A couple of interesting topics came up in this thread, and since juha said “I think dative objects in Icelandic is a topic worthy of a separate thread,” I thought I’d start one.
Verb classes and dative objects in Insular Scandinavian, by Jóhannes Jónsson:
Dative case in Insular Scandinavian (Icelandic and Faroese) exemplifies a fairly complicated relation between syntax and lexical semantics. Thus, monotransitive verbs selecting dative objects in Icelandic fall into various semantic classes and many of these classes also contain verbs with accusative objects (Maling 2002). The same is true of Faroese although the number of two-place dative verbs in that language is much smaller than in Icelandic. The reason is that dative objects of many verbs have been replaced by accusative objects in the history Faroese and this process is still ongoing.
Despite the complexities surrounding dative case selection in Insular Scandinavian, it is clear that some lexical semantic features are more strongly associated with dative case than others. This is also true cross-linguistically as can be seen by comparing two-place dative verbs across languages.
Dative in Icelandic: throw that ball!
And for general linguistic oddity, courtesy of PlasticPaddy, Quirky subject:
In linguistics, quirky subjects (also called oblique subjects) are a phenomenon where certain verbs specify that their subjects are to be in a case other than the nominative. These non-nominative subjects are determiner phrases that pass subjecthood tests such as subject-oriented anaphora binding, PRO control, reduced relative clause, conjunction reduction, subject-to-subject raising, and subject-to-object raising. […] Icelandic was argued to be the only modern language with quirky subjects, but other studies investigating languages like Faroese, German, Hindi, Basque, Laz, Gujarati, Hungarian, Kannada, Korean, Malayalam, Marathi, Russian, Spanish, and Telugu show that they also possess quirky subjects.
There’s even a Quirky Subject Hierarchy (QSH).
I’m baffled looking at the Wikipedia article at how the German „Ihm gefällt das Bild.“ supposedly has an oblique subject, but Spanish “Le gusta la foto.” does not.
The distinction seems to be a quibble at best over how the respective languages handle other features.
I’ve come across the passage below—illustrating different subject marking—in a book on adpositions by Claude Hagège:
(INV: involuntary)
More on Sinhala subject case marking
(I formatted the quote for better comprehensibility.)
Thanks!
The results from the two large-scale surveys discussed in this paper show that
a novel transitive verb in Icelandic takes a dative object if it (a) encodes some
kind of caused motion of the object referent, or (b) has a translational substitute
that takes a dative object. If neither (a) nor (b) holds, the object gets the default
accusative case.
Chapter 12 of Dative constructions in Romance and beyond
https://langsci-press.org/catalog/book/258
And there are alternations in ergative/active languages such as to slip: to glide/skate with absolutive vs ergative subjects:
Split-S in Tabasaran with person/number agreement
(31)
ʁurcʷura-za-wu
beat-1sgA-2sgP
‘I beat you.’
(32)
a. aldakura-zu
fall.down-1sgP
‘I fall down.’
b. urgura-zu
burn-1sgP
‘I burn.’
c. kc:ʷuχura-zu
slip-1sgP
‘I slip.’
(33)
a. aIlqura-za
laugh-1sgA
‘I laugh’
b. kc:ʷuχura-za
slip-1sgA
‘I skate.’
Ergativity_in_East_Caucasian
https://ilcl.hse.ru/en/east_caucasian/lectures
þágufallssýki
Icelandic
Etymology
From þágufall + sýki.
Noun
þágufallssýki f (genitive singular þágufallssýki, no plural)
1. (grammar) The act of erroneously using the dative case in place of another grammatical case, usually the accusative case.
erroneously…
Peevers of the world, unite!
þágufall /ˈθauːʏˌfatl/! Man, Icelandic spelling is not exactly transparent…
Accusative is the default (structural) case on objects in Icelandic but the number of
verbs taking direct dative objects is quite high, compared to e.g. German or Faroese.
According to Maling (2002), verbs with direct dative objects in Icelandic are almost
500. (For information about the token frequency of accusative and dative objects in
Icelandic, see Barðdal 2008.)
https://malvis.hi.is/sites/malvis.hi.is/files/San.Sebastian.2010.pdf
@Craig, the test that the German construction passes but the Spanish supposedly does not, is that Der Dame gefällt das Bild von sich is grammatical, and (reflexive) sich can only bind to the subject of a sentence. Spanish does not have a specific subject-relative pronoun, I think, so you can’t do the test. Notice that German has a difference between reflexive and non-reflexive that Spanish doesn’t, but I don’t know if there are other ways of testing the subject status of le that positively fail:
Der Dame gefällt das Bild von sich
A la dama le gusta la foto de ella misma
Ich gebe der Dame ein Bild von sie selbst
Le doy a la dama una foto de ella misma
(I will now be corrected by people who actually know German and Spanish).
I would think that would have to be “von ihr selbst”.
My Spanish isn’t strong but I’m pretty sure that “de sí misma/o” is the counterpart of “de ella/él misma/o”.
According to the Wikipedia article’s quirky subject criteria, if Spanish ‘gustar’ takes a quirky subject, a coördinated phrase like “Le gusta la foto y come el queso”* should be grammatical and mean “He likes the photo and eats cheese,” since syntactic subjecthood is necessary for the ellipsis in the second phrase. I’m guessing that this construction isn’t grammatical in Spanish, which is why ‘gustar’ isn’t a case of quirky subject, but rather would be a verb that takes an argument with subjecty semantics in a non-subject position.
This reminds me of the French verb ‘manquer,’ and the construction ‘Tu me manques.’ Even though an idiomatic translation would be ‘I miss you’ (with the syntactic positions of the two participants inverted), you couldn’t say, ‘Tu me manques et t’aime’ and mean ‘I miss you and love you.’
*I do not speak Spanish.
Does a compound predicate with mixing of verb types work in German? If I try to coordinate a non-quirky-subject verb with, “Der Dame gefällt das Bild,” it sounds wrong to me. However, I’m not a native, and my opportunities to use colloquial spoken German are even fewer these days than they had been in the past. (I don’t think I would ever put a reflexive at the end of a sentence link that, although it doesn’t sound exactly wrong to do so.) So what do the native speakers think?
Does it sound better if the quirky verb comes second? Something along the lines of “Die Dame schwimmt und gefällt das Bild”?*
*I speak German even less than I speak Spanish.
ə de vivre,
that German sentence is ungrammatical. And it’s an example of coordination, which has some quirks in itself.
In case it has not been mentioned yet, there is a whole book on the subject of quirky subjects.
You omitted the link.
ə de vivre, it would have to be “Die Dame schwimmt, und ihr gefällt das Bild.” Though to my ear, it would sound better as “Die Dame schwimmt, und das Bild gefällt ihr.”
Dang my greasy fingers! Here is the link for the book: https://benjamins.com/catalog/slcs.200.
Der Dame gefällt das Bild von sich
Actually, I can’t decide whether this sounds grammatical or ungrammatical to me.
On the coordination test, not only Spanish fails it, but German, too.
On the whole, calling the dative experiencer “subject” (even if quirky) looks like a mix-up of formal sentence roles (subject / object) with pragmatic roles. Formally (based on case, verb agreement, coordination) the picture is the subject here.
looks like a mix-up of formal sentence roles (subject / object) with pragmatic roles
I wondered about that.
I think this is piggybacking on all the work that has been done on ergatives, starting with Dixon, who clarified that an ergative subject really is a subject (as opposed to an instrumental, going with a basically passive verb) by looking at phenomena like syntactic pivots and the behaviour of reflexives. So it’s not simply a matter of pragmatics, but also of yer actual formal syntax, at least potentially.
The fact is, though, that the many and various criteria for “subjecthood” that people have come up with over the years don’t always align with one another (which is probably not all that surprising on first principles.)
Kusaal quite often has indirect objects where English has subjects, e.g. Li malisim (“I like it”, literally “It’s sweet for me”), but AFAIK such objects don’t behave syntactically like subjects at all; they aren’t omitted after ka “and”, for example, which is the rule for a personal pronoun when it refers to the subject of the immediately preceding clause:
Li malis o ka o bɔɔd ye o di.
it be.sweet her and she want that she eat
“She likes it and she wants to eat it.”
not
* Li malis o ka bɔɔd ye o di.
which would (I think) in fact have to be understood as “She likes it and it wants her to eat it.”
Georgian verbs can have
ergative subject, nominative direct object
nominative subject, dative direct object
dative subject, nominative direct object
depending on tense, mood and whether there is an R in the month.
Yes, that’s a truly stimulating language. It’s high time I tackled it again.
Same feeling here, but I think it is because the construction “das Bild von sich” (instead of “ihr Bild”) is normally considered to be part of uneducated usage.
I completely agree here. If an alleged subject passes all those “subjecthood tests” but fails the most important one — the subject has to be in the nominative — could it simply be that all those subjecthood tests are unrealiable or even simply invalid?
Croatian has the same thing that the Wikipedia article is talking about.
I won’t be updating the article though. I’ve given up updating Wikipedia on Croatian topics, due to its anti-Croatian bias. Editors regularly dispute the existence of Croatians before 1990, as well as the existence of the Croatian language. If I write that the phenomenon exists in Croatian and give examples, some f***wit will change it to Serbocroatian.
Double standards are alive & well on the English language Wikipedia though: Hindi is on the list of languages (in the Quirky Subject article) and no one has changed it, or is likely to change it, to Hindustani or Hindi-Urdu.
@ulr, the German examples have to use von sich because the possessive determiners don’t make the reflexive/non-reflexive distinction. (Unlike Danish sit vs hans, but we turned all the dative subjects into nominatives back in the 14th or so).
@ə: Yes, German fails the coordination test, but according to the Quirky Subject Hierarchy outlined in the WP article, lots of languages only do the reflexive reference bit and not the others like coordination — so it’s got lowkey quirk, not full-on quirk.
@Craig: I’m relying on GT to supplement my Spanish, but if these are both correct:
[ella] toma una foto de sí misma
le gusta la foto de ella misma
it does indeed demonstrate that the dative pronoun is not a subject. If you can say
¿le gusta la foto de sí misma
then Spanish has quirk–but GT refuses to produce it.
Me neither. I’ll try to wonder about this at greater length tomorrow.
And of pragmatic roles with English formal ones.
It smacks of Chomskyan introspection. I am naturally tempted to turn it on its head, declare German to be pure logic, point out the fact that das Bild gefällt mir really is something that the picture does to me – it impresses an emotion on me –, and declare English “quirky” for pretending that I have any agency when I like the picture.
The largely completed replacement of mir hat geträumt by ich habe geträumt would then reflect the change in the interpretation of dreams from visions (which happen to you because an outside agent throws them at you) to your very own brainfarts.
Oti-Volta substrate under the Balkan Sprachbund confirmed. It logically follows that Serbian, and very much Croatian, really is the oldest language because it’s underlyingly KONGO.
I think Haspelmath has written somewhere that “subject” doesn’t real – only “agent”, “patient”, “voluntary experiencer” and “involuntary experiencer” are universal, and languages differ in how they merge these roles (notably, none seems to distinguish all four, and distinguishing three is extremely rare). SAE puts agents and at least voluntary experiencers in the nominative and patients in the accusative; Basque, which used to be interpreted into the 20th century as “all verbs are passive”, puts agents in the ergative, patients and experiencers in the absolutive, and the instruments that the agents use in the instrumental case, proving that agents are not treated as instruments. (You just have to study Basque for longer than the seven years the devil wasted on it.)
There’s a third option: von ihr.
That’s not the-genitive-its death, i.e. “of her”, but more like “from her” and therefore grammatical in the standard. I’ll think more about it later.
And of pragmatic roles with English formal ones.
It smacks of Chomskyan introspection.
I had thought about writing something in this vein, but decided that would be too uncharitable for Christmas 🙂
@ulr: I agree with DM here, the von is not simply the genitive replacement but a prepositional construction licensed by Bild (or similar words like Porträt, Foto) that is used even in registers where the genitive is normal, simply to distinguish the roles of “depictor” and “depicted” – in Peters Bild von Paul Peter is either the owner or the creator of the picture, while Paul is the one depicted.
And I have no problem with Der Dame gefällt das Bild von ihr.
Added data point: I would interpret a sentence Der Dame gefällt das Bild von sich, das Maria gemalt hat as “The lady likes the picture of Maria that Maria has painted.” While if it would be das Bild von ihr, then my default interpretation would be that it’s a portrait of the lady.
I’m surprised no-one has mentioned phrases like качать головой ‘shake one’s head’, управлять машиной/самолётом ‘drive a car/fly a plane’, жонглировать шарами ‘juggle balls’, дирижировать оркестром ‘conduct an orchestra’, where direct (?) objects are in the instrumental case.
Swedish has something like it as regards body parts: skaka på huvudet ‘shake one’s head’, rycka på axlarna ‘shrug one’s shoulders’, röra på huvudet/på sig ‘move one’s head/move oneself.’
Asya Pereltsvaig looks at some of the Russian phenomena in her thesis.
There’s also the possessive dative. Here’s an example from Old English:
þæt him þæt hēafod wand forð on ðā flōre
so that to him the head fell down to the floor (=his head fell to the floor)
FYLOSC has this too, and I’m sure many other languages do as well.
(…tako da mu je glava pala… where ‘mu’ is DAT)
This can lead to ambiguity, such as in the following real-world example:
I onda mu se dete upiškilo u cipelu.
(and then to him the child peed in to the shoe)
It is unclear whether the DAT ‘mu’ refers to his child or his shoe or both. From the context you can presumably tell whether the child in question is his (as opposed to just some random child), but the question remains whether the shoe was also his or the child’s (i.e. whether the child peed into their shoe or their dad’s shoe).
@juha: That’s another good example of different languages encoding the same activity in different ways. German shows an “instrumental” construction in some of these cases as well, often depending on the verb: mit dem Kopf wackeln “to wobble one’s head”, but den Kopf schütteln “to shake one’s head”. It’s fine to say that some languages encode an activity with an instrumental construction while others use transitive verbs with direct objects, but it would be problematic in my view to call the instrumentals “quirky direct objects” just because English has constructions with direct objects in those cases.
I had a chance now to read the WP article on quirky subjects. The Icelandic material really is interesting, also its comparison to Faroese. I hope the examples are all correct, because the German examples often are borderline ungrammatical or at least quite unidiomatic, so I’d advise to treat all the conclusions about quirky subjects in German with caution.
Isn’t this rather a dativus (in)commodi than a dativus possessivus? In German at least, “Ihm fiel die Tasse zu Boden” (definitely a dativus incommodi) and “seine Tasse fiel zu Boden” do not necessarily mean the same thing. In Latin, the dativus possessivus always accompanies forms of esse, I think.
Re possessive dative:
ulr,
“Isn’t this rather a dativus (in)commodi than a dativus possessivus?”
In some languages, yes. But a Dative personal pronoun is one of the hallmarks of the Balkan Sprachbund and, as it so happens, it also occurs in my dialect of Eastern Slovak (way outside of the Sprachbund, which goes to show). Your examples have the same, ehm, surface form:
(1) “Findža mu spadla na žem.” (cup he.DAT fall-PART_PASS.F on ground)
But it still allows for two different readings, “ihm fiel” and “seine Tasse fiel”.
There are contexts where the Dative pronoun is unquestionably a possessive, e.g.
(2) “Ocec ci co robi?” (father you.DAT what do.3SG.PRES) = “What does your father do / is your father doing?”
All of this involves a rather complex interplay with agreement, between full forms and clitics and between same-subject reference (reflexive) and different-subject reference pronouns.
Also in Modern Greek:
Μεθαύριο θα πάω στο νοσοκομείο να μου κόψουν τα ράμματα.
The day after tomorrow I’m going to the hospital to have (me) the stitches taken out.
(The MG genitive has to stand in for the AG dative as well.)
juha,
in Slovak, we do have the Instrumental thing with body parts, but it pretty well matches with other uses of the Instrumental. So e.g. “triasť hlavou” for “shake one’s head” is the same construction as, say, “triasť fľašou” = “shake a bottle”. In fact, both are also subject to the colloquial/dialectal shift where the Instrumental requires the preposition “s”.
David,
and declare English “quirky”
Word. Bro, English quirky af. Just think about the whole negation thing.
All of this reminds me of the discussions I had when I was putting together the UD treebank of Maltese. I had some doubts concerning the status of verbal dependents which behave like objects semantically, but invariably are introduced by a preposition. All my colleagues who are native speakers of English were like “those are some non-object/oblique dependents”. All my Czech colleagues “yeah, definitely prepositional objects“.
(2) “Ocec ci co robi?” (father you.DAT what do.3SG.PRES) = “What does your father do / is your father doing?”
Yes, “Šta ti radi otac?” normally means “What’s your father up to?” (as opposed to “What is your father doing to you?”)
Andrej,
but note the word order, both that of the pronoun and its head, as well as that of the constituents.
(3) “Co ci robi ocec?” is also possible in standard Slovak as
(4) “Čo ti robí otec?”.
But both are ambiguous and can also mean “what is dad doing to you”. (2) cannot be read that way. A clearer example is a copular clause with such an NP as a subject:
(5) “Ocec ci nuka?” (father 2SG.DAT inside) = “Is your dad inside?”. This does not work in standard Slovak (even with a copula verb) and would immediately be read as dialectal.
And then there’s the 2SG.DAT pronoun in a presentative function, as in
(6) “A on ti tam len tak sedí.” (and he 2SG.DAT just so sits) = “And he’s just sitting there.”
The presentational dative is also possible in colloquial German: Und er sitzt dir da so rum.
If an alleged subject passes all those “subjecthood tests” but fails the most important one — the subject has to be in the nominative — could it simply be that all those subjecthood tests are unrealiable or even simply invalid?
Since “subject” is a syntactic concept, if something behaves syntactically like a subject but is morphologically marked differently from normal subjects, it’s sensible to think of it as an anomalous kind of subject.
Pragmatics doesn’t come into it, at least not necessarily: the argument that dative experiencers are quirky subjects in German but not Spanish is obviously not based on thematic roles but on syntax. For that reason I don’t think this is a case of blindly imposing English categories on other languages.
I agree.
…that, too, is of in English: I took a picture of her…
Agreed.
Yes. I don’t think I’ve seen an actual Latin-style dativus possessivus – mihi est “I have” – anywhere in Germanic.
Yup, that’s one, and I’d have completely misunderstood it (not necessarily as “what is he doing to you”, but at least as “what is he doing for you”).
Speaking of Balkan… oh, no, I was thinking of fildžan, which is what you drink
TurkishGreekSerbian coffee out of.That’s a dativus commodus again: um mir die Nähte herausnehmen zu lassen.
“Tun tut man nicht!”
“Tun tut man nicht tun!”
Quirkiest partitive I have ever seen.
Yup. It’s a dativus commodus: the listener is asked to imagine the situation vividly and have emotions about it. “Can you believe it? He’s just sitting there as if nothing had happened.”
fildžan, which is what you drink [Mediterranean] coffee out of
In Israeli Hebrew, findžan ended up as the word for the long-handled kettle used to make
ArabicArmenian[controversially named] coffee. This is taken to be due to a misunderstanding, though it’s not clear where and how that misunderstanding took place.Re: the Old English example with the head above, it’s from a Richard Hogg textbook. This is what he says about the example:
There is a special use of the dative in Old English to express possession: […] where the possessive pronoun has been replaced by him in the dative case. This can be compared with the French construction à la main ‘in his hand’. It is, of course, true that the French construction is rather different. But in one crucial respect it is parallel, for it shows the use of a special construction: possession involving parts of the body.
I accidentally the most important word.
This is just beyond my Spanish, I can see it through the fog in the distance…
Der Dame gefällt das Foto von ihr
Ich gebe der Dame ein Foto von ihr
are unremarkable;
Der Dame gebe ich ein Foto von ihr
is also unremarkable as a topic-and-comment sentence.
It’s not blind. It just betrays a profound unfamiliarity with the empathic dative/dativus (in)commodi and often with the languages in question.
I can’t find it on my harddisk, but I’m sure I once read a paper on supposed dative subjects in Icelandic and German with the conclusion that they’re not subjects in either language. I’m also sure Jóhanna Barðdal was involved; her publication list back to 2012, with links to most of the papers, is here, but I don’t have time to check them.
Yeah, bullshit. 😐 These datives are all empathic. The reason they’re lost in later English is not some morphosyntactic change, but the phonological fact that the dative has merged with the accusative and (except for pronouns except complications) the nominative.
Given that the possessor is already expressed by the empathic dative, it doesn’t need to be mentioned a second time, but can be and usually is left to context – Latin takes it a step farther and also leaves the possessor in “his friend” to context.
It’s not blind. It just betrays a profound unfamiliarity with the empathic dative/dativus (in)commodi and often with the languages in question.
How so? The argument isn’t “these datives are subjects because they stand for experiencers” but “these datives are subjects because they behave syntactically like subjects”.
juha, I am intrigued, what is surprising about instrumental case in your examples?
Eg, why is it качать головой ‘shake one’s head’ (Instr), but наклонить голову ’tilt one’s head’ (Acc)? OK, in чесать репу/голову ‘scratch one’s noggin’ (Acc), it’s not the head that’s doing the scratching, but the pair above looks pretty similar. It seems that one has to learn the verbs individually, with no obvious generalizations.
How so? The argument isn’t “these datives are subjects because they stand for experiencers” but “these datives are subjects because they behave syntactically like subjects”.
That approach (at least as outlined in the WP article) still ignores some markers of subjecthood (case, verbal agreement) and privileges other syntactical behaviours in the definition of what a subject is. The ones that are ignored are markers that are relatively unimportant in English compared to German or Icelandic (formal case, verbal agreement), while the ones privileged are syntactic markers, because those are important in English.
Now, what we can see is that, yes, certain constituents in the dative case do share syntactic behaviours with nominative subjects, which probably has something to do with their semantic / pragmatic role as experiencer, but is that enough to override the other markers and declare them subjects in place of the actual constituents that show the right case and the verbal agreement?
I don’t know Icelandic or Faroese, except that I can guess at deciphering the examples based on their similarity to other Germanic languages, but the bad quality of the German examples used give me low trust in the familiarity with the analysed languages of who came up with the analysis.
It seems reasonable to give greater weight to syntactic criteria when defining what’s essentially a syntactic concept. But in any case the point isn’t really to pronounce on whether such and such a constituent is a subject or not; the point is precisely that features of “subjecthood” that bundle together in English don’t necessarily so bundle in other languages. (Much of the early work on this phenomenon was done by native speakers of Icelandic, BTW.)
A thematic-role account doesn’t really work here, because the syntactic constructions in question don’t care about thematic roles: e.g. ellipsis can occur with a subject that’s agent, patient, or experiencer, but it can’t occur when those same roles are expressed by a non-subject.
I was thinking of fildžan, which is what you drink Turkish Greek Serbian coffee out of.
For Greek coffee, you’d have to drink it out of a φλιτζάνι.
þágufall /ˈθauːʏˌfatl/! Man, Icelandic spelling is not exactly transparent…
Here you can hear it—just click on the Framburður button.
þágufallssýki pronounced
It seems reasonable to give greater weight to syntactic criteria when defining what’s essentially a syntactic concept
But that’s exactly the question here. Are some fringe syntactic conditions more important in determining subjecthood than markers like case and verbal agreement that are there in any given sentence?
But in any case the point isn’t really to pronounce on whether such and such a constituent is a subject or not;
Well, calling something a “quirky subject” very much looks like doing exactly that to me. Maybe another name for the concept would be helpful, then.
the point is precisely that features of “subjecthood” that bundle together in English don’t necessarily so bundle in other languages
That’s something I can agree with.
Hans, have you seen “Dative Sickness” in Germanic?
@juha
Movements involving body parts are complicated. The case used seems to depend on whether (a) the movement is (perceived as) a full body movement with outward manifestation in the particular part, (b) whether the part is considered capable of independent action,(c) etc. (eg., is the movement extended, repetitive or part of a mutual or coordinated act?). So in your example in German you have both den Kopf schütteln (acc.) and mit dem Kopf winken/wedeln (prep. + dative–no instrumental available), also in die Hände klatschen for “clap one’s hands” (prep. + accusative–in+dative exists, but not for this expression).
Polysynthetic languages often do body-part noun incorporation even if they don’t do any other kind, with the the possessor of the body part typically turned into a direct object: “They head-wounded him.”
Kusaal isn’t polysynthetic, but it has a whole template of idioms of the form “verb indirect-object generic-noun” on the pattern
Ba nwɛ’ na’ab la nu’ug.
they hit chief the hand
“They’ve pleaded with the chief.”
This particular example is homophonous with “They’ve hit the chief’s hand”, but in general the constructions differ tonally, and also because some of the object pronouns are different segmentally from the corresponding proclitic possessives:
Ba nwɛ’ɛf nu’ug.
they hit.you hand
“They’ve pleaded with you.”
Ba nwɛ’ɛ fʋ nu’ug.
they hit your hand
“They’ve hit your hand.”
This discussion originally reminded me of when I first learned circa four decades ago that in German the idiomatic way to express the “experiencer-subject” version of English “I am cold” is “Mir ist kalt,” NOT “Ich bin kalt.” That’s not mentioned in the “Quirky subject” wiki article, but is mentioned in the https://en.wikipedia.org/wiki/Dative_construction article linked therefrom, together with an argument that an expanded version like “Mir war zu kalt um zur Kirche zu gehen” shows that “[m]ir behaves like a subject.” That “can control infinitives” test seems a bit different from the “quirky subject” tests for subjecthood, and I can’t figure out how persuasive it should be.
And in Russian it’s “Мне холодно,” just like German.
Neither of those are grammatical in Spanish; though they are intelligible, they are distinctly translationese.
To be idiomatic, the first would require a reflexive form (se toma una foto), the second a possessive determiner (le gusta su foto).
A la dama le gusta su propia foto.
Le doy a la dama una foto suya.
Both of these can be read as both subject and author of the portrait. To disambiguate you’d need a different construction, eg, a la dama le gusta la foto que le tomaron.
It seems that one has to learn the verbs individually, with no obvious generalizations.
It’s even worse than that. You have to learn verb-noun pairs with no clear rule. качать головой but качать ребенка. There are probably some subtle semantic clues that govern a good deal of the cases, but I cannot come up with one on the fly and it was probably researched to death by professionals.
A brief summary of noncanonical subjects and objects in Icelandic & German:
• Icelandic active sentences illustrate that we
cannot (necessarily) pair case and grammatical
function.
• Subjects need not be nominative.
• In passives, “non-structural” case is retained.
• In Icelandic, the passivized object behaves like non-structurally case-marked NPs in actives.
• Not so in German
• Nominative objects in Icelandic syntactically pattern like accusative objects.
• Nominative objects and embedded nominative subjects in ECM both allow for optional
agreement.
Dative Subjects and Nominative Objects
They’re totally groovy!
David,
Quirkiest partitive I have ever seen.
Oh we have that, too (both standard and Eastern Slovak, examples are from the latter):
(1) Pušči tam pľinu. (let-IMP.2SG there gas-GEN) “Let some gas in there.”
(2) Naľej mi paľenki. (pour-IMP.2SG me-DAT liquor-GEN). “Pour me liquor.”
I was thinking of fildžan
And you are right, except we kept the first n of “fincan”.
an argument that an expanded version like “Mir war zu kalt um zur Kirche zu gehen” shows that “[m]ir behaves like a subject.”
Funny, in English “It was too cold to go to church” works just fine with no semantic subject at all. As does “It was too cold for me to go to church”, with a non-subject controller.
@Alon, I know the examples are contrived, and even to me your sentences sound “better.” But if Spanish speakers refuse to evaluate constructions that would distinguish between reflexive (de sí misma) and non- (de ella misma), how can various researchers conclude that "le” is not in a subject role in le gusta la foto and that Spanish does not have quirky subjects? After all, that’s the whole point of the strained German sentences.
(Lest anyone is confused: su/suya look deceptively like reflexive 3rd person possessive determiners/adjectives, and their Classical Latin antecedents were just that, of course — but in current Spanish they are not marked for reflexivity or its lack. I’m not quite sure what Latin used for non-reflexive possessive deixis in the 3rd person, I’m guessing at illius but that didn’t survive into Romance AFAIK).
I’ll read the papers ASAP…
Literally this is “clap into the hands”, with in + accusative for directions (all the way back to PIE) as opposed to in + dative for places (“ablative” in Latin, prepositional in Russian etc.).
…but I’ve never heard anyone say it. Usually it’s just klatschen “clap, applaud” without any further material.
Change the word order, and behold what happens:
Es war mir zu kalt, um in die Kirche zu gehen.
A wild dummy subject appears!
This word order is uncommon, but does occur. I think the reason why the subject is omitted in the more common order that puts mir first is that es in mir war es zu kalt would be interpreted as a non-dummy subject, triggering misunderstandings about what “it” is.
Semantically, I have always taken the dative literally, as empathic/incommodi: “for me, to my mind, in my opinion, it was too cold” – definitely not “my own body temperature was too low”.
(…After all, you’re more likely to feel cold when your body temperature is too high, i.e. you’re having a fever.)
Compare:
Der Wind war mir zu kalt, um in die Kirche zu gehen.
Mir war der Wind zu kalt, um in die Kirche zu gehen.
“The wind was too cold for me to go to church.”
This is not a TED talk. I am not the subject of my presentation, and I’m not the syntactic subject of any of the above example sentences.
The dummy subject is the same as in the less idiomatic dative-free version:
Es war zu kalt für mich, um in die Kirche zu gehen.
“It was too cold for me to go to church.”
Oh wow.
Alon, thanks for the examples. My Spanish is very intermediate, and these are exactly the sort of examples that help one focus on how the syntax works and doesn’t work. This is the sort of thing Benjamin and Butt’s grammar of Spanish excels at.
Literally this is “clap into the hands”, with in + accusative for directions (all the way back to PIE) as opposed to in + dative for places (“ablative” in Latin, prepositional in Russian etc.).
…but I’ve never heard anyone say it
Let’s fill that gap 😉
@Alon, I know the examples are contrived, and even to me your sentences sound “better.” But if Spanish speakers refuse to evaluate constructions that would distinguish between reflexive (de sí misma) and non- (de ella misma), how can various researchers conclude that “le” is not in a subject role in le gusta la foto and that Spanish does not have quirky subjects? After all, that’s the whole point of the strained German sentences
Maybe use better examples? We’ve already seen for the German sentence that all of the three native speakers who commented here doubted its grammaticality, and when we are told that the Spanish sentences also wouldn’t be produced by a native speaker unprompted, then this just makes the conclusions rather doubtful.
Hans, have you seen “Dative Sickness” in Germanic?
No, I hadn’t. As I don’t have JSTOR access and they want me to pay $44 for it, I’ll pass 🙂
You can always use sci-hub, like everyone else. 😀
@Hans, I’m trying to poke at what counts as evidence. Both for German and Spanish we have sentences (with von sich and with de sí misma) where the question is if the coreference of a dative with a reflexive pronoun is grammatical. And the natives go “nah, we don’t like that construction in that context so we won’t tell you if it’s grammatical.” I’m just going by the sample sentences offered by WP, assuming that those are as good as it gets, and if those are all ruled out for other reasons, do we have to conclude that the researchers are talking through their hats when they say German does, Spanish doesn’t?
(For both languages, the issue boils down to possessive determiners being preferred to preposition+personal pronoun when the distinction only exists for pronouns. I started out by trying to explain why the samples for German had to look like they did, I don’t have any investment in positive or negative answers).
@Lars Mathiesen: I didn’t refuse to evaluate the construction. I explicitly said it was ungrammatical.
Probably.
I tried to read one of the articles linked from the WP article and the author’s reaction to multiple L1 speakers saying “this construction is unacceptable” was to insist it was grammatical, just “strongly deviant”. From that kind of reasoning anything might follow.
the author’s reaction to multiple L1 speakers saying “this construction is unacceptable” was to insist it was grammatical, just “strongly deviant”.
That’s… not good scholarship.
Oh, and full disclosure about my own native language: You can in fact prepose an object for emphasis even though it’s a bit strained (and not really grammatical unless you fill its “nexus slot” with something like a negator), but the reflexive/non-reflexive distinction is well alive in possessive determiners so clear cut examples are not hard to construct. (At least for singular coreference, but that’s all we need).
Reflexivity distinction (where German and Spanish use sein Bild/su foto in all cases):
Hanᵢ beundrer sitᵢ billede
Hanᵢ beundrer hansᵢᵢ billede
Hanᵢ giver hamᵢᵢ ikke sitᵢ billede igen
Hanᵢ giver hamᵢᵢ ikke hansᵢᵢ⸝ᵢᵢᵢ billede igen
Putting the indirect argument in the subject slot does not affect the coreference of a reflexive possessive determiner:
Hamᵢᵢ giver hanᵢ ikke sitᵢ billede igen
Hamᵢᵢ giver hanᵢ ikke hansᵢᵢ⸝ᵢᵢᵢ billede igen
The closest to German gefallen is probably tiltale whose single object is probably “direct” — with only a nominative/oblique distinction in pronouns, and no noun cases at all, you can’t really tell. But a reflexive form is still not grammatical:
Hansᵢ billede tiltaler hamᵢ⸝ᵢᵢ ikke
Hamᵢ tiltaler hansᵢ⸝ᵢᵢ billede ikke
So it’s provable that Danish doesn’t have quirky subjects. But as the proverb goes, the sum of the sins is a constant.
@Lars: For the German examples, the natives expressed doubt on the grammaticality first, and then only speculated on whether these doubts may have something to do with questions of register.
Sometimes you can separate questions of grammaticality and idiomaticity (the classical “yes, that’s grammatical, but nobody says it like that”), but if in order to establish whether a condition (as here for subjecthood) is valid you have to construct examples that make people feel queasy, and you don’t get a clear “yep, that’s fine, no problem with that”, that means you can’t establish grammaticality with these examples. And if other examples aren’t possible, the grammaticality must remain doubtful. Accordingly, if establishing quirky subjects in German rests only on one of three conditions being fulfilled (as per WP), and the grammaticality of the example establishing the validity of that condition is under doubt, stating that German has quirky subjects is a bold thing to do, maybe even close to talking through one’s hat.
BTW, that’s why I find the Icelandic evidence much better – there is a wide range of subjecthood conditions that are fulfilled, and you also get the feeling that the interplay between noun case, verb valency, and syntactic roles in other parts of the grammar is different from other European languages with case in ways that makes it useful to speak about non-nominative subjects. It doesn’t all rest on a couple of examples in fringe constructions.
@juha: Thanks for the link to the “groovy” paper.
@Alon, is there no middle ground between fully idiomatic and ungrammatical? I admit I focused on you saying the Spanish examples would have to be rephrased to be fully idiomatic, but to me that doesn’t mean you can’t distinguish degrees of badness between de sí misma and de ella misma in that unidiomatic construction. You may have to fiddle a little with the framing and / or put a few extra adverbs in like I did for Danish.
EDIT: OK, Hans answered that. And leaving it at “we can’t really know” for both languages is fine by me.
@Lars: Just to make my position clear, I’d rather go for a “not proven” than a “we don’t know” regarding quirky subjects in German.
Professor Ussery, of the “groovy” piece linked to above, has lots of other work on Icelandic and Faroese that may be of interest to the sort of people interested in that sort of thing. I haven’t clicked through to see if the same dizzying range of fonts is employed in her other work. https://www.carleton.edu/people/cussery/research/icelandic-research/
@Hans, I think my quibbles boil down to wanting to know what the difference between those two are.
@Hans & Lars Mathiesen: Yeah, I was going to ask what the difference was there. An allusion to Scottish law?
Oh.
{Dem Fritz}ᵢ gefällt das Bild von sichᵢ
Grammatical – but not standard, because articles with personal names aren’t standard; omitting it would leave the dative unmarked. You’ve elegantly avoided this problem by replacing the Fritz with the lady.
Ich sehe ihnᵢ {den Mann}ᵢᵢ mögen
Borderline grammatical, because German is much more reluctant to use accusativus cum infinitivo than English and much happier to introduce new clauses. This contraption is more likely to be interpreted as a questionable attempt to say “I watch him liking the man”, whatever exactly that might mean, than “I see him like the man” (doubtful even in English) / “I see that he likes the man”. The German article has Ich sehe ihn das Haus verlassen “I see him leave the house”; that’s completely unobjectionable. I think I’ll simply replace that in the English article.
…and there aren’t any other German examples in the article.
At least one of the Basque ones is claimed to contain a nominative, BTW.
In the section about PRO control, I have no idea why this PRO is postulated for these sentences, let alone why it’s supposed to have the invisible cases it’s supposed to have.
The bottom of the talk page contains this unanswered comment:
Indeed English has both I like it and it pleases me, and German has both ich mag es and es gefällt mir. The only difference is idiomaticity, not one having a quirky subject there and the other not.
Clicking on the link to the German article redirects to Lexikalischer Kasus, which has a one-sentence section on Quirky case, italics in the original, at the bottom; that sentence says Icelandic has it and German lacks it, flatly contradicting the English article.
In that sentence, there’s also a link to the relevant section in the article on grammatical subjects that shows things Icelandic can do and German can’t. I don’t understand all of them, but I find it striking that leiðast “be bored” is both reflexive (-st), like German sich langweilen, and puts the people who are bored in the dative, like regional German fad sein (mir ist fad “I’m bored”). Similarly, the English article has mig vantar peninga, “I need money”, with both “I” and “money” in the accusative; I can’t make German sense of that either.
(The closest I can get is mir fehlt es an Geld “I lack sufficient means”, with a dummy subject. Mir fehlt Geldes seems like it should have been grammatical in the 17th century, but I’m just speculating here, and the money there would be in the genitive, not the accusative or dative.)
However, on the mir ist kalt phenomenon, some of these articles mention mich friert (“I’m cold”, literally “freezes me” with accusative). Its dummy subject often surfaces if the sentence goes on: mich friert es an den Zehen “my toes are cold”, lit. “it freezes me at the toes” – that doesn’t seem to be obligatory, but it’s much more idiomatic than es ist mir kalt.
And now I’ll read two of the papers…
In the section about PRO control, I have no idea why this PRO is postulated for these sentences, let alone why it’s supposed to have the invisible cases it’s supposed to have.
It’s a ghost subject which must exist if you accept the doctrine that all clauses have subjects. Those of us who are agnostic about such questions can simply think of it as a way of talking about what case the subject would be expected to take if there was one.
Ah, so the idea is that all languages underlyingly practice Balkan-style infinitive avoidance. Why not. It’s a nice change from the idea that all languages are underlyingly English. 🙂
It was maintained by Mark Baker at one point that polysynthetic languages (as defined by his own comparatively precise criteria) don’t have infinitives. Unfortunately for the beautiful theory, some of them actually do.
{Dem Fritz}ᵢ gefällt das Bild von sichᵢ
Grammatical
I’m still not sure about the grammaticality of the reflexive. In any case, von ihm would work better for me.
Ich sehe ihnᵢ {den Mann}ᵢᵢ mögen
Borderline grammatical, because German is much more reluctant to use accusativus cum infinitivo than English and much happier to introduce new clauses.
I agree with your assessment and reasoning.
@Hans, I think my quibbles boil down to wanting to know what the difference between those two are.
The claim that German has quirky subjects has not been proven, so my opinion is it doesn’t have them. That’s different to being agnostic about the question.
@Hans: So that’s the opposite of the “Scottish verdict” meaning then. Although it did not originate this way, the “not proven” verdict is mostly used now for cases where the jury finds that the evidence indicates that the defendant is probably guilty, yet still falls short of the legal standard for a conviction.
To me, the most striking thing about Icelandic datives is ditransitive verbs with two datives such as:
(9)
a. Þú græjar þér bara útilegudrasli ef þú átt það ekki.
you procure you.DAT just camping stuff.DAT if you own it not.
b. Þú græjar þér bara útilegudrasl ef þú átt það ekki.
you procure you.DAT just camping stuff.ACC if you own it not.
‘You just get yourself camping stuff if you don‘t have it.’
The test examples for the verbs meila and smessa are given in (10):
(10)
a. Gætirðu meilað mér þessu/þetta sem fyrst?
could.you e-mail me.DAT this.DAT/ACC as first
‘Could you e-mail this to me as soon as possible?’
b. Geturðu ekki bara smessað honum reikningsnúmerinu okkar?
can.you not just SMS him.DAT the.account.number.DAT our
c. Geturðu ekki bara smessað honum reikningsnúmerið okkar?
can.you not just SMS him.DAT the.account.number.ACC our
‘Can‘t you just send him our account number by SMS?’
[…]
The intended goal of verbs of instrument of communication can be expressed not only as a dative DP but also as a PP headed by the preposition til ʻtoʼ (Barðdal 2008: 128–132) but this does not effect the case variation with the direct object:
a. Gætirðu meilað þessu/þetta til mín?
could.you e-mail this.DAT/ACC to me.GEN
‘Could you e-mail this to me?’
b. Geturðu smessað númerinu/númerið til hennar?
can.you SMS the.number.DAT/ACC to her.GEN
‘Can you send her the number by SMS?’
(from Chapter 12 above)
A major argument against the feasibility of reconstructing syntax for proto-stages is the widely
discussed lack of directionality of syntactic change. In a recent typology of changes in argument
structure constructions based on Germanic (Barðdal 2015), several different, yet opposing,
changes are reported. These include, among others, processes sometimes called dative sickness, nominative sickness, and accusative sickness. In order to tease apart the roles of the different processes, we have carried out a phylogenetic trait analysis on a predefined data set of twelve predicates found across the Germanic phyla using the MULTISTATE method. This is, as far as we are aware, the first application of the MULTISTATE method (Pagel et al. 2004) in historical syntax. The results clearly favor one of the models, the dative sickness model, over any other model, as this model is the only one that can accurately account for both the observed diversity of case frames and the independently proposed philological reconstructions.Methods of evolutionary trait analysis can be used to model evolutionary paths of argument structure constructions, and they provide the perfect testing ground for hypotheses arrived at through philological reconstruction, based on classical historical-comparative methods.
Dative sickness:Aphylogenetic analysis of argument structure evolution in Germanic
I agree it would, but von sich should work at least if he’s taken the picture himself.
That looks to me like an ongoing wholesale replacement of the accusative by the dative, as has happened in the Berlin dialect:
“Küsse mir, Kasimir!”
“Mich!”
“Also jut, küsse mir, Kasimich!”
Finally someone does ancestral-state reconstruction in linguistics properly! 🙂 Methinks there are a few minor mistakes in the dataset, though, and the paper keeps calling the method “MULTISTATE” even though in the cited source (Pagel et al. 2004) that string only occurs twice – both times in the name of the software, BayesMultiState. I’m also surprised the concept of partial uncertainty is never mentioned, but computationally it’s the same as polymorphism anyway, so maybe the software doesn’t even let you choose.
I’ve now read the groovy presentation. Far out, man. The word-order restrictions on slides 8 and 12 simply don’t apply in German. And example 13d is mind-blowing.
Eliding datives like on slide 21 isn’t an option in German either, and neither is any of the optional agreement in the whole presentation.
The grammaticality judgments about German on slide 23 are all correct, though Er hofft, wegzugehen. is rather nonsensical – “he hopes to be going away”. Try Er hofft, weggehen zu können. – “he hopes to be able to leave”. In the last example, both options for “he hopes to be helped” are correctly stated to be ungrammatical; how this situation is really handled is with an extra clause – er hofft, dass ihm geholfen (werden) wird or rather er hofft, dass ihm jemand hilft.
Slide 24 is all correct, too.
The examples in the image on slide 15 are mind-blowing about something completely different: while festnehmen “arrest” is a verb with a separable prefix in German, its Icelandic calque is a verb with not even an adverb, but an adjective that gets declined!
…and p. 17 shows that you wish the genitive like in Russian.
Disclaimer: I have not yet read anything linked to in this post or the comments.
I was sure I had discussed Jóhanna Barðdal and Icelandic oblique subjects here before, but it seems that we’ve only touched on it. I also thought I had saved several decade-old papers on the subject, but alas.
Anyway, my understanding (loosely based on memory) is that on one level this is about free word-order and fronting for emphasis or topicalization. When the object of a certain verb is routinely moved to the front, the verb may be said to have changed direction, i.e. have become deponent. If **me feeds he becomes the (or a) regular syntax of the verb ‘feed’, it can be said to mean “be fed by” in that particular construction, In Icelandic, this is quite common — and the verbs are interpreted accordingly. So far so good. However, quite a number of verbs may never have had a non-deponent form or may never have required a (nominative) subject at all. Rather, active meaning of cognate verbs in related languages are (or would seem to be) secondary developments.
In the oblique-subject analysis, the experiencer or patient of a verb, marked by an oblique morphological form, takes on the role as the syntactic subject of the sentence. If that is a good description would depend both on complicated tests and on definitions of the terms — or maybe really be irrelevant. The interesting question (at least to me) is whether this construction (or this use of pronoun morphology) can be reconstructed for earlier stages of the language. What I read a decade ago seemed to want to go that way — maybe all the way back to PIE — as a natural predecessoir of both ergative and accusative languages. That would mean exploring other Germanic and IE languages for vestiges of the construction. I don’t understand why that would require testing the limits of synchronic grammaticality in German and Spanish,
Trond’s comment got cut off, but:
That would mean exploring other Germanic and IE languages for vestiges of the construction. I don’t understand why that would require testing the limits of synchronic grammaticality in German and Spanish
It wouldn’t — the purpose of doing that is for synchronic analysis, not reconstruction.
I don’t think non-nominatives in Ancient Greek or Latin pass any of the subjecthood tests except (occasionally) coreferentiality with reflexives, which FWIW fits in with the “Quirky Subject Hierarchy”, but is probably better taken as grounds for saying that reflexives in those languages don’t necessarily have to refer to subjects.
Tangentially to what Trond says, there were what looks like accusative subjects in Old Danish, so it’s not just West Nordic. Maybe even datives, I think the cases had merged a bit earlier: mik drømde is well known. For a verb like tykke, the original tycker mig changed to tyckes mig around 1500*, and then to jeg tykkes in the 18th. (It’s now obsolete, but syner/synes is still alive and must have more or less the same history. I can still say han synes mig lidt forsigtig — it feels archaizing, but it allows putting a predicate in, but with the modern slot assignment it has to be jeg synes han er lidt forsigtig. [We do have object predicates, they just don’t work with this verb]).
Anyway, I don’t have anything that talks about reflexive coreference with such oblique experiencers in Old Danish and my Sprachgefühl is of course useless — it would be interesting to know when Icelandic developed this feature, or if it was present in ON already.
____________________
(*) cf the funky Ic leiðast mer.
TR: Trond’s comment got cut off
That was me moving the “but” upfront as a disclaimer.
Rereading in broad daylight, I have a couple of corrections:
Rather, active meaning of cognate verbs […]
Rather, active syntax/semantics of cognate verbs […]
(or this use of pronoun morphology)
(or this use of morphological markers)
Lars M.: For a verb like tykke, the original tycker mig changed to tyckes mig around 1500*, and then to jeg tykkes in the 18th. (It’s now obsolete, but syner/synes is still alive and must have more or less the same history.
Me thinks this would be the nominative (or at least non-oblique) subject becoming obligatory. Instead of the verb requiring an experiencer and a predicate, the predicate now (briefly) takes the role of subject, reflexively referenced by the verb. The next step would be a pronoun subject with the old predicate added later for clarity; Det tykkes meg, at han ikke vil blive gift. This is soon bleached to a dummy subject, and the predicate is a predicate again.
But me also thinks the path from Det tykkes mig at […] to Jeg tykker at […] is less a development than a leap. It’s easier to bridge if we suppose the coexistence of Det tykkes mig and Mig tykkes, and only the latter became Jeg tykkes.
@Trond, as far as I could see from the very brief treatment, the construction with a clausal argument went more like Tykker mig at […] (without a dummy subject) -> (~1500) Tykkes mig/Mig tykkes at […] -> (~1730) Jeg tykkes at […] (and now extinct). I don’t have any way to look for constructions with simple nominal or pronominal subjects, so I don’t know if you could say Han tykker mig led vel sim back in the 14th.
These datives [in OE] are all empathic.
Double positive. There is nothing empathic about 9C sticode him mon þa eagan ut ‘then someone gouged him the eyes out’, at least if empathic means what I think it means. It’s a dative external possessor, and it’s part of the SAE Sprachbund now lost in English, even though we still have a (pre)positional dative.
The reason they’re lost in later English is not some morphosyntactic change, but the phonological fact that the dative has merged with the accusative
What is phonological about mec > me, uncit > unc, usic > us, þec > þe, incit > inc, eowit > eow, and the more confusing 3p? There is no such phonological process as “loss of -it/-c.” Note particularly that this change happened within the OE period, while vowels were still full, and the older forms remained current in East Anglian. The dative ate the accusative, that’s all.
…and from that he has a disadvantage. Something is done to him, and that’s expressed by the dativus incommodi. Corresponds 1 : 1 to stach man ihm die Augen aus except for word order.
I was thinking of the nouns, but apparently that’s wrong:
Offenbar sind wir alle Berliner.
David M.: That looks to me like an ongoing wholesale replacement of the accusative by the dative, as has happened in the Berlin dialect:
At least some of those examples do seem to have datives where I’d expect an accusatve.
Also note how the genitive-cum-partitive is used when the referent unspecified for quality or quantity, as in negation.
[Disclaimer: I’m always wrong about Icelandic.]
The paper on ancestral-state reconstruction does find “dative sickness” to fit the data best (nominatives stay nominatives, datives become nominatives, accusatives become nominatives or datives).
That’s shared with Slavic, too.
Obliquely related to the overall topic of noncanonical actants:
В печати (2010). [Ю.А. Ландер, А.П. Выдрин] Неканоническое маркирование субъекта ситуации в языках Северного Кавказа: модальные конструкции
In this paper, we explore the phenomenon of non-canonical subject marking in modal constructions of languages of the North Caucasus which belong to the three families: Northwest Caucasian (Abkhaz-Adyghe), Northeast Caucasian (Nakh-Daghestanian) and Indo-European (Ossetic).
The term ‘subject’ is mainly understood here semantically, typically it refers to the most agentive and the most topical participant. Somewhat simplifying the actual situation, it can be said that with monovalent (one-place) predicates, subject canonically should have nominative coding in accusative languages and absolutive coding in ergative languages, while with polyvalent predicates, subject is normally encoded as the nominative argument in accusative languages and as the ergative argument in ergative languages. All other kinds of marking of subject (including, for example, those observed in passive constructions in accusative systems, constructions with multivalent intransitive predicates in ergative systems, etc.) are considered non-canonical.
Three types of non-canonical marking can be distinguished, namely:
— lexically-motivated non-canonical marking, which is tied to particular verbal lexemes,
Ossetic
dæwæn æmbæl-ǝ а-sæw-ǝn
you(SG).DAT should.PRS-PRS.3SG PRV-go.PRS-INF
‘You should go away.’ (example 13)
Chechen
suna v-axa meg-a
I.DAT I-go.INF can-PRS
‘I can (am permitted to) go.’ (example 17)
— derivational non-canonical marking, which is motivated by the use of particular grammatical forms or complex analytical constructions,
Chechen
süöga d-a-lwo
I.ALL III-do-POT
‘I am able to do (this).’ (example 35)
Kryz (Authier 2009: 309)
va-var za šimbi ula-z ša-ba-da-b
you-AD.EL I.GEN brothers eat-INF be-H.PL-NEG.FUTP-H.PL
‘You will not be able to eat my brothers!’ (example 32)
— non-derivational non-canonical marking, which is not grammatically motivated by any other elements of the clause.
Adyghe
[Ø-fe]-tə-š’t-ep
3SG.IO-BEN-give-FUT-NEG
‘He will not be able to give (this).’ (example 44)
(Cp. with ə-tə-š’t-ep [3SG.A-give-FUT-NEG] ‘He will not give (this).’)
Hinuq (Diana Forker, field notes)
hadze-qo hago uži aši-yo gom
they.O-POSS.ESS this boy find-PRS be.NEG
‘They cannot find this boy.’ (example 49)
As is seen, all of these types are found in modal constructions of the languages of the North Caucasus.
Lexically-motivated non-canonical marking is actively used in Ossetic, but in other languages of the North Caucasus it virtually plays a less important role (thus, for example, many Lezgic languages lack a genuine verb denoting possibility and hence cannot display this kind of noncanonical marking in the relevant constructions). Moreover, in many Northeast Caucasian languages lexically-motivated non-canonical marking instantiates a general tendency of noncanonical marking of less agentive subjects usually associated with experiencers, although in some languages (e.g., in Tsakhur, Lak, Hinuq), modal subjects show a distinct pattern of marking.
Derivational non-canonical marking is quite widespread in the Caucasus. It is especially prominent in Northeast Caucasian languages, where it can be yielded by both synthetic modal constructions (primarily with the morphological potentialis category) and non-compositional analytic modal constructions (employing verbs like ‘be’, ‘become’ etc.). Besides ergative North-east Caucasian languages, derivational non-canonical marking is found in Ossetic where it is used in the cases of expressions of external possibility and obligation.
Non-derivational non-canonical marking, which involves the change of the syntactic role of the agent without any other marking of the modal semantics, is observed in the Northwest Caucasian family (in particular, in Adyghe and Kabardian), but also marginally in some Tsezic languages belonging to the Northeast Caucasian family. But in Abkhaz and Abaza (Northwest Caucasian) only traces of non-derivational non-canonical marking are found: in fact, these languages demonstrate how derivational non-canonical marking may evolve from non-derivational one.
Interestingly, in several languages throughout the Caucasus – e.g., in Chechen, Tsez, Hinuq and Khinalug (Northeast Caucasian) and in Adyghe and Kabardian (Northwest Caucasian) – we observe effects of ergativity in the domain of non-canonical subject marking: in particular, such marking appears with transitive subjects but not with intransitive ones. Basing on this fact as well as on general facts concerning the distribution of various kinds of non-canonical subject marking, we hypothesize that ergativity may correlate with the range of types of the phenomena that is found in a given language. Thus we propose the scale LEXICALLY-MOTIVATED MARKING vs. DERIVATIONAL MARKING vs. NON-DERIVATIONAL MARKING and suggest that accusative languages prefer non-canonical marking which is closer to the left end of the scale, while ergative languages prefer non-canonical marking which is closer to the right end of the scale.
Interesting stuff, thanks!