Icelandic: verb of motion + dative object

Discussion in 'Nordic Languages' started by Gavril, Mar 18, 2013.

  1. Gavril Senior Member

    English, USA
    Sæl,

    I've noticed that, for many Icelandic verbs in which the subject moves (displaces) an object, that object appears in the dative case:

    lyfta "lift": Hann <lyfti> steini "He lifted a stone"
    sópa "sweep": Hún sópaði saman brotnu glerinu "She swept up the broken glass"
    kasta "throw": Hann kastaði steini í loftið "He threw a stone in the air"
    spyrna "kick" : Styrmir spyrnti knettinum"Styrmir kicked the ball"

    rugla isn’t (necessarily) a verb of motion, but it takes the dative when the meaning is “to mix (things/ideas) up”: Ég ruglaði saman hinum tveim orðum "I mixed the two words up".

    I have a few questions about this pattern of dative objects:

    1) Does the pattern primarily apply to verbs of transitive motion (= displacement), or is it broader/more complex than this?

    2) What is thought to be the historical reason for this development from accusative to dative? E.g., did these dative forms originate as instrumental datives in double-object constructions?

    I'm assuming that the development was from accusative to dative because, based on the little searching I did, the cognate verbs in other Germanic languages that have an accusative/dative distinction don't seem to put the objects of these verbs in the dative: e.g., it seems as though Old English swápan "sweep" and weorpan "throw" took accusative objects, in contrast to Modern Icelandic sópa and verpa.

    Thanks for any info,
    Gavril
     
    Last edited by a moderator: Mar 19, 2013
  2. Alxmrphi Senior Member

    Reykjavík, Ísland
    UK English
    Hello Gavril,

    1) It is a little bit more broader than that, yes. You can see a pattern of alternation between accusative and dative in certain verbs (like skjóta) which pattern differently/semantically depending on the case of the object, but not necessarily in the sense that it's case that is 'phonemic' in that nuance of singularly causing an alternation - different nouns would also be used. Some Icelandic linguists have equated this (quite often the skjóta example) to be that a dative shows the initiation of an action while accusative has an effect on the object of the verb (i.e. not instrumental dative) that lasts until some sort of different state has been accomplished. So, to say skjóta + animal (ACC) is how you illustrate the OBJECT has been affected, but if you wanted to talk about what you fired, then the instrument is in the dative. In that sense it's not really an 'alternation' but it's often described as such. If you wanted to go into the weird semantics of shooting animals out of a gun, you could argue for a true alternation but the nonsensicalness of that would just cause a native to find it weird and reject it anyway. Dative usually indicates movement in this kind of example of forced motion (typically only when PPs accompany it). If things are in the same location, they are accusative (directly opposing the alternative of subject-orientated alternation where dative means being stationary and accusative, movement). On that same line of thinking you also get the ballistic versus accompanied motion distinction related to case in Icelandic where dative indicates separated caused motion but the accusative indicates accompaniment. So, if I throw something up into another area (typical example is usually a wagon) then the object is in the dative, but if I walked up with it and took it there and accompanied it, that then requires the accusative case.

    2) As far as I know, yes that is the case.

    Oh right, I didn't read the last paragraph until just now. I get why you said that now. I think when you bring Old English into the mix, taking any sort of 'this must have been the original' is not advisable. If you take a look at the cognate verb in German (werfen), well, that's a minefield when searching for a linguistic explanation as to what case it selects as it varies between transitive and ditransitive and in itself is weird compared to semantically similar verbs. It could very well be in Icelandic this object movement derived from the semantics of ditransitive constructions that became productive. I am not aware of if accusative was ever the original case in Proto-Germanic though. It itself is built on words mainly and retracing syntax with such grammatical behaviours in the daughter languages leads me to hypothesise we're not looking more than an educated guess that has a huge chance of being wrong. Archaic dialects of Swedish and Norwegian are said to pattern with Icelandic a lot, but also in ways that English and German behave - read into that what you will.

    In any case, seeing the complications of the semantics even in Icelandic today, regarding case, and how much they are debated in the literature, to not even have that down and for some people arguing for classes of frames that these verbs fall into (some say four classes - others go up to seventeen of them), this kind of shows the minefield of mapping this out for a modern language. God only knows how people would trace this in other languages. It's not sufficient to find a few examples of what case a verb selects in another language without matching the context, the frame, and that is where I believe the problem lies in anyone getting anywhere near a good hypothesis as to what the wholly reconstructed language of PG would have done, and due to that fact we can't tell what was the original and which daughter language reflects the oldest syntactic behaviour.
     
    Last edited: Mar 18, 2013
  3. Gavril Senior Member

    English, USA
    Thanks (eins og alltaf) for the response.

    Do you mean that transitive motion verbs tend to take dative object + a prepositional phrase, because the PP serves to clarify the direction of movement? (E.g., in the sentence, Ég ýtti steininum fram af bjarginu, the PP (fram) af bjarginu shows that I wasn't pushing the stone away from the cliff, or downward, or in another direction.)

    On the other hand, a verb like lyfta doesn't seem to require a PP because the direction is automatically understood. I found the example, Hann lyfti annarri hönd sinni og gaf þannig áhorfendum til kynna að í lagi væri með hann. "He lifted one of his hands [no PP], indicating to the spectators that he was alright."

    I have a sense that verbs meaning "throw" are somewhat variable across languages, in terms of the case forms assigned to the different semantic roles. So verpa/weorpan/etc. probably wasn't the best example I could have chosen.

    I think I understand your point about tracing syntactic development: it's harder to label a syntactic form as "archaic" than it is to label a phonetic form this way.

    Still, if

    1) you have a large corpus of related languages, and

    2) only one of these languages shows a given syntactic pattern, and

    3) this pattern seems to have a fairly regular/productive correlation with semantic properties (in this case, dative -> verb of transitive motion)

    it may be reasonable to guess that this syntactic pattern is an innovation. I think you're saying (below and elsewhere) that #3 is not necessarily true in this case, but the relationship between dative and transitive (separating) motion does seem fairly widespread based on the searching I've done so far.

    (This may be more of an Etymology and History of Languages discussion at this point -- should I request that the thread be moved?)
     
    Last edited: Mar 19, 2013
  4. Alxmrphi Senior Member

    Reykjavík, Ísland
    UK English
    Regarding the aspect of using the dative with PPs - yes I think you got what I meant. When you have the PP it allows you to expressly state something like the location or a way of saying [TO WHERE] the object is going and this is when you use the dative. In some sort of patterns you can't put an object in the dative without this accompanying PP. To take an example quoted in a paper by Peter Svenonius, he uses a pair of example in the spray/load semantic category to show this relationship:

    1) Hann spreyjar málningu (DAT)
    2) * Við hlóðum heyinu.
    3) Við hlóðum heyinu [á vaginn].

    So the 'movement verb' takes an object in the dative to express movement, but only when a direction is supplied. This restriction isn't on movement verbs that don't link to a sort of 'completed event' (so for example, you can shoot a bullet and in that semantic frame it's only about the movement of the bullet and no link to an affected object so you don't need a PP here, but in verbs like 'load' then you would). In that case, when you push a stone as in your example, you can push a stone and it enters into a relationship where the PP does clarify the direction of movement, but when you don't use it there is a different interpretation, i.e. it's just an effect upon an object that is being pushed rather than motion in an event that has an endpoint.

    I'm gonna have a little refresh on some of the studies to make sure I haven't missed anything important and to see if there is anything else interesting that can be added.
    I don't think it's an EHL thread really. You're certainly correct in Icelandic that dative is very typical when there are two parts dedicated to motion and it's often the semantic frame and not specifically the verb that licences the case. Note this example (also taken from Svenonius):

    4) Hann sló köttinn.
    5) Hann sló kettinum í veginn.

    Affected object only = Accusative.
    Two-part frame with motion of object expressed in the dative.

    I think it's easier to say 'a given syntactic pattern' but when it comes down to defining it and specifically looking at cases it wouldn't be linked with and others that it would, that's when the complications will come out. That sort of 'on paper/in theory' list is completely understandable but it's exactly one of those cases that linguists have started off with for about 50 years and the data just doesn't want to behave and there are always people providing more data to say "Well, what about this case?" and then it just turns into some perpetual self over-complicating mess. I'm really nowhere near competent enough to talk about whether it truly is an innovation or not - that'd just be a big guess. What can be said with confidence is this dual sort of 'tool' usage is dative because dative took over the instrumental case and that did come from Proto-Germanic and it is quite possible this did become a productive generalisation to take over verbs of motion, but I don't think this is *it* - I think lexical case assignment also has a part to play in some cases where we can't draw a blanket generalisation over the data which are of a different sort / where the case either comes from a variable alternation (i.e. the cat example) or something like að kasta which would have the specification that all objects are followed in the dative. In that paper, Svenonius says "Whatever the historical source of the construction, it is clear that modern Icelandic uses dative on objects which undergo (certain kinds of) motion." This is extremely typical - people just don't even want to go there when it comes to suggesting this stuff, point to the possibilities but stick to the modern data.

    It doesn't mean you're wrong in this assumption, however.

    Late edit
    :
    Just as an addendum, I thought you'd want to know that sópa can take complements in both cases: sópa gólfið / sópa snjónum burt. When you have two sub-events that merge (basically it's what they say when you can't identify a moment when you have done one sub-event [put brush to object] and when the second one starts [move object with instrument]. So, because 'brush the floor' is one where it all happens in a unified manner, whereas 'brush the snow away' contains two sub-events of brushing the snow, and then moving it away, you have this "in-between object of actual displacement" which as you know, takes the dative. This sort of backs up what I said about taking a few examples of verbs and the cases they select, because you noted sópa takes the dative but someone else could have searched and seen a group of accusative objects. Anyway, that was just to reference what you said before to point out this pattern there as well. :)
     
    Last edited: Mar 19, 2013

Share This Page