedicated to Irish (mostly Old) words for crazy.

    The one that came instantly to my mind was baile, buile.

  76. David Eddyshaw says:

    I agree that Irish does crazy better than Norse does.

    Welsh tends to the whimsical-fantastic rather than the crazy. I expect it’s something to do with the Roman conquest.

  77. Well, Patrick: we should start by dropping every other vowel

    Nay, not Patrick: he was a Romano-Briton, as his name Patricius loudly proclaims. Indeed, he brought two things to Ireland, Christianity and /p/.

    In any case the conspirators would have said “We will be after dropping every other vowel.” In Primitive Irish, of course.

    Old Irish literature is crazier though

    But Old Norse literature is funnier. Here’s Shippey again:

    Old Norse myth is strangely funny. I don’t mean “comic,” exactly, I mean amusing. Thórr is often a figure of fun, in a way which is not true of Zeus or Jupiter. Think of him disguised as Freyja when he tries to recover his hammer in Þrymskvida, with the giant asking:

    Hví eru öndótt augu Freyju?
    Þykki mér ór augum eldr of brenna

    and Loki craftily replying:

    Svaf vætr Freyja átta nóttum,
    svá vas hon óðfús í Jötunheima.

    Think of him struggling to drain the drinking-horn in the house of Útgarða-Loki, which is connected to the sea, or to pick up the cat, which is really the Miðgarðsormr. This is not the kind of story we are told about Hercules.

    But there are plenty of other examples. The Krakumál ends with Ragnar Loðbrók saying “Hlæjandi skal ek deyja”, and in another of the versions of his death in the ormgarðr his last words are “Gnyðja munu grísir ef galtar hag vissi”, “if they knew how the old boar died, the little pigs would grunt.” But gnyðja is surely a vulgar word, and “the little pigs” is a funny way to refer to Ívarr hinn beinlaussi and Sigurðr orm-i-auga. They do not say things like this in Virgil’s Aeneid.

    Nevertheless these vulgar or amusing ways of telling mythic or heroic story are not intended in any way to diminish the status of Norse gods or heroes, just the opposite. And Norse saga and edda is perfectly capable of reaching out to the sublime and the magnificent, as we see from the Völuspá or the Sólarljóð. You will find the funny, and the heroic, and the sublime, all very close together in the pages of [Gordon …].

    I would suggest, in fact, that this book shows very well a second reason for the attraction of Old Norse literature in the learned world, which is that as well as being funny, it rejects the classical notion of decorum: of keeping the styles separate, high style, middle style, low style. This is notoriously a native English trait as well – it is what made Shakespeare unacceptable to Voltaire – but Old Norse literature gave this English failing a distinguished ancestry. (Let me note, en passant, that in this Introduction Gordon gives a strangely composite account of the Battle of Stiklastaðir, which is highly “indecorous,” and reminds me in a way of the end of Gerpla.)

  78. David Eddyshaw says:

    Not convinced: there are plenty of comic stories about the Greek gods from the Iliad onwards, and the Dionysus of Aristophanes’ Frogs is the same terrifying god as the Dionysus of the Bacchae: he’s not afraid to let his hair down, act the fool and party. But don’t mess with him if you know what’s good for you. (It would be a great misreading of Aristophanes to think of him as at all irreligious.)

    Virgil is not the place to look for these things. Great poet though he was, he had a stick up his arse* when he wrote the Aeneid. Pious bloody Aeneas …

    *As we say in Latin. Satura tota nostra est.

  79. baile, buile.

    I am aware of 4 people doing Celtic studies in Moscow in 80s (and 90s, but in 90s many more people joined in). Maybe they were the only ones in whole USSR. They all focused on Ireland, and their work suffered a lot from inaccessibility of many foreign publications. Studing a country from behind the Iron Curtain without being able to come and work there is nothing easy.

    One is a philologist, with her main focus closer to literature. One was a philologist (historical linguistics and poetry). One was what is called a linguist here. One was a historian. When the Iron Curtain fell and everyone found himself/herself in love either with Ireland specifically or with Celtic things in general (I do not know why) people who used to have one student a year (collectively) found themselves surrounded by a little crowd. Mikhaylova, the only one who worked in the university rather than in a research institution, said: before she was trying to lure people in the field, now it will be like a Chinese Buddhist monastery and she is going to be beat students with bamboo sticks (one of those students later became a rock-star, January First-of-May mentioned her recently.)

    She built her course around a seminar, in turn, dedicated to Buile Shuibhne. I do not know how long this seminar continued, I attended it in 90s, but it could easily have began in 80s (with one student?) and maybe is still thriving. The text was basically an excuse to study anything. Madness too.


    (I later realized that with all my love to Ireland and Brittany – I, unfortunately, haven’t dealt with Welsh – I do not want a career in Celtic linguistics. And by now hardly remember any Old Irish or Breton)

  80. PlasticPaddy says:

    @jc
    re “I will be after…”, this is not cromulent in Irish or Hiberno-English. It is difficult to explain, but it may be because the principal sense of the construction is to remind someone of something they witnessed, or to provide context for or explanation of an action or occurrence. To give an analogy, “I will have just come back from shopping” is only cromulent if one is describing one’s premonition, perhaps under the influence of one’s preferred hallucinogenic.

  81. Trond Engen says:

    John C.: sentences like Norw. Leveren matte de fjerne på ham lit. ‘liver.DEF must they remove on him’

    Almost. The fronted liver is clearly for emphasis, and måtte (with å) is a preterite. “His liver had to be removed”. De måtte fjerne leveren på ham “They had to remove his liver”.

  82. Not convinced

    Me neither. Talk about special pleading! And Irish literature is full of humor from the beginning, unless you confine your attention grimly to glosses.

  83. David Marjanović says:

    Not So Different

    almost certainly none in Gothic (the syntax is artificial, so we don’t really know)

    I once read a paper that pointed out the syntax follows the Greek except when it doesn’t, and argued the translation was literal wherever the result was within the range of Gothic possibilities, but not otherwise, so we can actually learn something about Gothic syntax from the Bible. And that turns out to be coherent and realistic-looking.

    Welsh tends to the whimsical-fantastic rather than the crazy. I expect it’s something to do with the Roman conquest.

    Caesar, that most reliable of narrators, implied that rather heavily.

    Indeed, he brought two things to Ireland, Christianity and /p/.

    Not /p/, that came later – he ended up as Cothraige and similar forms at first, not to mention the Ogham inscription with qatrikias.

  84. DEPs cannot be PIE, because Anatolian and Celtic lack them completely; their absence in English is probably yet another British Sprachbund effect. They are rare to nonexistent in Skt (partly because of dat/gen merger in pronouns, which are the most common kind of possessors). and in early Gmc; there are none in Runic Norse, almost certainly none in Gothic (the syntax is artificial, so we don’t really know). Homeric Greek has them, but they disappear in or before Classical times, only to return in Modern Greek. They are not found in Turkic or any Caucasian
    That history doesn’t IMHO exclude convincingly that they had been a feature of PIE – it can be argued that both in Celtic and Anatolian they fell out of use due to substrate influence on syntax; in Gothic they may not be attested due to the texts being translations from a language that doesn’t have them (the only conclusive evidence would be if the Greek text had them and the Gothic translation avoided them); the Runic Norse corpus is so small that their absence isn’t evidence of their non-existence. The rest of the evidence rather speaks for them being an inherited feature that was lost in later stages of individual IE languages. (The absence in Anatolian can also mean that DEPs developed only in core PIE after Anatolian split off.)

  85. J.W. Brewer says:

    I take it there’s no consensus of the learned as to the relative importance of funniness and craziness in ranking the objective quality or world-historical importance of a given language’s early literary canon?

  86. Well, Schüttpelz in his Geschichte des Welthumors in der Perspektive der Wahnsinnsliteratur uses a complex system of ranking that no one else has been able to decipher, let alone use, but it might provide a provisional approach to the question. Further research is needed.

  87. Unity and diversity in grammaticalization scenarios
    Walter Bisang Andrej Malchukov
    https://langsci-press.org/catalog/book/152

    (Open access)

  88. Dative constructions in Romance and beyond
    Anna Pineda Jaume Mateu

    https://langsci-press.org/catalog/book/258

  89. and from there to the Russian preposition u, which now competes with the original DEP construction.

    I would not call it competition:/ “By” for u:

    “I borrowed by him 3 roubles”
    “I gave to-him 3 roubles”

  90. Lars Mathiesen says:

    I’m late to the party, but Danish represent: LEPs work fine in most cases where DEPs are used in Spanish, for instance, but they are not the only option: De tog hans lever ud is even slightly higher register than De tog leveren ud på ham. If the possessor is an inanimate noun, however, the LEP version is preferred: Han tog nummeret på bilen vs Han tog bilens nummer — the latter is a bit stilted.

    The exception is reflexives (subject = possessor). While you can in fact say Jeg børster tænder på mig selv, it is very marked relative to Jeg børster mine tænder and you can’t omit selv — pragmatically it can only be used when denying the possibility that someone else is doing the brushing. (In Spanish, of course, Me cepillo los dientes is the expected form).

    There is a difference between Swedish and Danish here (I don’t know which side Norwegian picks): Jag tar på mig klärna vs Jeg tager tøjet på. One language insists on including the pronoun, the other insists on leaving it out, but the locative has sort of been baked into a phrasal verb. (The placement of the adverbial parts of phrasal verbs is another difference, in case you wondered).

  91. Trond Engen says:

    Norw. Jeg kler på meg.

    Kle på is very much felt as a phrasal verb, but the preposition seems to be applicative in origin, since without it the verb is differently jected: Den kler deg “It suits you”.

  92. Lars Mathiesen says:

    Interestingly, Danish does use a pronoun with that verb: Jeg klæder mig på. I have no idea why this is.

  93. There is this recent volume Reconstructing Syntax (open access) from Brill, and the Chapter 4 by Silvia Luraghi is External Possessor Constructions in Indo-European.

    I am reading it in the moment.

  94. Reflexive constructions are an exception in Russian as well – while it normally used DEPs, in reflexive constructions concerning body parts, every indication of ownership is omitted – ja moju golovu “I’m washing my hair”, lit. “I wash head”. If the topic is someone else’s body part, the DEP is used ja moju golovu Petru “I wash Peter’s hair”.

  95. David Marjanović says:

    every indication of ownership is omitted

    An interesting parallel to the noun incorporation in, uh, whatever parts of nonstandard German, including my dialect: staubsaugen, haarewaschen, zähneputzen…

  96. A question to folks speaking languages with DEPs.
    Modern Hebrew uses DEPs extensively (probably under the influence of European languages). One type of construction omits the possessed item, leaving it to be inferred or determinate. For example:

    nigmar li
    finish.PASS.3.M.SG.PAST 1.SG.DAT
    ‘I’m out’ (of something understood from the context).

    ro’im lexa
    see.3.M.PL.PRES 2.M.SG.DAT
    ‘They (indefinite) see yours’ i.e. one can see some part of your body or clothing that should be hidden.

    axlu li, šatu li
    eat.3PL.PAST 1.SG.DAT, drink.3PL.PAST 1.SG.DAT
    ‘They (indef) ate up mine, they (indef) drank up mine’; an expression used to mock North African/Middle Eastern Jewish immigrants and their descendants complaining about their dispossession and maltreatment in Israel.

    Is that a common or universal construction in DEP-loving languages?

  97. Since I do not see it mentioned it above— Malchukov has a second piece about case on the same site: serious-science.org/grammatical-case-morphology-syntax-and-word-order-9354 (They appear to have been posted on the same day.)

  98. @Y: I can’t think of any parallel to the constructions you quote in German. A reason may be that in German, the dative case is formed without a preposition, only with case endings, so such a construction would just consist of the verb with not even a trace of a complement.

  99. David Marjanović says:

    Is that a common or universal construction in DEP-loving languages?

    Oh, literal dative possessors? I’ve never encountered that before.

    Standard Average European “dative external possessors” aren’t really underlyingly possessors. They’re emphatic datives. As it happens, they mention the possessors, so the possessive pronoun that is obligatory in English can be downgraded to an article (and usually is).

  100. Hans: that’s not so different from the Hebrew. Can you have a German construction consisting of a verb, the “possessor” as a noun or a pronoun in the dative, but omitting the possessed/affected part?

    DM: not literal dative possessor. It still has the semantics of control, not necessarily exactly the same as possession using genitives and such, similar to European languages.

  101. Standard Average European “dative external possessors” aren’t really underlyingly possessors. They’re emphatic datives.

    Yes. I think the term comes from possessor raising/ascension/promotion in transformational grammar.
    See.

  102. As it happens, they mention the possessors, so the possessive pronoun that is obligatory in English can be downgraded to an article (and usually is).

    And yes, I guess it is inpired by English. In Russian:

    Wash hands!

    “Your” is not “omitted”. In just not needed here:/ Pragmatically (how else?). It is essentially an article, and why articles in Russian? I can’t know what they are really underlyingly and can’t speak about SAE as a whole, but I can not see how they are possessors in Russian in any concievable way.

    Sometimes the dative actant is indeed a possessor of the direct object. But sometimes your agent is also the possessor of her (instrumental) tool. I am quite irritated that many people write that EP “expresses possession”.

  103. David Eddyshaw says:

    Although Kusaal doesn’t do the “I wash me the hands” thing, nevertheless there are aspects of “possessive” pronouns which do have properties reminiscent of datives.

    The scare quotes around “possessive” are because Kusaal has no distinctive possessive pronoun forms: the exact same forms appear as are used as verb subjects, so on the face of it there seems to be no basis for analysing (say) m biig “my child” and dau la biig “the man’s child” differently: the “possessive” meaning simply arises from the concatenation of the two NPs.

    However:

    The cases are actually subtly different: unlike definite NP possessors, personal pronoun possessors don’t make the whole NP definite.

    In dau la biig the “child” has to be interpreted as definite: although it is grammatical to repeat the la after the second noun, la then has to be interpreted as deictic “that child of the man.” (This, the original sense of la, survives in a couple of other syntactic contexts as well.)
    On the other hand, if you want to say “a child of the man” you have to spell it out specifically: dau la bisɔ’ “a certain/some/another child of the man’s.”

    But if the possessor is a non-contrastive personal pronoun, the article is used normally, in just the same way as with an unpossessed noun, so m biig is more accurately “a child of mine”, and “my child” has to be m biig la. You even have contrasts like

    M biig kae.
    1sg child:SG not.exist
    “I have no child.”

    M biig la kae.
    1sg child:SG ART not.exist
    “My child’s not there.”

    Similarly, a very common introductory strategy in stories involves constructions like

    Pu’a da bɛ n mɔri o biig …
    Woman:SG TENSE exist CAT have 3sg child:SG
    “There was a woman who had a child …”
    (literally, “A woman existed, having her child …”, where subsequent mentions of the child would go o biig la.)

    This all makes sense if you think of the “possessive pronoun” in m biig or o biig as a sort of “embedded dative”: as it were “child which-is-to-me/her.”

    [There is also a tonal distinction between pronoun “possessors” and other NPs as possessors, in that the possessed noun is subject to a tonal neutralisation, but only after NPs other than personal pronouns. I’m not sure how this fits in to the picture exactly.]

  104. A Greek example:

    Μεθαύριο θα πάω στο νοσοκομείο να μου κόψουν τα ράμματα.
    =
    The day after tomorrow, I’m going to the hospital in order to get the stitches removed.

    μου is technically the genitive of M εγώ, but here it’ s behaving as an indirect object.

    More on ethical datives:

    https://core.ac.uk/download/pdf/83938741.pdf

    https://www.researchgate.net/publication/271827075_Ethical_Datives_A_Puzzle_for_Syntax_Semantics_Pragmatics_and_Their_Interfaces

  105. Two modes of dative and genitive case assignment: Evidence from two stages of Greek

    https://link.springer.com/article/10.1007/s11049-020-09465-z

    (Open access)

  106. Lars Mathiesen says:

    FWIW, Danish optionally has the same thing as Russian for reflexives: Vask ‘hænder!, Vask hænderne! (definite suffix) and ˈVask dine hænder! are equally cromulent, and I don’t even sense any pragmatic difference. (Actually the stress is constrained, as marked, so if you want contrastive stress your choice is restricted. Also the third one is better for annoyed parent voice because you can have two stresses). This choice runs through the verb paradigm, except maybe for the “centaur” verbal noun — vasken hænder is the only one that feels right. (Does German have something like a noun Händerwaschen?)

  107. Händewaschen
    As for the reflexive dative object, it can be dropped as well, but it’s not like Russian where actually using it would imply a strong emphasis.

  108. Martin Haspelmath, External possession in a European areal perspective, in : Barshi, Immanuel & Payne, Doris (eds.) External possession. Amsterdam: Benjamins, 109-35.

    I think it is the source of “areal”. Haspelmath does not say it is not between “areal” – “early”- “areal with inherited source”. He is synchronic. Naturally, a shared retention is also an object of study for areal linguistics. But: horizontal effects are extremely interesting and hard to model, in diachrony especially, retentions especially and semantics especially.

    Here the possessive NPs expressing ‘child’s’, ‘his’ and ‘my’ are not constituents of an NP, but contrary to the English translation, they occur NP-externally.

    And this is what irritates me.

  109. PlasticPaddy says:

    @dm
    My primary association for phrases like Händewaschen is with when they are given in a Kommandostimme or Befehlston and followed by an exclamation mark. 😊

  110. David Marjanović says:

    emphatic datives

    Empathic, I mean. *facepalm*

    Can you have a German construction consisting of a verb, the “possessor” as a noun or a pronoun in the dative, but omitting the possessed/affected part?

    I can’t think of any where the possessed part can be omitted (to leave it vague for example?). Unlike in English (gimme!), you cannot omit “it” in “give it to me”, unless you omit “me” too – gib her!, literally “give towards here”.

    The only possible exception is that I vacillate, in my dialect, between dative and accusative in cleaning my nose: mir ~ mich schnäuzen. It can be argued that there is an omitted accusative object there, my nose, but if so, that’s historical and not synchronic. (In the standard, mich schnäuzen and mir die Nase schnäuzen both occur, but the possibly mixed form does not. – The third-person reflexive pronoun, sich, somehow lost the dative form sir a thousand years ago, and the accusative form sich is now used in both cases.)

    μου is technically the genitive of M εγώ

    It’s the identical genitive and dative, and occurs twice in sentences like “my husband told me”.

  111. Vanya wrote above on a nearby wall „lutsch mir“ ( suck me DAT). I instinctively supplied a missing „den Schwanz „

    In Russian „lutsch mir“ works just fine.

  112. Google gives for “empathic” encouraging 61 result. And 260 for “emphatic”, including a paper by E. Guest from 1844.

    (“Datives of empathy” 9, but “dative of empathy” 193000. Crazy)

  113. @drasvi: All that’s showing is the meaninglessness of Google counts. I currently get 9.13 and 14.00 Mghits for “empathic” and “emphatic” (both with quotes; the number without quotes, which includes other inflected forms of the same lexemes, are 3–6 times higher).

  114. languagehat: But the Icelandic sagas are not archaic at all

    One specific element of Eric Brighteyes that I think of as an archaism emulating some of the Norse sagas—albeit a stylistic archaism, rather than a linguistic one—is switching the narration between past and present tense in the midst of a single scene. It is easy to overlook the fact that the modern standard of sticking to past or present tense narration (usually past) is a convention, not a literary necessity. (And more modern authors have certainly shown that interesting things can be done by violating this convention.) Similarly—although less prominently—Eric Brighteyes occasionally departs from the convention of telling the story in the third person, ; and the same kinds of departures can again be found in Icelandic, Greek, and Irish epics as well.

  115. As I am collecting names for this DEP:

    Havers’ term dativus sympatheticus is often translated as ‘dative of affection’, or ‘dative of interest’.
    (from Luraghi, 2020)

    Havers has:

    Die bisherigen Benennungen dieses Dativs sind meist recht unglücklich gewählt: Günther bezeichnet ihn in seiner Abhandlung über den Dativ bei Homer S. 58 als Dativus possessivus mit der den Tatsachen vollkommen widersprechenden Begründung “quod germanice fere semper pronomine aut genetivo possessivo eum reddimus”. Im Lateinischen ist die Bezeichnung Dativus dynamicus1) oder Dativus energicus üblich; die letztere geht, so viel ich sehe, auf Holtze, Syntaxis adumbratio S. 18 zurück, von dem dieser Kasus früher in der Syntaxis prisc. script. Lat. § 61 als Dativus personae, cui Studium est bezeichnet worden war. Am besten wäre die von H. Schwartz Zsch. f. dtsch. Phil. Bd. 17, 77 vorgeschlagene Benennung Dativus ethicus, ‘hätte man sich nicht’, wie Schwartz selbst sagt, ‘aus der Syntax der alten Sprachen her gewöhnt, nur diejenigen Dative so zu nennen, die im Satze eigentlich überflüssig sind’. Ich nenne den mit dem Genitiv im Austausch stehenden Dativ Dativus sympatheticus (Dat. symp.) im Anschluß an die Bemerkung von Gildersleeve zu Pindar Pyth. IIΙ 46 ἀνθρώποισιν ἰᾶσθαι νόσους : “ἀνθρώποισιν more sympathetic than ἀνθρώπων”.

    1) In einem anderen Sinne gebraucht Krüger Att. Synt. § 48, 15 diese Bezeichnung; der dynamische Dativ ist nach ihm identisch mit dem instrumentalen Dativ.

    Havers.
    Havers via sci-hub.

    and:
    S. 19 ff. Der von E. H. Miles in seinem Aufsatze The Dative of the Possessor (Classical Review Vol. XI, 1897, S. 142 f.) versuchte Nachweis, daß idg. *moi, *toi, *soi had in early times not only a dative use, but also a possessive use (which was not derived from this dative use) überzeugtin keiner Weise.

  116. […] applicative constructions in fact divide into two different types semantically. In one type the applicative head denotes a thematic relation between an individual and the event described by the verb. This type will be called a high applicative ((1) in Table 1), since the applicative head attaches above the VP. The other type of applicative is low; the head combines with the direct object and denotes a transfer of possession relation between the direct object and the applied argument. From this proposal various applicative asymmetries fall out naturally, including new data on the combinatorics of secondary predication with the two different types of applicatives (§2.1.3). Further, it will be argued that low applicatives come in two varieties: one describes a recipient-relation between the indirect and direct objects and the other a source relation. It will be argued that so-called adversity constructions, which otherwise constitute a puzzling syntax-semantics mismatch, are in fact ordinary double object constructions except that they exemplify the source variety of low applicatives.

    http://aix1.uottawa.ca/~asalanov/UBA/pylkkanen-thesis.pdf

  117. There was a joke about

    rúchka krýshki liúka báshni tánka, “handle of-lid of-hatch of-turret of-tank”. It works well with “DEP”:

    Did we break to-hatch-of-turret-of-tank the handle-of-lid?

  118. John Cowan says:

    It’s pretty clear in English that the possessory sense of the dative died out long before the “ethical” sense (this is the word usually used in English grammars of Latin and Greek). Gouged him the eyes out doesn’t survive OE, but Shakespeare can still have Casca (who is playing the Plain Roman, since he is speaking prose) say he [Caesar] plucked me ope his doublet. The conflict of person eliminates anything possessive, and the context (speaking in front of a crowd) eliminates a plain benefactive reading ‘he plucked open his doublet for me’. Instead, it has to indicate Casca’s strong feelings about Caesar: the text continues “[…] and offered them [the crowd] his throat to cut. An [if] I had been a man of any occupation, if I would not have taken him at a word, I would I might go to hell among the rogues.”

    It’s noticeable that English He put his hands in his pockets is redundant from the perspective of more-SAE languages like German and French, where the determiners are normally articles. This was explained to me in my first German class with two rhetorical questions: “Whose hands would he put in his pockets? Whose pockets would he put his hands in?” (note the obligatory alternation in the position of in). And yet his hands/his pockets are what we say, and the hands/the pockets sounds totally L2.

    There is a similar criticism of the overuse of respective(ly) in English: “They both went to their respective seats” may be true, but hardly worth saying: if they exchanged seats it would be mentionable, but if not, “They both went to their seats” is much more natural, if it is necessary at all.

  119. By the way.

    Of course you can find DEP in Berber and Arabic too. Haspelmath’s areal can easily be extended. I am just not that experienced with Arabic and Berber dialects to say “how many”. Maybe Lameen can?

    It does not help us much with history though.

  120. Just quick (very quick) googling: Christian J. Rapold, Beneficiary and other roles of the dative in Tashelhiyt (2010) in Benefactives and Malefactives: Typological perspectives and case studies, Fernando Zúñiga, Seppo Kittilä (eds)

    p. 365 ff. (3.14 Possessor). The book can be found in Library Genesis.

  121. The guy seem to have the same doubts as I have:

    Such constructions alternate with cases where the possessor is encoded NP-internally by a possessive pronoun sufx or by means of a genitive phrase (36’ below). However, it is questionable whether in the so-called external possessor construction possession is really the main semantic role of the dative phrase. Impressionistically, its role is rather that of a general affectee (benefciary/malefciary), while ‘possessor’ is at best an implicature that can be cancelled. Affectedness, by contrast, cannot be cancelled in cases like (36). Whether this hypothesis is indeed correct for Tashelhiyt is a subject for further research. One argument against viewing “pure” possession as the main semantic role of the dative in such cases is the fact that the dative may co-occur with a possessive afx in the same construction (Šarić 2002:16; for Tashelhiyt data see below (41, 42).

    For the sake of cross-linguistic comparison, however, such cases are listed as “external possessor constructions” here, but this should be taken as no more than a (perhaps misleading) label for constructions in which the dative phrase can be interpreted as a possessor.

    But he is more polite;)

  122. David Eddyshaw says:

    The two “dative” roles are very much distinct. Kusaal doesn’t do external possessive constructions, even though it can add an “ethical dative” after almost any verb, including “exist”:

    Alaafʋ bɛɛ ba.
    Health exist 3pl
    “They are well.”

    Kusaal also has a lot of expressions where the “logical” object is expressed as an indirect object before a fixed-formula noun as the direct object, like

    O kadi ba sariya.
    3sg chase 3pl law
    “He’s judged them.”

    Ba niŋ o yadda.
    3pl do 3sg trust
    “They’ve believed her.”

    M zɔt o dabiem.
    1sg run:IPFV 3sg fear
    “I’m afraid of her.”

    O nwɛ’ɛ ba nu’ug.
    3sg strike 3pl hand:SG
    “He’s pleaded with them.”

    The pronouns are definitely not possessors in these constructions (although in some cases the difference from “possessive” pronouns is purely tonal.)

  123. But I do not have a problem with languages that do not have DEP. I have a problems with claims that some langauges do have this! Especially when I am told that in Russian she babisat to-me children “expresses possession”. Do we have a definition of this possessive function?

    If it is when we have A B C and D, and the linguist thinks that A an agent, D is a “verb” and C is an actual possession of B, then it sure tells us something about the linguist.

  124. The phrase “external possessor” origianated in early 80s under influence of “possessor raising” which comes from “possessor acsension” which tends to co-occurs with verbs like “to Chomsky-adjoin”, that is, very model-specific. When one needs examples freom old IE languages, one usually (it seems) cites Havers 1911. He defines it based on situations like this:

    ἐπὶ δ᾽ οὔατ᾽ ἀλεῖψαι ἑταίρων
    ….
    ἑξείης δ᾿ ἑτάροισιν ἐπ᾿ οὔατα πᾶσιν ἄλειψα.

    “on ears smear of-comrades” – [wax, Odyssey, 12 (μ), 47)
    “in-turn then to-comrades on ears to-all smeared” (12 (μ) 177)

    That is, the logic is that: if and only if a possessive from can be used in the same situation, it is dative of sympathy. Does it create a coherent class?

    P.S. “Wir begegnen in verschiedenen idg. Sprachen einem Dativ, für den auch der Genitiv eintreten kann; so heißt es z. B. bei Homer μ 177 ἑτάροιϲιν ἐπ᾽ οὔατα πᾶσιν ἄλειψα scil. τὸν κηρόν, ib. 47 steht aber in derselben Wendung der Genitiv: ἐπὶ δ᾽ οὔατ᾽ ἀλεῖψαι ἑταίρων | κηρὸν δεψήσας. “

  125. P.P.S.: “….Ebenso können wir im Nhd. wechseln zwischen Ausdrücken wie “die Kugel durchbohrte dem Feinde das Herz” und “die Kugel durchbohrte das Herz des Feindes”. Dem Wechsel von Dativ und Genitiv beim Nomen kann beim Pronomen ein solcher von Dativ und Pronomen possessivum entsprechen, z.B. nhd. “er hat mir die Hand verwundet” oder “er hat meine Hand verwundet”. Da Pron. poss. und Gen. in diesen Fügungen grammatisch auf gleicher Stufe stehen, gebrauche ich im folgenden für beide die gemeinsame Bezeichnung possessive Ausdrucksweise (Gegensatz: dativische Ausdrucksweise). Nicht für jeden beliebigen Dativ kann die possessive Ausdrucksweise eintreten; der Dativus commodi in einem nhd. Beispiele wie “der Yater baute seinem Sohne ein Haus” kann z. B. nicht durch den Genitiv ersetzt werden. Es handelt sich also bei dem mit dem Genitiv in Austausch stehenden Dativ um eine besondere Art dieses Kasus. ”

    —-
    “….smeared”
    I-smeared

  126. PlasticPaddy says:

    @drasvi
    I think the affective dative is perceived by the individual as possessive to the extent that the individual is possessive in his/her affections 😊. As an example “Sie nahmen uns Österreich” said by a L1 Czech speaker with regard to the Anschluss could be identical in meaning to “Sie nahmen unser Österreich” or “Sie nahmen Österreich [weg] von uns”, depending on the speaker’s individual psychological makeup.

  127. David Eddyshaw says:

    That is, the logic is that: if and only if a possessive from can be used in the same situation, it is dative of sympathy. Does it create a coherent class?

    But languages can have unequivocal “datives of sympathy” without normally using the construction to replace a possessive (Kusaal being a clear example.) So I would say that being able to use “datives of sympathy” to replace possessives is indeed an identifiable distinct coherent language trait* (and indeed we don’t seem to be having any difficulty in identifying which languages have it: French does, English doesn’t, etc etc.)

    We shouldn’t get distracted by Chomskyite fantasies about underlying syntax. The feature in question is apparent (or not) on the surface.

    *That’s not to say that it’s necessarily a particularly profound language trait, in the sense of being correlated with all kinds of other syntactic features. On the contrary, it seems that quite closely related and otherwise syntactically very similar languages may differ with regard to this feature.

  128. David Marjanović says:

    As an example “Sie nahmen uns Österreich” […] could be identical in meaning to “Sie nahmen unser Österreich” or “Sie nahmen Österreich [weg] von uns”, depending on the speaker’s individual psychological makeup.

    The first would rather be doubled up, though: sie nahmen uns unser Österreich.

  129. David Eddyshaw says:

    With regard to drasvi’s point about form versus semantics:

    Languages with noun incorporation often incorporate body parts particularly, or indeed exclusively. Bininj Gun-wok has a closed (though extensive) set of non-body-part nouns which can be incorporated, but can incorporate any body part noun, e.g.

    A-bikbik-bakme-ng.
    1-rib-crack-PP
    “I cracked my ribs.”

    Gun-dulk an-bid-djudme-n.
    IV-splinter 3/1-finger-stick.in-NP
    “A splinter is sticking into my finger.”

    In his grammar, Nicholas Evans has an interesting discussion (pp456ff) on how this relates to “possessor raising”, which is often assumed to be what’s going on in polysynthetic languages in cases like these. He comes to the interesting conclusion that in Bininj Gun-wok this is not appropriate as a description of a process of syntactic derivation but simply as a label about semantic interpretation, because the same argument structure is present whether or not the noun is incorporated.

    Thus, he contrasts this with the case of “benefactive” constructions (one sort of applicative), which Bininj Gun-wok can use to do the “wash me the hands” kind of thing:

    Bi-yau-melme-ng ngarduk na-beiwurd.
    3/3hP-child-touch.with.foot-PP my I-child.
    “He kicked my child.”

    Ngan-marne-yau-melme-ng na-beiwurd.
    3/1-BEN-child-touch.with.foot-PP I-child.
    “He kicked my child/He kicked the child on me.”

    But

    Ngan-melme-ng an-gorn.
    3/1-touch.with.foot-PP III-crotch.
    “He felt my crotch with his foot.”

    Ngan-gorn-melme-ng. (NOT Ngan-marne-gorn-melme-ng.)*
    3/1-crotch-touch.with.foot-PP
    “He felt my crotch with his foot.”

    He points out that even the fairly closely related language Rembarrnga does have possessor-raising incorporation, however, e.g. nga-njarra-bolhminj “I father-arrived”, i.e. “My father arrived.”

    *Interestingly, Evans says that Nganmarnegornmelmeng is not ungrammatical, however: “it could mean ‘he touched (her) crotch with his foot’ with the implication that ‘her’ refers to a close female relation (e.g. wife, sister) whose interests coincide with the speaker’s.” **

    **Hey, I didn’t pick these examples.

  130. being able to use “datives of sympathy” to replace possessives is indeed an identifiable distinct coherent language trait”

    I agree. But it is Havers 1911 and it is vague. 100 years have passed, and a typologist Haspelmath says “express possession” which is not just more vague. It is false:) Of course a part of the problem is inherent: every feature only makes sense in context [of a language or utterance], and langauge is the context for itself. You won’t find Russian dative in Klingon and likely nothing that even feels like Russian dative (we are back to universals here).

    Obviously, typology of interrelations between features is more likely to be productive.

    An attempt to anchor typology of semantics in “real world” is understandable. Haspelmath is too optymistic though:

    “Since conceptual similarities are by definition universal (if conceptualization is defined as the language-independent level of meaning), it should be possible to arrive at a universal representation of the mutual relations of meaning/conceptualization by comparing polsysemy patterns in different languages (thus, an alternative term for these universal semantic maps is “cognitive maps”).”

    ههه

  131. David Eddyshaw says:

    I share (what I take to be) your doubts about whether these “conceptual universals” actually exist at all, beyond a trivial level. “Language-independent levels of meaning”, eh? I think they have those in El Dorado …

    “By definition” is a bit of a giveaway, I think. We’re assuming what really needs to be proved. Defining something doesn’t conjure it into existence.

    However, while I agree that Haspelmath is too optimistic about ever arriving, I think that the journey could be interesting (and educational) anyway. Partial successes may be of immense value even when complete success is wholly impossible.

  132. By the way.

    Take me by the hand, is it accusative external possessor?

  133. David Eddyshaw says:

    Now it’s interesting you should ask that, because that is (mutatis mutandis) pretty much exactly how Nicholas Evans ends up analysing body-part incorporation in Bininj Gun-wok: viz that the “me” and the “hand” are effectively appositional.

    Part of his evidence for this is that the degree of overlap between the two is variable in such a way that it does not seem possible to say which element is the head, and interpretation requires real-world knowledge, not just reading off the syntax. He cites

    Abanmani-bid-garrme-ng daluk.
    1/3du-hand-grasp-PP woman.
    “I grabbed the two women by their hands” NOT *”I grabbed the woman by her two hands.”

    In a somewhat similar way, I think it would be possible* to analyse the French Je me lave les mains as “I wash (myself) with respect to the hands”, with mains as a sort of “accusative of respect”, like saucius pectus “breast-wounded” in Latin.

    *Of course, just because you can do something, doesn’t mean that you should do it …

  134. By the way-2. And back to “suck me”. I hope it is clear to everyone, that:

    “She made to-me blowjob” has a precise Engish semantical counterpart:
    “She gave me a blowjob”

    The very same thing that we express with dative (that makes us use dative here, that we mean by dative), IS actually expressed with dative in English as well.

  135. David Eddyshaw says:

    Evidently I spoke too soon about universals …

  136. David Eddyshaw says:

    My French example doesn’t really work, unfortunately. Je lui lave les cheveux, not *le lave. Oh, well. No PhD for me … the proposal didn’t have enough about theta roles in it anyway.

  137. My French exmaple is more French.

  138. David Eddyshaw says:

    Peut-être.

  139. David Eddyshaw says:

    Looking at Dimmendaal’s Turkana grammar, which is one of those good enough to address such questions thoroughly, I see that the construction with body-part objects differs according to the definiteness of the object, with the possessor being promoted to object if it’s definite (p356):

    ŋwòonɪ̥̀, ɛ̀àbùnì emunɪ̥̀ nàwuyè̥ kaŋ̀, tokɔɳ̀ ɪkɔ̀kʊ nàkɛ̀jʊ̀
    other.day, 3:PA:come:A snake in.home my, 3:bite child in.leg
    “the other day a snake entered my homestead, and it bit the child [object] in its leg”

    ŋwòonɪ̥̀, ɛ̀àbùnì emunɪ̥̀ nàwuyè̥ kaŋ̀, tokɔɳ̀ nàkɛ̀jʊ̀ à ɪ̀kɔkʊ̀
    other.day, 3:PA:come:A snake in.home my, 3:bite in.leg of child
    “the other day a snake entered my homestead, and it bit some child [possessor] in its leg”

    (the differing tones on ɪkɔkʊ “child” mark respectively the absolute and genitive cases; like the rest of Eastern Nilotic, Turkana is a marked-nominative language, and marks case entirely by tone changes*.)

    *How cool is that?

  140. David Eddyshaw says:

    Hausa has applicative verb forms, but doesn’t use them for “I wash my hands” type constructions, which instead work just as they do in English or Kusaal.

    (The situation is a bit complicated with applicatives in Hausa: there isn’t a unified “applicative” derivation, but many verbs must alter “grade” to be used with an indirect object; the grades are rather like Semitic binyanim.)

  141. Nom. ɪkɔ̀kʊ vs. gen. ɪ̀kɔkʊ̀… what is going on there? Are all the tones ‘flipped’ between neutral and low?

  142. David Eddyshaw says:

    The genitive has a fixed tone overlay (varying according to number and to the segmental shape of the word, but predictable), whereas the absolute shows the word’s basic lexical tones. The nominative in principle has an initial imposed low tone, but this triggers fairly complicated internal tone sandhi changes. Grammars of other Eastern Nilotic languages mostly seem to give up trying to give rules and just say that the changes in the nominative are lexically given, but Dimmendaal is teh hardcorez (either that, or the Turkana system is just more transparent than most.)

    Eastern Nilotic languages also just lurve having lots of unpredictably different plural (and singulative) forms for nouns; it’s as bad as Welsh or Hausa. And they have grammatical gender, because they would, wouldn’t they?

  143. Thanks… add this to the list of accidental inversions in the world’s languages (like German wer/wo and English who/where).

  144. differs according to the definiteness of the object

    I am not surprised. In Russian you have dative experiencer, in other langauges it can be topicalization or definiteness or any other important cathegory.

  145. David Eddyshaw says:

    It supports your contention that, for all that there are evidently common principles at work, there is no single magic bullet that will neatly account for all the different ways these principles are instantiated in real languages.

    In fact, it may be that there cannot be such a magic bullet: it’s not just that we haven’t been clever enough to discover it yet. (On this point, I am myself agnostic.)

  146. der Yater

    ORLV?

    “By definition” is a bit of a giveaway, I think. We’re assuming what really needs to be proved. Defining something doesn’t conjure it into existence.

    No, Haspelmath is perfectly right in his own terms (literally). A lower-case comparative concept is what Humpty Haspelmath says it is, because he creates it for the sake of making cross-linguistic comparisons. If the comparisons that can be made aren’t useful, he drops the term or changes its definition. His (title case) language-specific categories, on the other hand, are inferred from the evidence and cannot be used for comparison.

    As he says in his early (2010) paper, suppose we want to test the claim “In all languages with a dative and an accusative case, the dative case marker is at least as long as the accusative case marker.” In order to do so, we need to have inter alia a definition for the comparative concept “dative”.

    Now on a Chomskyite view, “dative” is a (universal) category and therefore a comparative concept, but Haspelmath cannot swallow this: “It is relatively easy to see that dative cases (or dative-like cases) cannot be equated across languages, i.e. there is no [such thing as a] cross-linguistic dative category. The Russian Dative [titlecased because language-specific], the Korean Dative and the Turkish Dative are similar enough to be called by the same name, but there are numerous differences between them and they cannot be simply equated with each other. Clearly, their nature is not captured satisfactorily by saying that they are instantiations of a cross-linguistic category ‘dative’.”

    So what is to be done? Haspelmath introduces this definition of dative: “A dative case is a morphological marker that has among its functions the coding of the recipient argument of a physical transfer verb (such as ‘give’, ‘lend’, ‘sell’, ‘hand’), when this is coded differently from the theme argument.” That is not, or only partly, inferred from the languages mentioned above: it is devised by Haspelmath qua typologist. Indeed, it turns out that the Finnish Allative is a dative, but the Nivkh Dative-Accusative is not, for it fails to meet the criterion.

    Now obviously if this typological definition had been manipulated to work solely with the languages it was based on, it would indeed be an exercise in futility. But Haspelmath thinks that it does in fact work well across a variety of languages, though not all. So why do comparison at all? Because it’s the only legitimately theoretical part of synchronic linguistics, given the non-existence of Chomskyite comparative categories. Everything else is descriptive-factual (and of course none the worse for that), because as David M says “everything is the way it is because it got that way”. So here’s Haspelmath between Scylla and Charybdis, or Ossa and Pelion, or whatever, brightening the corner where he is.

  147. Drasvi: I don’t think “take me by the hand” means ‘take my hand’, more like ‘lead me by holding my hand’.

  148. Lars Mathiesen says:

    EDIT: I tried to update the thread before replying to drasvi, but it didn’t show anything new. But after posting there are 14 comments since then, so read this as a reply to Y just above instead…

    You can’t tell if it’s accusative or dative, since English only has an oblique case (in pronouns) besides the nominative. But if you analyse it as containing Take me, then me is accusative by function but not a possessor — conversely, if it’s equivalent to Take my hand, then it’s a possessor but you might as well call it a dative. Note that you can say Take me by my hand (I think) without losing me.

    Now this makes me notice that Danish has Jeg gav ham hånden = ‘I shook hands with him’ or ‘I gave him my hand’ — where the possessor is wholly implicit. You can use a possessive pronoun, though that’s very high style: Jeg giver dig min hånd derpå!, but not an external possessor. (The locative preposition has a special meaning with this construction, in case you were wondering, you can give hånd på a deal or a promise to confirm it, and derpå is one of the associated deictic pronouns).

    Så giv da hinanden hånd derpå! is what the pastor says where the lewd Anglicans have You may now kiss the bride. (Also less redolent of the patriarchy and all). Here derpå refers to the vows in their entirety.

  149. Take me by my hand is very odd. I defer to native speakers to say if it’s at all acceptable. It would take some odd scenario to compel one to use it.

    Take my hand means ‘hold my hand’. Take me by the hand means ‘lead me (somewhere), utilizing my hand’. Two different senses of “take”.

  150. Lars Mathiesen says:

    Take me by my hand may be weird, but Take by my hand is just wrong. I think. If native speakers concur, I think that shows that me is not a possessor in that construction.

    @Y, I’m not disagreeing with you, I just like my argument better.

  151. Take by my hand is just wrong. I think. If native speakers concur, I think that shows that me is not a possessor in that construction.

    Why?


    In Russian:
    (stress in “by hand” shifts from hand to the preposition)
    Take!
    Take by hand! – Not ungrammatical. Omitted direct object can be, for example, someone we both are looking at. Say, we are helping our drunk freind and after several clumsy attempts I advice you to take his arm.
    Take-self by my hand – no comments (it is common)

    Take by his hand – clumsy, but not terrible. See above.
    Take by my hand – oops.

  152. Trond Engen says:

    Take me by hand.

  153. David Eddyshaw says:

    I don’t think there is a problem with the grammaticality of “take me by my hand” as such; what confuses the issue is the existence of “(take me) by the hand” (and so forth) as an idiom. “Take me by my hand” would appear in a case where the literal or specific nature of the act was being emphasised, perhaps in the context of dance instruction, say, or perhaps in correction of some oaf who was attempting to take the speaker by the elbow.

    The less common the taking-by-that-appendage is, and (consequently) the less stereotyped the expression, the more readily the possessive is likely to be included. The use of more specific and less idiom-generating verbs than “take” also makes this more likely. “They strung him up by his heels.”

    [I wish to preemptively rule out examples involving Donald Trump; observing merely that if he had been referring to a specific individual rather than half the human race, he would have been more likely to use a possessive pronoun. On the other hand, he is a dehumanising creep, so perhaps not. Appendages float free of any actual personal ownership in his mindscape.]

  154. David Eddyshaw says:

    An analogy to the “by the hand” structure in Kusaal:

    M kɛŋ nɛ nɔba.
    1sg go with leg:PL.
    “I’ve gone on foot.”

  155. Finnish examples:

    Pojan lähestyessä miestä tämä tarttui häntä kädestä ja veti pojan lähemmäksi.
    ‘As the boy approached the man the latter grasped/grabbed him by the hand and drew the boy closer.’

    häntä: him/her
    kädestä: from hand

    https://en.wiktionary.org/wiki/tarttua
    käsi

    Hänen mukaansa mies oli ottanut häntä kädestä, vetänyt väkisin metsään ja pakottanut sukupuoliyhteyteen.
    ‘According to her, the man had taken her by the hand, led (her) forcibly into the woods, and forced (her) into intercourse.’

    https://en.wiktionary.org/wiki/pakottaa
    https://en.wiktionary.org/wiki/sukupuoliyhteys

  156. Dark are Finnish woods and so are their exmaples…

  157. I have a feeling that some linguists understand this construction (“DEP”) as I do, while some understand it differently. They do not give definitions, and I can not understand what I am dealing with. When I read that Russian dative in “babysat me children” is a possessive NP, expressing possessor, my Russian jaw drops and face palms. Either it is:

    – wrong intuition in their part (they misunderstand what Russians mean here), or
    – wrong analyzis, or
    – a specific analyzis/definition.

    The latter option does not help, because in literature one can find more than one description. My problem is rather practical: I have no idea what exactly Haspelmath is comparing, or what John Cowan means here.

    John groups as DEP this The neck to him [Finn MacCool] was as the bole of a great oak, knotted and seized together with muscle-humps and carbuncles of tangled sinew, the better for good feasting and contending with the bards. The chest to him… (in a different thread) and Haspelmath’s DEP.

    But Haspelmath’s Russian exmaples are experiencer/affectedness. And I know that in Spanish it is different – how I am supposed udnerstand Basque then, if they all are “possessive NP’s expressing ‘my'”?

  158. As for “intuition”, enough to say that for me dative in the oral sex example feels about the same as in the handwashing example. They both imply body parts, but what dative expresses is the same thing both times for me: experiencer/affectee. Of course hands (or penis) are somehow implied irrespectively of whether they are mentioned (wash hands) or not (make blowjob).

  159. Another example:

    Hän huusi, että joku puree häntä kädestä.
    ‘She screamed that someone/something bit her on the hand.’

    https://en.wiktionary.org/wiki/purra#Finnish
    https://en.wiktionary.org/wiki/joku#Finnish

  160. David Eddyshaw says:

    I think there are two sources of confusion.

    Firstly:

    English “I wash my hands” and French “Je me lave les mains” have quite different structures: in the English, “my” is a possessive pronoun and part of the NP object of “wash”, in the French, “me” is an indirect object, and is not part of the NP “les mains” in any way.

    However, many Chomsky-descended approaches to syntax operate at a level of abstraction so rarified that it will describe the French as a sort of transformation of an underlying Englishoid structure (never the other way round, for some reason): hence all this stuff about “raising.” Those of us who have not drunk the Kool-Aid see no reason to suppose that anything has been “raised” at all. It is where it is. Basta. (More sophisticated latter-day Chomskyan approaches don’t entail any sort of actual movement, but by playing around with the labelling of verb arguments for individual verb types they end up in the same place, postulating a Super Secret Real Structure not visible to profane eyes, which speakers nevertheless use all unknowingly.)

    Secondly:

    The “possessive” relationship within NPs is itself neither a single homogeneous thing, nor does it copy over automatically between languages. (As I was saying above, a “possessor” in a Kusaal NP has some properties more reminiscent of a dative than a genitive*; many languages have different kinds of possession, such as alienable and inalienable, and sometimes even more subtypes.)

    So when Chomskyites and their camp-followers talk of “raising”, they are not necessarily talking about raising exactly the same kind of constituent. Even. That’s before you enter into further complications regarding verb transitivity, whether you can add indirect objects without formally changing the verb into an applicative, and the fact that there is more than one kind of applicative anyway.

    It’s hardly surprising that the resulting accounts are confusing-to-incoherent.

    *The exact same formal structure in Kusaal is also used for the vast range of meanings expressed by English “of”, as in e.g. salima bʋtiŋ “gold cup” (“cup of gold”), where salima “gold” is a noun, not an adjective.

  161. David Eddyshaw says:

    In Jarawara inalienable-possessive constructions, the possessor is the head (and the possessum agrees with it in gender.)

  162. @David Eddyshaw, thank you. If I were sure that it is what Haspelmath means, no problem (well, there is a problem: those theories of syntax evolve). I just need to know what people mean. But “DEP”, I think, originated with ascension/raising but is not based on it now.

    The problem with Chomsky is the echo. I am fine with that girl-professor referenced by Bathrobe. She LOVES that syntax and it makes me curious to know what he loves about it. But the trace Chomsky left in descriptive linguistics is a mess, honestly.

  163. A naïve analyzis that leads to “expresses possession” is this. variables:

    NOM VERB DAT ACC.

    now fix the values for 3 of them:

    she washes DAT hands

    When DAT is an unknown, you find yourself wondering whose hands are these. Now why these 3 values first? And why do you ask yourself this particular question (as opposed to “who is affected” or “who is the second participant”)? Are you sure that it is not because of your own native langauge?

    A counter-analyzis (also naïve) would be this. We fix the situation. We know what is going on. We know what we refer to and what action we reconstruct from the utterance. Then of all utterances that refer to this situation I have at least these two:

    she washes to-me hands
    she washes my hands.

    They differ in perception and maybe attitudes.

  164. David Marjanović says:

    Take me by the hand = nimm mich bei der Hand, where mich is unambiguously accusative.

    as David M says “everything is the way it is because it got that way”

    From D’Arcy Wentworth Thompson’s (1918) book On Growth and Form, an early and very theoretical attempt at evo-devo.

    though that’s very high style: Jeg giver dig min hånd derpå!

    I guess it’s another calque from German: Ich gebe dir meine Hand darauf!

    Omitted direct object can be, for example, someone we both are looking at.

    Germanic languages seem to insist on mentioning all direct objects, even if they’re reduced to clitics that consist of a lone consonant.

    The neck to him […] The chest to him […]

    That’s not at all like German. Is the original Irish more like Latin, where dative + esse means “have”?

  165. David Eddyshaw says:

    Germanic languages seem to insist on mentioning all direct objects

    Kusaal can omit retrievable direct or indirect objects, which complicates the analysis rather; however, you can tell that a verb is ditransitive by the fact that if only one object is explicitly present, it has to be interpreted as indirect:

    M tis biig la daka.
    “I’ve given the child a box.”

    M tis biig la.
    “I’ve given it to the child.”

    M tisya.
    “I’ve given it to her.” (This ya follows all clause-final perfective forms which carry the tone overlay marking the clause as independent: it’s not a pronoun.)

    ?? M tis daka.
    “I’ve given it to a box.”

    Monotransitive and intransitive verbs can add an indirect object as an “ethical dative”; naturally there is no anaphoric implication if it’s omitted:

    Li malis biig la.
    “The child likes it.”

    Li malis.
    “It’s sweet.”

  166. Omitted direct object can be, for example, someone we both are looking at.

    I think in this case there have been several attempts to take this drunk guy by various body parts. But anyway, when I offer an apple, the imperative is just “take!”.

    And:
    “I took him by right hand a she by left.”.
    or
    “I took him by right hand a she took by left.”.
    by taste.

    a is a conjunction used (among other things) for [T1 C1]a[T2 C2]:

    I sober a Vasya drunk
    sober I a drunk Vasya (the one who’s sober is me…)
    here cold a there warm
    I am-cleaning room a she is-baking pie

    Not necessarily contrasting. Accordingly:

    Yesterday David Eddyshaw was-working over grammar of-langauge Kusaai, a today over Klingon

    Of course one can repeat the long part. But not:

    *….a today David Eddyshaw over Klingon
    only
    ….a today David Eddyshaw is-working over Klingon
    and again not:
    *….atoday David Eddyshaw is-working Klingon

  167. PlasticPaddy says:

    @dm
    For modern irish:
    In the standard language, only two words Éire (“Ireland”) and fiche (“twenty”) have distinct datives – Éirinn and fichid, respectively
    Dative appears after certain prepositions.
    Prepositions that govern the dative: a/as, do, de, ar, ó, ós, ag; and ar, fá and i except when used with verbs of motion, in which instances they govern the accusative.
    There is no bare dative, i.e. “he gave me the book” is “Thug sé an leabhar dom”, where dom = the preposition do + first-person pronoun “clitic”. As you have guessed “to have” = “to be”+ dative (using the preposition “ag”) and is used also with intangibles: “Tá Laidinis aicí” = “she knows (lit. has) Latin”. My maleficient dative uses the preposition ar, which also enters into phrases like “cuireann sé sin iontas orm” = “that surprises me (lit. that puts surprise on me”). I would have considered that last accusative, but cuir is not strictly a verb of motion (or is it?).
    Old Irish did not use a bare dative (or you have to go very far back for examples of such usage). So those JC examples would have used a preposition (I think “ag”).

  168. David Marjanović says:

    In the standard language, only two words Éire (“Ireland”) and fiche (“twenty”) have distinct datives – Éirinn and fichid, respectively

    Certe est, quia absurdum.

  169. David Eddyshaw says:

    Here’s to Ireland! Here’s to the Twenty!

  170. Old Irish did not use a bare dative (or you have to go very far back for examples of such usage).
    If I remember my Thurneysen correctly, the bare uses of the dative in Old Irish are instrumental, due to the old PIE instrumental having been merged into the Old Irish dative.

  171. Certe est, quia absurdum.

    Nota bene the words “in the standard language”. The province of Munster is the home of many things linguistic, including:

    (a) a functioning dative case, which is why the English insisted on calling the Island of Doctors and Saints Erin, as they tended to hear it most often in dative constructions like in Eirinn ‘in Ireland’, go hEirinn ‘about Ireland’, and faoil Eirinn ‘from Ireland’;

    (b) an indecent superfluity of diphthongs;

    c) stress on the first heavy syllable (including those in -ach for whatever reason) instead of proper ruthlessly initial stress;

    (d) swearing by one’s baptism, palms, or oddness (all now meaning ‘Indeed!’ or ‘My God!’);

    (e) a suspicious resemblance not only to Manx but to Scots Gaelic, and

    (f) the silly notion that the name Eoghan ‘born from the yew’, as in County Tyrone / Tír Eoghain ‘land of Eoghan’ and my own name Eoghan Mac Eoghain, is < Eugenius (I am a Connachta and know better).

    Here’s to the Twenty!

    I think you mean the Twenty-Six, if not the Thirty-Two.

  172. David Eddyshaw says:

    I think you mean the Twenty-Six, if not the Thirty-Two.

    But that wouldn’t be distinctively dative, and sub specie aeternitatis, is that not what really matters?

Speak Your Mind

*