Case in coordinated conjuncts

Several works have focused on case in coordination, such as Munn (1994), Johannessen (1994, 1997, 1998), Quinn (2005), and Parrot (2009, 2010). The literature shows that it is not the case that all conjuncts have the same grammatical features, not even with respect to case. Coordination is not symmetrical; it is very often true that conjuncts have different case (this is called unbalanced coordination in Johannessen 1994, 1997, 1998). We also find coordination in which both conjuncts have the same case, but that this case is unexpected given the position of the coordinated phrase in the clause (called extraordinary balanced coordination by Johannessen, op.cit.). The ScanDiaSyn survey has tested four sentences in Swedish (Sweden and Finland). These four sentences are given below.


Introduction
Several works have focused on case in coordination, such as Munn (1994), Johannessen (1994Johannessen ( , 1997Johannessen ( , 1998)), Quinn (2005), and Parrot (2009, 2010).The literature shows that it is not the case that all conjuncts have the same grammatical features, not even with respect to case.Coordination is not symmetrical; it is very often true that conjuncts have different case (this is called unbalanced coordination in Johannessen 1994Johannessen , 1997Johannessen , 1998)).We also find coordination in which both conjuncts have the same case, but that this case is unexpected given the position of the coordinated phrase in the clause (called extraordinary balanced coordination by Johannessen, op.cit.).The ScanDiaSyn survey has tested four sentences in Swedish (Sweden and Finland).These four sentences are given below.

Nordic Syntax Database (NSD)
Below we present the four maps that represent the judgments for sentences (1a-d).
Map 1: Accusative case in the second conjunct of a coordinated subject.Map 4 shows that there is variation in the extent to which Swedish informants accept nominative case in the second conjunct of a coordinated prepositional complement.Nominative in this position is accepted in Northern Sweden and across the Baltic Sea in Finland as well as in the island of Gotland in the south east of Sweden and across the sea to the Southern parts of Finland.

Nordic Dialect Corpus (NDC)
Given that none of the other Nordic languages, apart from Swedish, has been tested for these questions on coordination, we need to find data for the other Nordic languages.Here we will look at the transcriptions and recordings of spontaneous speech, as it appears in Nordic Dialect Corpus (Johannessen et al. 2009).
Starting with Norwegian, a simple string search for the conjunction og 'and' followed by meg The same search with respect to Danish gives 18 relevant hits (subjects and subject predicates).Since the Danish part of the corpus is only a tenth of the size of the Norwegian part (211 000 words and 2 190 000 words, respectively), it is obvious that the 18 hits in the Danish part as against the 9 hits in Norwegian, suggest that this phenomenon is much more common in Danish.The hits are from all across Denmark, and there seems to be no dialectal variation.
( The Danish results are corroborated by Parrot (2010); he found nearly 18 % pronoun mismatches amongst coordinated NPs.There are, in comparison, no hits for a possible Swedish sequence och mig ('and me') in the NDC.Clearly the maps in section 2.1 is a true reflection of Swedish as opposed to Norwegian and Danish.The latter two can have non-nominative pronouns in subject position as long as the pronoun is the second (or third or fourth etc.) conjunct in a conjunction phrase.To conclude, both Norwegian and Danish have sentences like #1444 and #1446, but Swedish does not.
Looking for sentences like #1447 in the NDC is easy.The search string for Norwegian and Danish is simply preposition + any word + og 'and' + jeg 'I', and for Swedish the equivalent with och jag 'and I', which would have given examples of a nominative pronoun conjunct in a conjunction phrase following a preposition.There were only irrelevant examples of VP or CP coordination.
Finally, for the sake of completeness, we have tested whether the pattern in #1445 is found in the corpus.
Not surprisingly there are many examples for Swedish, given map 2.An example is this: (4) a. för mor och jag vi ansåg at föräldrar skulle vara kvar (Swe.) for mother and I.NOM we.NOM thought that parents should be behind 'for mother and I, we thought that parents should stay...' (boda_om1) In Norwegian, there were also countless examples.However, Danish had only a handful of hits.I present one example of each of these: ( It is clear that overall, Danish is a real mismatch language, while Swedish is not, and Norwegian is somewhere in between.It is clear that Danish and Norwegian share the possibility of having an accusative second conjunct pronoun in subject position.However, it is also clear that Danish has relatively many more hits in the corpus.Likewise, although Danish has a handful of examples of nominative second conjuncts in subject position, these are very few.There are many more of these in Norwegian.Swedish only has nominative and no accusative pronouns in subject position, and thereby confirms the Swedish maps.However, no corpus example was found in any of the languages of a nominative pronoun conjunct in the complement of the preposition.Here the prediction of Map 4 was not borne out.We have not investigated Faroese, but Parrot (2010) shows that this kind of case mismatch is completely unacceptable for his 40 informants.

Discussion
The NSD shows that in Swedish, case in conjunction phrases follows the same rules as case in simplex phrases; nominative is the subject case, non-nominative (accusative) is the case for objects, though there is some variation after prepositions.However, in Norwegian and Danish, the NDC shows many examples of unexpected case (called unbalanced coordination in Johannessen 1994Johannessen , 1996Johannessen , 1998)).Here the second conjunct in a conjunction phrase in subject position can have accusative case.In Danish there are actually very few examples of the expected nominative case in this position, showing that the unbalanced coordination may be more the rule than the exception in Danish.Swedish, according to the NSD, reveals some acceptance of nominative case in the second conjunct of coordinated complements of prepositions, but there is no corpus example of this phenomenon in any of the languages.Icelandic does not show anything unexpected in the corpus.The Faroese part of the corpus has not been investigated, but Parrot (2009) shows that Faroese behaves like Swedish.

(# 1444 :
Min man och mig ska prata om...'My husband and me are going to talk about…') (Black=low score) Map 1 shows that no informants accept a coordinated subject where the second conjunct has Nominative case in the second conjunct of a coordinated subject.(#1445: Min fru och jag brukar åka.. 'My wife and I tend to go...') (White=high score) Map 2 shows that all informants accept a coordinated subject in which the second conjunct has nominative case.Map 3: Accusative case in the second conjunct of a coordinated subject where both conjuncts are pronouns (#1446: Den dagen satt hon och mig... 'That day, she and me sat…').(Black=low score) Map 3 shows that no informant accepts accusative case in the second conjunct of a coordinated subject where both conjuncts are pronouns.Nominative case in the second conjunct of a coordinated prepositional complement.(#1447: Det är två år mellan min bror och jag.'There are two years between my brother and I') (White = high score, black = low score) me', gives several interesting results.This search would give 9 results equivalent to sentences #1444 and #1446 (i.e., subjects and subject predicates) if indeed they can be found.There are not many hits, but some are presented in (2).They show that Norwegian has subject pronouns that do not have nominative case.The results are all from Southern Norway, but here they span across the whole width, from Hordaland and Rogaland in the west via Vest-Agder to Akershus and Hedmark in the east.The fact that there are no examples from Trøndelag or Northern Norway may indicate a north-south divide, but there and Anne and Martin and me.ACC we.NOM have well known each other in 'and Anne and Martin and me, we have known each other for...' (nordjylland4)