“Jag vill vara dig”: Nominative and oblique case in Modern Swedish
CASE marking yields contrasts like "JAG valde Maria" / "Maria valde MIG", where JAG and MIG are the nominative and the oblique forms. Many languages have numerous cases, many have no cases, and some are case poor, like Swedish and most of the other Germanic languages. However, case marking in Swedish differs from case marking in other case poor languages in a number of ways, Swedish, e.g., generally applying the nominative as a predicative case ("Det är JAG", cf. Danish "Det er MEG"). Nevertheless, the Swedish case system is in flux, partly heading towards a more English/Danish-like system and partly developing along lines special among the Germanic languages.
Some of the Swedish case variation is well known, commonly raising debates about what is correct language ("Hon är större än JAG/MIG", etc.). However, most of the variation has actually gone unnoticed and scientific research on it is sparse. The overarching research question of this project is: Where is the Swedish case system heading and what are the probable effects of the ongoing changes? The goal is thus to accumulate knowledge about the Modern Swedish case variation and to develop a coherent understanding of its nature, limits and probable effects. For this purpose the project will elicit native speaker data and also excerpt written language corpora. - An important issue the project might throw some light on is the question of whether language development bears on trends in complex systems in general.
Some of the Swedish case variation is well known, commonly raising debates about what is correct language ("Hon är större än JAG/MIG", etc.). However, most of the variation has actually gone unnoticed and scientific research on it is sparse. The overarching research question of this project is: Where is the Swedish case system heading and what are the probable effects of the ongoing changes? The goal is thus to accumulate knowledge about the Modern Swedish case variation and to develop a coherent understanding of its nature, limits and probable effects. For this purpose the project will elicit native speaker data and also excerpt written language corpora. - An important issue the project might throw some light on is the question of whether language development bears on trends in complex systems in general.
Final report
P15-0389:1 Jag vill vara dig.
Final report
1 The goals of the project and its development
Languages commonly CASE mark pronouns and nouns depending on their function, see (1) and (2).
(1) JAG valde Maria.
(2) Maria valde MIG.
JAG and MIG are the nominative vs. the oblique forms of the first person singular pronoun (hereafter NOM vs. OBL). Within Germanic, one can distinguish between rich and poor case languages. Swedish, Norwegian, Danish, English, Frisian, Dutch, and Afrikaans are case poor. Despite the overall similarities of the case systems in the Germanic case-poor languages there are some striking dissimilarities between them, Swedish differing from the other.
The goal of the project was to map the distribution of NOM and OBL forms in Swedish, across a number of variables: Age, regions, and education, with focus on the most central constructions that show case variation:
A Predicates in finite clauses (Det är bara JAG/MIG)
B Predicates in infinitives (Det är inte lätt att vara JAG/MIG)
C Comparatives (Hon är större än JAG/MIG)
D Objects of V and P (Jag har inte sett HON/HENNE idag)
E Clefts, etcetera (Det var HAN/HONOM som hon skulle välja)
F Vocatives (Dumma DU/DIG)
G Coordinated subjects (Anna och JAG/MIG)
This has all been accomplished.
Case is limited to pronouns in Swedish – it is a pronominal feature. The project was thus gradually linked to a more general project on pronouns and pronominal features, including studies on Gender, null pronouns, and the nature of plural pronouns. The publication list includes all the writings and presentations of the project leader in 2016-2020.
2 Proceeding
The project leader and principal investigator was Halldor Sigurdsson, Senior Professor, SOL, Lunds Universitet. A co-worker, taking care of statistics, was Joost van de Weijer, researcher, SOL, Lunds Universitet (funded by SOL).
Four investigations were made. First, a preparatory survey among students at Lund University (45 informants, 42 sentences). Second a preparatory survey in written language corpora (Korp, Google). Third, a major online survey across the whole Swedish language area in Sweden and Finland (5,315 informants, 85 sentences, altogether over 450,000 answers), called The Swedish Case Survey, SwCS. Fourth, a supplementary online survey, also across Sweden and Finland (417 informants, 69 sentences). As far as known, comparable surveys over the whole native language area have never been made for any of the other case-poor Germanic languages. The overall results of SwCS are published on the project’s website:
https://jagvillvaradig.files.wordpress.com/2018/03/the-overall-results-tables1701301.pdf
In addition, the results have been presented at nine workshops and conferences (see the publication and presentation list), and a workshop on “Case-impoverished Germanic” was held in October 2017 (with 12 talks on all the case-impoverished Germanic languages, funded by The Birgit Rausing Language Programme). Six of the papers given will be published in 2020 in Acta Linguistica Hafniensia (editors Halldor Sigurdsson and Lars Heltoft).
3 Major results – conclusions
Due to the huge amount of replies to the major survey (SwCS), it gives a fairly reliable picture of the present “Swedish case situation”. As it turned out, there is much more variation in case marking in Swedish than has been previously known, but the variation is systematic. There is limited correlation with education (which can possibly be seen as a success of the Swedish educational system). Regional variation is also limited in general (which might be related to the limited educational variation), but there are some interesting exceptions. NOM direct objects are most widespread in Norrland (as expected), and, secondly, in Österbotten (new knowledge). In general, however, Finland-Swedish is markedly more conservative than Sweden-Swedish (also new knowledge). In contrast to the regional and educational variables, the age variable reflects huge variation, much more so than previously known. Three major changes are markedly on the increase:
A NOM > OBL in certain predicates. This is a very rapid change (for example, 52% of the youngest group (24 years or younger) accepted “Jag låtsas inte vara DIG”, only 20% of the oldest group (65 years or older).This has not been previously noticed, so this is the most striking new knowledge provided by the project.
B NOM > OBL in comparative phrases. This change has been widely noticed, but it has not gone as far in the spoken language as has commonly been assumed (for instance in SAG); it is still very much taking place, with younger speakers accepting the OBL form much more frequently than older speakers (for example, 76% of the youngest speakers accepted Hon är större än MIG, only 31% of the oldest speakers; the respective figures for Hon är lika stor som DIG are 51% vs. 13%).
C OBL > NOM in direct third person objects, in particular complex ones. Strikingly, this change goes in the opposite direction, but it is yet another rapid change (for example, 96% of the youngest speakers accepted Känner du HON som ska träna oss?, only 35% of the oldest group).
At first sight, NOM>OBL in certain predicates might seem to be a radical system change. However, it is evidently not. Swedish is not on its way to “Det är bara MIG”, corresponding to Danish “Det er kun MEG” or Engilsh “It is only ME” (only 0,8% accepted “Det är bara MIG”, with practically no age correlation). The change is largely confined to predicates where the predicative pronoun expresses role semantics (and not identity semantics), so, these predicates are by and large parallel to som-phrases. Jag försöker inte vara DIG ˜ Jag försöker inte vara som DIG. The project shows that there is a strong statistical correlation between OBL vs. NOM in in comparative phrases and predicates. Speakers who accept ”Hon är lika stor som DIG” are highly likely to also accept ”Jag låtsas inte vara DIG” (and vice versa; those who prefer ”Hon är lika stor som DU” are very likely to also prefer ”Jag låtsas inte vara DU”).
The opposite change in complex objects, OBL > NOM, seems to also be a “logical” change within the system (and not really a system change). The central conclusion is that the basic system remains largely intact – there are interesting ongoing changes within the system but not of the system itself.
4 New research questions
There is clear evidence in the data that individual speakers are variably consistent: Some reject the new varieties quite regularly, while others commonly accept both varieties, and yet others (especially the youngest informants) show clear preference for the new varieties. It would be very interesting to map the individual behaviors more accurately, by separately studying the behavior of a selected number of individual speakers. Another interesting question is how the ongoing development in Swedish compares to the development in the other case-poor Germanic languages, in the present and in the past.
5 The distribution and presentation of the results, and cooperation
The overall results have been published on the project’s homepage and presented at numerous workshops and conferences. A paper with the title “Predicative case in Swedish: Default or not?” has been published on Lingbuzz, and two other works are in the making: “Case drift in Swedish”, and an edited special issue on “Case-impoverished Germanic” that will be published 2020 in Acta Linguistica Hafniensia (where the final version of “Predicative case in Swedish: Default or not?” will also appear).
Cooperation: By taking care of statistics, Joost van de Weijer made an invaluable contribution. Lars-Olof Delsing, SOL, and Jeffrey K. Parrott Palacký University Olomouc, helped organize the workshop “Case-impoverished Germanic). Lars Heltoft is the coeditor of the Acta volume.
Final report
1 The goals of the project and its development
Languages commonly CASE mark pronouns and nouns depending on their function, see (1) and (2).
(1) JAG valde Maria.
(2) Maria valde MIG.
JAG and MIG are the nominative vs. the oblique forms of the first person singular pronoun (hereafter NOM vs. OBL). Within Germanic, one can distinguish between rich and poor case languages. Swedish, Norwegian, Danish, English, Frisian, Dutch, and Afrikaans are case poor. Despite the overall similarities of the case systems in the Germanic case-poor languages there are some striking dissimilarities between them, Swedish differing from the other.
The goal of the project was to map the distribution of NOM and OBL forms in Swedish, across a number of variables: Age, regions, and education, with focus on the most central constructions that show case variation:
A Predicates in finite clauses (Det är bara JAG/MIG)
B Predicates in infinitives (Det är inte lätt att vara JAG/MIG)
C Comparatives (Hon är större än JAG/MIG)
D Objects of V and P (Jag har inte sett HON/HENNE idag)
E Clefts, etcetera (Det var HAN/HONOM som hon skulle välja)
F Vocatives (Dumma DU/DIG)
G Coordinated subjects (Anna och JAG/MIG)
This has all been accomplished.
Case is limited to pronouns in Swedish – it is a pronominal feature. The project was thus gradually linked to a more general project on pronouns and pronominal features, including studies on Gender, null pronouns, and the nature of plural pronouns. The publication list includes all the writings and presentations of the project leader in 2016-2020.
2 Proceeding
The project leader and principal investigator was Halldor Sigurdsson, Senior Professor, SOL, Lunds Universitet. A co-worker, taking care of statistics, was Joost van de Weijer, researcher, SOL, Lunds Universitet (funded by SOL).
Four investigations were made. First, a preparatory survey among students at Lund University (45 informants, 42 sentences). Second a preparatory survey in written language corpora (Korp, Google). Third, a major online survey across the whole Swedish language area in Sweden and Finland (5,315 informants, 85 sentences, altogether over 450,000 answers), called The Swedish Case Survey, SwCS. Fourth, a supplementary online survey, also across Sweden and Finland (417 informants, 69 sentences). As far as known, comparable surveys over the whole native language area have never been made for any of the other case-poor Germanic languages. The overall results of SwCS are published on the project’s website:
https://jagvillvaradig.files.wordpress.com/2018/03/the-overall-results-tables1701301.pdf
In addition, the results have been presented at nine workshops and conferences (see the publication and presentation list), and a workshop on “Case-impoverished Germanic” was held in October 2017 (with 12 talks on all the case-impoverished Germanic languages, funded by The Birgit Rausing Language Programme). Six of the papers given will be published in 2020 in Acta Linguistica Hafniensia (editors Halldor Sigurdsson and Lars Heltoft).
3 Major results – conclusions
Due to the huge amount of replies to the major survey (SwCS), it gives a fairly reliable picture of the present “Swedish case situation”. As it turned out, there is much more variation in case marking in Swedish than has been previously known, but the variation is systematic. There is limited correlation with education (which can possibly be seen as a success of the Swedish educational system). Regional variation is also limited in general (which might be related to the limited educational variation), but there are some interesting exceptions. NOM direct objects are most widespread in Norrland (as expected), and, secondly, in Österbotten (new knowledge). In general, however, Finland-Swedish is markedly more conservative than Sweden-Swedish (also new knowledge). In contrast to the regional and educational variables, the age variable reflects huge variation, much more so than previously known. Three major changes are markedly on the increase:
A NOM > OBL in certain predicates. This is a very rapid change (for example, 52% of the youngest group (24 years or younger) accepted “Jag låtsas inte vara DIG”, only 20% of the oldest group (65 years or older).This has not been previously noticed, so this is the most striking new knowledge provided by the project.
B NOM > OBL in comparative phrases. This change has been widely noticed, but it has not gone as far in the spoken language as has commonly been assumed (for instance in SAG); it is still very much taking place, with younger speakers accepting the OBL form much more frequently than older speakers (for example, 76% of the youngest speakers accepted Hon är större än MIG, only 31% of the oldest speakers; the respective figures for Hon är lika stor som DIG are 51% vs. 13%).
C OBL > NOM in direct third person objects, in particular complex ones. Strikingly, this change goes in the opposite direction, but it is yet another rapid change (for example, 96% of the youngest speakers accepted Känner du HON som ska träna oss?, only 35% of the oldest group).
At first sight, NOM>OBL in certain predicates might seem to be a radical system change. However, it is evidently not. Swedish is not on its way to “Det är bara MIG”, corresponding to Danish “Det er kun MEG” or Engilsh “It is only ME” (only 0,8% accepted “Det är bara MIG”, with practically no age correlation). The change is largely confined to predicates where the predicative pronoun expresses role semantics (and not identity semantics), so, these predicates are by and large parallel to som-phrases. Jag försöker inte vara DIG ˜ Jag försöker inte vara som DIG. The project shows that there is a strong statistical correlation between OBL vs. NOM in in comparative phrases and predicates. Speakers who accept ”Hon är lika stor som DIG” are highly likely to also accept ”Jag låtsas inte vara DIG” (and vice versa; those who prefer ”Hon är lika stor som DU” are very likely to also prefer ”Jag låtsas inte vara DU”).
The opposite change in complex objects, OBL > NOM, seems to also be a “logical” change within the system (and not really a system change). The central conclusion is that the basic system remains largely intact – there are interesting ongoing changes within the system but not of the system itself.
4 New research questions
There is clear evidence in the data that individual speakers are variably consistent: Some reject the new varieties quite regularly, while others commonly accept both varieties, and yet others (especially the youngest informants) show clear preference for the new varieties. It would be very interesting to map the individual behaviors more accurately, by separately studying the behavior of a selected number of individual speakers. Another interesting question is how the ongoing development in Swedish compares to the development in the other case-poor Germanic languages, in the present and in the past.
5 The distribution and presentation of the results, and cooperation
The overall results have been published on the project’s homepage and presented at numerous workshops and conferences. A paper with the title “Predicative case in Swedish: Default or not?” has been published on Lingbuzz, and two other works are in the making: “Case drift in Swedish”, and an edited special issue on “Case-impoverished Germanic” that will be published 2020 in Acta Linguistica Hafniensia (where the final version of “Predicative case in Swedish: Default or not?” will also appear).
Cooperation: By taking care of statistics, Joost van de Weijer made an invaluable contribution. Lars-Olof Delsing, SOL, and Jeffrey K. Parrott Palacký University Olomouc, helped organize the workshop “Case-impoverished Germanic). Lars Heltoft is the coeditor of the Acta volume.