Content uploaded by Simon Kasper
Author content
All content in this area was uploaded by Simon Kasper on Aug 01, 2016
Content may be subject to copyright.
1
PREPRINT: Considerably revised version published 2015 as “Linking syntax and semantics
of adnominal possession in the history of German” in Gianollo, Chiara/Jäger, Agnes/Penka,
Doris (Hrsg.): Language change at the syntax-semantics interface. Berlin/New York: de
Gruyter (Trends in Linguistics: Studies and Monographs 278), 57–99.
Structural and functional aspects of adnominal possession in German.
A historical perspective
Simon Kasper
Research Center “Deutscher Sprachatlas”
Philipps-University Marburg
simon.kasper@staff.uni-marburg.de
Abstract: The present article is an attempt towards a unified picture about some central structural and functional
changes within complex (possessive) NPs in German. Firstly, the expressive adnominal strategies for possessive
relations in German are presented and the concept of possession characterized. After that major changes in the
relationship between structure and function within German adnominal (possessive) constructions throughout
different historical periods are traced as well as changes that are of a purely structural kind. In what follows these
changes are modeled using the Role and Reference framework and unification of the observed data is attempted.
The main components of this unification are: 1) The prenominal position in complex German NPs becomes
successively reanalyzed as a position for article expressions, i.e., RRG operators. 2) This development competes
with the tendency to have the more referential/definite/agentive entity in the relation to be expressed before the
less referential/definite/agentive expression in the NP. 3) The relationship between expressive strategies and
conceptual domains underlying possession in the periods of German can partially be explained as a result of this
competition.
1. Introduction
2. Possession and the spectrum of expressive strategies
3. Expressive strategies for adnominal possession in the diachrony of German
4. Structural changes in the German noun phrase: an RRG perspective
5. A tentative unification of structural and functional changes in German adnominal
possession
1. Introduction
In the New High German Standard language there are a number of strategies available for the
expression of possessive relations by means of (complex) noun phrases.1 The most important
ones are given in (1). In general, the head, or nucleus (adopting Role and Reference Grammar
(RRG) terminology)2, of these constructions is the possessum expression, abbreviated “H”. Its
1 The notion of possession is clarified in section 2.
2 See van Valin & LaPolla 1997, van Valin 2005.
2
case is that of the whole NP in the syntactic context of the clause. In the following examples,
this is the nominative case. The dependent, or argument, of the head/nucleus is, if present,
abbreviated “D”. Its form depends on the construction as a whole.
(1) a. D-H:
Autoreifen
car-tire
‘car tire’
b. Dgen H (prenominal genitive construction):
Haralds Hund
Harald.GEN dog
‘Harald’s dog’
c. H Dgen (postnominal genitive construction):
der Hund des Mannes
Det.NOM dog Det.GEN man
‘the man’s dog’
d. H von Ddat:
der Hund von dem Mann/von Harald
Det.NOM dog of Det.DAT man/of Harald
lit. ‘the dog from the man/Harald’ / ‘the man’s/Harald’s dog’
e. H an/bei D:
der Nachteil an/bei Harald
Det.NOM downside at Harald
lit. ‘the downside at Harald’
f. H zu D:
der Bruder zu Harald
Det.NOM brother to Harald
lit. ‘the brother to Harald’/‘Harald’s brother’
g. Poss H
sein Hund
his.NOM dog
‘his dog’
In non-Standard varieties, ranging from colloquial German “down” to local dialects
(cf. Schmidt/Herrgen 2011) this number reduces by the variant (1c) due to the almost
complete absence of the genitive in these varieties (cf. Behaghel 1923: 479, Schirmunski
2010: 496, Mironow 1957: 391–398). There are only residues of constructions like (1b)
3
containing proper names, but only in some regional varieties/dialects, e.g., Low German and
the Wallis (cf. Wipf 1910, Bohnenberger 1913, Henzen 1932, Bart 2006). At the same time
there is one construction in non-Standard German that is completely absent in the standard
language but that can be found in almost all regional varieties/dialects. It is given in (2).
(2) Ddat Poss H (adnominal possessive dative)
dem Mann/(dem) Harald sein Hund
Det.DAT man (Det.NOM) Harald his.NOM dog.NOM
lit. ‘the man his dog’/‘the man’s dog’
Case is mostly indicated by means of the determiner system in German. Taking the
singular, masculine, definite article as the paradigmatic case, Standard German can be
described as a four case system exhibiting nominative, genitive, dative, and accusative. Non-
standard varieties display, as mentioned above, an almost complete loss of the genitive. Most
Low German varieties lack the dative as well which developed syncretism with the accusative
case. Western Middle German and Western High German varieties show a wide-spread
syncretism between the nominative and the accusative with a preserved dative. Some varieties
show a complete loss of case distinctions, resulting in a common case (cf. Shrier 1965, Koß
1983). The construction in (2) is usually called an “adnominal possessive dative” but it is
obviously a nomenclature that draws from the Standard German case system, where we do not
find this construction. This should be kept in mind throughout this article.
In the present article I will demonstrate some changes in the structure-function
relationship that have come about in the expression of adnominal possession in German since
the Old High German period with a special focus on the adnominal possessive dative and its
closest relatives from a functional perspective: the attributive genitive constructions (1b, c)
and the postnominal von construction (1d). In connection with these changes earlier research
suggests that major, primarily structural changes happened in the German noun phrase (or
determiner phrase, respectively) to the extent that the definite article developed in the Old
High German period (cf. Oubouzar 1992), that possessive pronouns (like those in (1g) and
(2)) and genitive attributes (like those in (1b) and (c)) changed their grammatical status (cf.
Demske 2001), and that the adnominal possessive dative came into existence (cf. Zifonun
2003, Fleischer & Schallert 2011: 96–99), whose origins are not quite clear (e.g., Weiß 2012).
From a functional perspective, it is expected that the loss of the genitive in most German
varieties triggered a functional pressure to the degree that some means of expressing the
associated meaning had to be “found” or “invented”, that the Middle High German split
between pronominal and postnominal genitive attributes (see (1b) and (1c), cf. Ebert 1986:
89–98) and the grammaticalization of the formerly ablative/locative von ‘of’/‘from’ (see (1d))
had functional repercussions.
In the remaining parts, I will first characterize the concept of possession (section 2),
then I will trace the changes in the relationship between structure and function within German
adnominal (possessive) constructions throughout different historical periods (section 3) as
well as changes that are of a purely structural kind (section 4). Section 5 models these changes
using the Role and Reference framework and attempts a unification of the observed data. The
4
main components of this unification are: 1) The prenominal position in complex German NPs
becomes successively reanalyzed as a position for article expressions, i.e., RRG operators. 2)
This development competes with the tendency to have the more referential/definite/agentive
entity in the relation to be expressed before the less referential/definite/agentive expression in
the NP. 3) The relationship between expressive strategies and conceptual domains underlying
possession in the periods of German can partially be explained as a result of this competition.
2. Possession and the spectrum of expressive strategies
The term possession suggests that there is an easily definable concept ‘possession’. This
impression is deceptive because several different conceptual domains contribute to what we
call possession (cf. Chappell & McGregor 1995, Heine 1997, Lehmann 1998, Seiler 1983,
2009). Instead of giving an exhaustive analysis of the conceptual sources, or ingredients, of
possession (for this, see Heine 1997), I confine myself to pointing out its most important
conceptual sources on the basis of the lexical or original meaning of the “construction
markers” (Koptjevskaja-Tamm 2003) occurring in the constructions in question, i.e., (1b) to
(g) and (2).3 This procedure yields results that are easily comparable to those of Heine (1997).
It is important to note that in contemporary German the prepositional construction markers
found in (1b, d, e, and f) are nearly fully grammaticalized in these constructions with only
little spatial meanings left. That means what today are grammatical constructions expressing
possessive relations make use of markers that once indicated spatial relations and continue to
indicate spatial relations in other constructions in German. This is given in Figure 1:
H TO D / H FROM Ddat / H AT D / H Dgen / Dgen H / Poss H / Ddat Poss H
‘allative’ ‘ablative’ ‘locative’ ‘meronymic/partitive’ ‘possessive
(ownership)’
(1f) (1d) (1e) (1b, 1c, 1g) (1b, 1c, 2)
Figure 1: Conceptual sources of possession
In the H TO D pattern (which conforms to (1f)), zu ‘to’ is originally and in other
constructions in German a goal marker. von ‘from’ in the H FROM Ddat construction (cf. (1d))
is originally and in other constructions in contemporary German a source marker. an/bei ‘at’
in the H AT D construction (cf. (1e)) indicates location in less grammaticalized constructions.
The old adnominal genitive (cf. (1b) and (c)) is usually and as far as we know considered to
express meronymic and partitive and possessive (ownership) relations (besides other relations
that do not belong to the realm of adnominal possession, cf. Behaghel 1923: 485ff., Wilmanns
1909: 575).4 The dative is usually connected to possession (ownership) and participation
3 Construction markers are elements within the NP that indicate a possessive relation between a H and a PR.
4 The genitives named objectivus, qualitatis, explicativus, and definitivus belong to the class of non-partitive and
non-possessive (ownership) genitives (cf. Eroms 2000: 282/283). The genitivus subjectivus is sometimes treated
as a type of the genitivus possessivus (cf. Wilmanns 1909: 600. Extending this argument, one could also treat the
5
(besides other relations that do not belong to the realm of adnominal possession, cf. Behaghel
1923: 609ff.). On the far right of the conceptual sources of possession there is ‘possessive
(ownership)’. This is a terminological inconvenience which I do not know how to circumvent.
That means the complex conceptual domain of possession has at its core the concept of
‘possession as ownership’, besides the diverse spatial relations and part-whole relations
presented in Figure 1.
When looking at the conceptual sources of possession as presented in Figure 1, it
seems to have a spatial perceptual basis: allative, ablative, locative, and meronymic/partitive
relations can well be characterized as spatial relations. Only possession as ownership does not
immediately conform to this pattern. It seems to have cultural rather than spatial perceptual
origins. However, when looking into possession as ownership it might be possible to reduce
this concept to location plus something like control (cf. Stolz et al. 2008: 17ff.). That means
the possessum “is at” (i.e., location) the possessor, who has the possessum “at his/her
disposal” (i.e., control). This also seems to lie at the heart of the verb haben ‘have’. When we
look into the semantics of haben we actually find that English have and German haben derive
from Germanic *habǣ- which is the durative form of Germanic *hafja- ‘lift’ (‘heben’) (cf.
Kluge 252011). Thus, have has gone through a process of meaning extension at the beginning
of which it meant something like ‘hold’ (German halten) which can very well be treated as
the durative of ‘lift’. Applying these insights to the spatial conceptual structures underlying
possession, one can identify it as the combination of H’s being located at D and at the same
time as D’s controlling H.
Obviously, the relations we are talking about are binary ones, either spatial or
possessive. The corresponding predicate-argument structures could look like (3a) and (b):
(3a) GO-TO/COME-FROM/BE-AT/BE-PART-OF (x, y)
(3b) POSSESS (y, x)
The reversal between the arguments is the result of the fact that if x is at y and if y controls x,
then y possesses x.
Another dimension of the domain of possession is the alienability/inalienability
dimension: When looking at (4a) and (b) we find a potential meaning difference:
(4a) Haralds Bein
Harald.GEN leg
‘Harald’s leg’
(4b) das Bein an Harald
Det.NOM leg at Harald
lit. ‘The leg at Harald’/‘The leg attached to Harald’
genitivus auctoris as a possessive genitive.) However, as will become evident in the text, possession (ownership)
is reduced to concrete and spatial relationships between objects. The subjective genitive and the genitive of the
auctor do not express relations of that kind because they involve events or abstract entities that the “subject” or
“auctor” bring about.
6
(4a) is quite natural an expression of the inalienable relation between Harald and his
leg. When hearing (4b), in contrast, we infer that the leg in question need not be, or can even
be excluded to be, Harald’s “inherent” leg. The rationale seems to be that the Dgen H
construction is usually used to code inalienable relations where the leg is initially understood
to be inherently relational, while the H an/bei D construction is usually used to code alienable
relations, where the leg is initially understood to stand in an established spatial relation with
Harald (cf. Seiler 1983).
Besides the conceptual sources of possession and the alienability/established vs.
inalienability/inherent opposition, animacy also seems to play a role. Note that from the
perspective of the conceptual sources of possession depicted in Figure 1, all possessors except
that in the ‘possessive (ownership)’ and – maybe – the ‘allative’ relations should be open to
any degree of animacy as listed in (5):
(5) human > kin > animate > inanimate > abstract
This is indicated by the examples in (6):
(6a) *dem Haus sein Balkon
Det.DAT house its balcony
lit. ‘the house its roof’/ ‘the roof of the house’
(6b) ?der Balkon zu dem Haus
Det balcony to Det.DAT house
(6c) der Balkon des Hauses / vom Haus / am Haus
Det.NOM balcony Det.GEN house.GEN / from-Det.DAT house / at-Det.DAT house
‘the balcony of the house’ / ‘the balcony at the house’
(6d) sein Balkon
its balcony
‘its [= that of the house] balcony’
In other words, ‘possession (ownership)’ seems to be restricted to at least animate
entities, since its possessor must be capable of executing control (6a). The status of (6b) is
difficult to assess. Its acceptability seems to increase when nominalizations of allative motion
verbs appear as possessa: die Einfahrt zu dem Haus ‘the gateway to the house’. The H zu D
construction thus seems to be the least grammaticalized one of the spatially motivated
possessive constructions.
Taken together, the expressive strategies with respect to adnominal possession found
in different periods of the German language have to be located in the abovementioned
functional spectrum ranging from the conceptual sources of the possessive relation over the
alienability/established vs. inalienability/inherent opposition to the animacy dimension. One
7
can expect that the conceptual domains remain constant through history and require
expressive strategies, while these syntactic strategies change. The next step is therefore to
look at the structural options that are available in OHG and MHG for the adnominal
expression of possession. I will focus on the basic strategies represented by the H von D
construction, the genitive constructions, and the adnominal possessive dative.
3. Expressive strategies for adnominal possession in the diachrony of
German
3.1 Expressive strategies for adnominal possession in Old High German
Taking the position as a starting point that speakers of OHG need expressive strategies for the
expression of ablative, meronymic/partititive and possessive (ownership) relations, we find
that the corresponding structural options differ from those found in contemporary Standard
German and the non-standard varieties (see (1) and (2)).
The OHG analogue of the contemporary German H von D construction is the H
fon(n)(e/a) D construction. It is usually assumed not to express meronymic/partitive and
possessive (ownership) relations, but mainly ablative ones, even if metaphoric (e.g., temporal)
ones (cf. Behaghel 1924, 33–35). This observation is related to another observation according
to which the source of an ablative relation is often inanimate, often even a location (or point
in time), and that the relation is an established one, not an inherent one. In addition, the
preposition fon(n)(e/a) is mostly used in adverbal contexts. (7) below demonstrates that if the
noun governed by fóne ‘from’ is in Latin, it bears the ablative.
(7) Fóne déro questione chúmet si ad phisicam disputationem
From this.DAT question come.3 she.3NOM to physiological.ACC discussion.ACC
‘From this question she comes to a physiological discussion.’
Notker, Consolatio, p. 182, l. 12
(8) gibót iz ouh zi wáru \ ther kéisor fona Rúmu.
command.3 it.ACC and indeed Det.NOM emperor from Rome
‘And indeed, the emperor commanded it from Rome.’
Otfrid, I, 11, 2
(9) Sih scéident […] gilíabe [:] Múater fona kínde […].
each-other separate.3PL loved ones.NOM mother.NOM from child.DAT
‘Loved ones … separated from each other: … Mother from child…’
Otfrid, V, 20, 39–41
When we look at the analogues of NHG genitive constructions in OHG we find that
there are two functionally separated constructions. In partial revision of Behaghel (1932: 181–
193), Carr (1933) could show that in OHG non-partitive genitives almost always appear in
8
prenominal position. They occur in postnominal position only when a Latin original exhibits
this order. That means all genitive types (see note 4) as well as the one expressing possession
(ownership) appear in front of the head/core. An example for a (probably) genitivus qualitatis
is given in (10), a genitivus possessivus is given in (11). On the other hand, partitive (and
meronymic) genitives usually follow the core/head (Behaghel 1932: 178).
(10) uzs fona paradises bliidhnissu Isidor, 5, 10
away from paradise.GEN joys.DAT
‘away from the joys of paradise’
(11) dhiz ist chiquhedan in unseres druhtines nemin Isidor, 3, 3
this.NOM is said in our.GEN lord.GEN name.DAT
‘this is said in the name of our lord’
Some changes occur with Notker in the late OHG period. From this time on there
seems to be a tendency to place names and designations of persons in the prenominal/pre-
head position and to place inanimate and abstract entities in the postnominal/post-head
position (cf. Carr 1933). A consequence of this development should be that non-partitive
inanimate genitives should now be placed after the head/core of the complex NP, differently
from what we find in (10) and in the time before Notker.
(12) díu geskáft téro dingo Notker, Consolatio, p. 81, l. 16
Det.NOM creation.NOM Det.GEN things.GEN
‘the creation of things’
Turning to the OHG analogue of the adnominal possessive dative (Ddat Poss H) in
most contemporary German varieties, we can with some certainty say that this construction
does not yet exist. The only OHG constructions that come close to that construction are
exemplified by (13) and (14):
(13) Sámenont ímo sîne heîligen Notker, Psalter, p. 172, l. 5
gather.3PL him.DAT his.NOM.PL saints.NOM
‘His saints gathered for him‘
(14) du uuart demo balderes uolon sinuuoz birenkict
there was Det.DAT Balder.GEN colt.DAT his/its.NOM-foot.NOM aff.-by-disl.PTCP
‘The foot of Balder’s colt was affected by dislocation’
(2. Merseburger Zauberspruch)
Usually, these constructions should be read as involving an adverbal free dativus
(in)commodi or a dative of pertinence (Pertinenzdativ). In the above case the syntactic
structures allows an adnominal possessive interpretation, but it also allows a free dative
reading in the sense of ‘his saints gathered for him’ and ‘it happened to the disadvantage of
9
Balder that the foot of his colt was affected by dislocation’. Therefore, we cannot be sure
about the status of these constructions. Other reasons why they should probably not be
considered a Ddat Poss H construction are discussed in later sections.
The picture that emerges with respect to constant conceptual domains and expressive
requirements for OHG is that the prenominal genitive (Dgen H) is the primary, if not only
construction by means of which adnominal possession (ownership) is expressed in OHG. The
postnominal genitive (H Dgen) construction is reserved for partitive/meronymic relations in the
time before Notker and to relations with inanimate Ds from Notker’s time on. There is no
adnominal possessive dative in OHG. These considerations are condensed in Figure 2 below
which is based on Figure 1 above. In the topmost line there are the conceptual sources of
today’s possessive constructions. They correspond to meaning domains (spatial ones like
‘allative’, ‘ablative’ etc. and not only spatial ones like ‘possessive (ownership)’). For these
meaning domains and within them for the specific meanings, Old High German requires
expressive strategies. These are indicated in the lowermost line. Through the triangles, these
expressive strategies cannot only be related to the specific meanings they are used to express
but also to their own conceptual sources.
H TO D / H FROM Ddat / H AT D / H Dgen / Dgen H / Poss H / Ddat Poss H
‘allative’ ‘ablative’ ‘locative’ ‘meronymic/partitive’ ‘possessive
(ownership)’
/ H fon(n)(e/a) Ddat / / H Dgen / Dgen H / / Ø
(& adverbal von) (partitive)
Figure 2: Old High German expressive strategies for adnominal possession
Taking the example of the ‘ablative’ meaning we see that there is a one-to-one
mapping between the conceptual source H FROM Ddat and the expressive strategy H
fon(n)(e/a) Ddat. We do not expect the ‘ablative’ meaning to disappear or to change through
history, but it is expected that syntactic constructions change with respect to the range of
meanings they can express. Taking OHG, we do not find that the H fon(n)(e/a) Ddat construction
deviates from its ablative source meaning.
3.2 Expressive strategies for adnominal possession in Middle High German
MHG too requires expressive means for the meanings depicted in Figure 1 above, i.e., from
‘allative’ to ‘possessive (ownership)’. Looking at the MHG analogue of the NHG H von D
10
construction first, we find it in the H von(e) D construction. It seems that the adnominal
von(e) phrase is more frequent than its OHG analogue – perhaps due to the more frequent
mentioning of (fictional) persons genealogically stemming from specific locations (e.g., (15)).
Besides that there is often an ambiguity between an ablative and a non-ablative reading:
Because MHG geschëhen ‘happen’ in (16) is not a motion verb, a purely adverbal ablative
meaning of rede von in (lit. ‘speech from him’) is unlikely and an overlap of the H von(e) D
construction with the meaning of the adnominal genitivus auctoris seems possible. The use of
the latter would have required a genitive form of the personal pronoun of the 3rd Ps. Sg.
Masc., i.e., something like sîner rede ‘the speech of him’, a construction that is
indistinguishable from one with a possessive pronoun. As Ebert (1986: 92) notes, the von
periphrasis becomes the primary strategy for expressing genitive functions, especially in
pronominal contexts where genitives are not marked distinctly. (17) seems to have a partitive
meaning, since an adverbal ablative reading of ein teil von dîner nœte ‘part of your sorrows’
seems to be excluded. In (18) the difference between the genealogical origin (ablative) and a
possessive reading begins to blur: dem vogte von den sahsen could be understood as ‘the
reeve of the Saxons’ and as ‘the reeve from the Saxons’.
(15) Sus sprach der künec von Brandigân Parzival 215, 15
thus spoke Det.NOM king.NOM of/from Brandigan
‘Thus spoke the king of/from Brandigan’
(16) dô diu rede von in geschach Parzival 389, 3
after Det.NOM speech from/of him.ACC happened
lit. ‘After the speech from him happened’
(17) «Obylôt, nu sage mir / ein teil von dîner nœte.» Parzival 373, 16/17
Obylot now say me.DAT a part from/of your.DAT sorrows.DAT
‘Obylot, entrust part of your sorrows to me.’
(18) Dem vogte von den Sahsen was daz wol geseit
Det.DAT reeve.DAT from/of Det.DAT saxons was that well known.PTCP
‘This was well known to the reeve from/of the Saxons.’
Nibelungenlied (A 208, 1; B 207, 1; C 210, 1 (vogete))
These observations are backed by Kiefer (1910; see also Behaghel 1924: 62, Ebert
1986: 92) who claims that from the 12th century on, the original genitive meanings begin to
be expressed by the H von(e) D and other constructions discussed below.
With respect to the genitive constructions the changes found in the OHG period
continue into the MHG period. According to Ebert (1986) and Demske (2001), building on
analyses of Carr (1933), expressions designating persons and names occur as prenominal
genitives, whereas expressions designating things and abstract entities are placed after the
11
head/core of the complex phrase as postnominal genitives.5 This development comes to an
end as early as in the 15th century. From the 16th century on (Early New High German period)
two other developments take place: Firstly, appellative genitive expressions designating
persons begin to “move” to postnominal position. This development will come to an end in
the 17th century, when these expressions are highly preferred in postnominal position (cf. Prell
2000, Demske 2001: 215–230). Secondly, prenominal partitive/meronymic genitives are
reanalyzed as appositional nominatives or accusatives due to the loss of genitival morphology,
e.g., ein fesla guten alten wein ‘a keg of good old wine’ (cf. Behaghel 1923: 532). These
reanalyzed genitives “need not” move into the postnominal position.
As indicated above, the first of these developments seems to have begun in pre-OHG
time. Germanic attributive genitives probably occurred prenominally, independent of their
semantics. In early OHG the partitive ones tend to occur postnominally. The data for ENHG
in the 17th are thus part of a process that has already been lasting for maybe a thousand years
then, and that is not yet finished today.
Turning back to MHG again and to the MHG analogue of the contemporary
adnominal possessive dative (Ddat Poss H) we find that MHG is the period where this
construction appears first and where we can find unambiguous examples. However, case on
the possessor is ambiguous between genitive and dative in (19). It is unambiguously a dative
in (20) but one should not infer from that that the former is also a dative because there are also
constructions with unambiguous genitives, as in (21).
(19) nv sihe ich rot von blv
ͦt Hagnen sein gewant.
now see I red of/from blood Hagen.GEN/DAT his.NOM clothes.NOM
‘Now I see Hagen’s clothes (are) bloodred.’
Nibelungenlied, A 1992, B 2052
(20) dô sach man trüebe unde naz / dem Bernaer sîniu ougen
Then saw one hazy and wet Det.DAT one-from-Berne.DAT his.ACC eyes.ACC
‘One saw the hazy and wet eyes of the one from Berne there.’
Dietrichs Flucht, 7817
(21) swaz ich von Parzivâl gesprach, / des sin âventiur mich wiste,
what I of Parzival told DET.GEN his.NOM adventure.NOM me.ACC know
‘what I told about Parzival whose adventure I came to know‘
Willehalm 4, 20
These constructions are few in MHG and they are not easy to find (cf. Weise 1898,
Kiefer 1910, Behaghel 1923: 638–640). However, what is remarkable is that in the few
examples we know the dependent possessor expression is mostly one designating a person
and mostly a proper name. The marginal MHG Ddat Poss H construction thus shares the
tendency with the attributive genitive constructions to place expressions designating persons
5 This process is carried through earlier for MHG prose than for verse. The latter seems to be more conservative
and influenced by considerations of metrics and rhyme (cf. Prell 2000).
12
in the prenominal position. In fact, there is no evidence for postnominal possessive datives in
the history of German.
The resulting pattern, i.e., the relationship between meanings that need expression and
the actual expressive strategies for MHG and ENHG are given in Figure 3 below.
H TO D / H FROM Ddat / H AT D / H Dgen / Dgen H / Poss H / Ddat Poss H
‘allative’ ‘ablative’ ‘locative’ ‘meronymic/partitive’ ‘possessive
(ownership)’
/ H von(e) Ddat / / H Dgen / Dgen H / / Ddat Poss H
(prose)
Figure 3: Middle High German/Early New High German expressive strategies for adnominal
possession
The dotted lines indicate that some expressive strategy begins to develop or to extend
its semantic range. In the case of MHG and ENHG, the H von(e) D construction begins to be
used for the expression of meronymic/partitive and possessive (ownership) meanings. Its
original ablative meaning lives on. The H Dgen construction, in OHG restricted to
meronymic/partitive meanings, begins to be extended to possessive (ownership) relations as
well. This happens in those contexts where appellative person expressions “move” into the
postnominal position. Possessors (ownership) are necessarily among these expressions. The
Dgen H construction is thus more and more restricted to relations with human possessors
expressed by proper names (and some kin expressions). At the same time we witness the
coming into existence of the first unambiguous adnominal possessive dative constructions. Its
MHG/ENHG characteristic is the occurrence of both dative (in MHG and ENHG) and
genitive (possibly in MHG, definitively in ENHG) possessors.
Without attempting a causal explanation, it seems noteworthy that functionally highly
relevant processes take place nearly simultaneously between the MHG and ENHG times: The
H von(e) D and the Ddat Poss H constructions begin to be used for originally genitival
functions (see Figure 3) in higher frequency at the time where the adverbal genitive object
begins to come under pressure by other objects (mostly prepositional and accusative object
types) (cf. Fleischer & Schallert 2011: 87–94 for an overview). Importantly, these changes
from genitive expressions to accusative and prepositional expressions result in the loss of the
original expressive strategy for partitivity. In the adverbal domain, partitivity thus has to be
13
expressed by other means, then, e.g., by means of the determiner/quantifier system (Brot
essen ‘eat some bread’ vs. ein Brot essen ‘eat a bread’, by prepositional objects with von (vom
Brot essen ‘eat some of the bread’; cf. Ebert 1986: 37/38).
3.3 Expressive strategies for adnominal possession in contemporary German non-standard
varieties
In the introduction I have mentioned that German non-standard varieties exhibit different
patterns of case syncretisms and different degrees of case loss. Very few of them have a
preserved – and productive – genitive, e.g., the highest alemannic dialects of Walser German
(cf. Wipf 1910, Bohnenberger 1913, Henzen 1932, Bart 2006). Three-case systems in Middle
and High German and two-case systems in Low German varieties are dominating. The nearly
complete missing of the genitive has important repercussions for the expressive strategies
associated with the constant meanings in Figure 1 above. The H von D construction is (with
its regional characteristics) omnipresent and has – from the perspective of its areal distribution
– become the primary strategy for the expression of meronymic/partitive and possessive
(ownership) relations (cf. Mironow 1957: 395, Schirmunski 2010: 495–500). The original
ablative meaning of this construction seems to be functionally outweighed by the
partitive/meronymic and possessive (ownership) meanings.6 In the H von D construction the
case of the possessor is governed by the preposition in dependence of the case system of the
variety in question. It may thus be a dative where this is a distinct case category (i.e., most
West Middle German and High German varieties) or an accusative (i.e., East Middle German
and most Low German varieties) or a common case (i.e., some Low German varieties). The H
von D construction can also be found in Walser German (cf. Henzen 1932: 100, 101/102) as
an expressive strategy for meronymic/partitive and possessive (relations).
Turning to the Ddat Poss H construction we find that it is the other main strategy for
expressing meanings formerly associated with the genitive in German non-standard varieties
besides the H von D construction. For colloquial German and many regional and dialectal
varieties it is assumed that the adnominal possessive dative is restricted to animate possessors
(e.g., Wegener 1985: 49, Mironow 1957: 496). This would indicate its restriction to “real”
possessive (ownership) relations. However, such claims neglect a considerable amount of
data. It seems that it is common in Low German varieties to have expressions like (22).
(22) mien Huus sien (Lindow et al. 1998, 160)
my house his/its
lit. ‘my house its [H]; the [H] of my house’;
Here, it is perfectly natural to use inanimate possessors. In the context of the research
project “Syntax hessischer Dialekte (SyHD)” ‘Syntax of Hessian Dialects’ (cf.
www.syhd.info and Fleischer, Kasper & Lenz 2012) we have collected data from Hessian
6 The ablative meaning can also be expressed using the preposition aus ‘out of, from’, as in der Schneider
aus/von Panama ‘the tailor from Panama’.
14
dialects which show that coming from the Southern Rhine-Franconian parts of Hesse the
construction seems to become increasingly grammaticalized. While the construction with
animate possessors like (23) can be found in the whole Hessian area (and most High and
Middle German areas), there are a number of constructions of the type exemplified by (24).
(23) de Gertrud ihr Bröll East-Hess., E2, 14, Kerzell_2
Det.DAT Gertrud her.NOM glasses.NOM
lit. ‘the Gertrud her glasses; Gertrud’s glasses’
(24) de Bopp ehrn Fuß Rhine-Franc., E3, 13, Ernsthofen_1
Det.DAT puppet.DAT her.NOM foot.NOM
lit. ‘the puppet her foot; the foot of the puppet’
These constructions with inanimate but human-like possessors in some respects seem
to enter Hesse from the South. That similarity to humans probably plays a role is indicated by
the fact that only one of about 700 informants produced an adnominal possessive dative for
the relation in (25) with an inanimate and non-anthropomorphic possessor (here: a (toy)
crane). The von construction was by far the most frequent variant here. (25) is the response of
the same informant as (24).
(25) de Oarm vun deum Kran Rhine-Franc., E3, 24, Ernsthofen_1
Det.NOM arm of your.DAT crane.DAT
lit. ‘the arm from your (toy) crane; the arm of your (toy) crane’
The Low German and Hessian data indicate that at least in some German varieties the
adnominal possessive dative may code partitive/meronymic relations as well, if they have a
possessor that bears some similarity to human beings.
It is also often purported that the Ddat Poss H construction does not appear with first
and second person possessors (e.g., Zifonun 2003: 101, Behaghel 1923: 638).7 However, in
Low German (predicative constructions) and Berlin German one can find them:
(26) dat is mien/dien sien? Lindow et al. 1998: 165
that is my/your his
lit. ‘that is my/your his; is this mine/yours?’
(27) meiner/deiner seiner Schiepek 1898/1909: 221, cit. in Weiß 2008: 393
my/your his
lit. ‘my/your his; mine/my / your(s)’
7 Behaghel even considers a possessor expressed by a third person pronoun impossible, i.e., ihm sein Hut (lit.
‘him his hat’. This is clearly contradicted by a wealth of data (e.g., Henn-Memmesheimer 1986: 144–146).
15
A last observation concerns the status of the possessive element. In some areas it lacks
the typical gender agreement with the possessor (see (22) to (24)). Instead, a default element
is placed between D and H, as exemplified by (28):8
(28) de Gertrud soi Brell
Det.DAT.F Gertrud his/its.NOM.F/N glasses.NOM
lit. ‘the Gertrud his/its glasses; Gertruds glasses’
Central Hessian, E2, 14, Niederweidbach_4
In the introduction it has been mentioned that the designation “adnominal possessive
dative” draws from Standard German nomenclature. However, depending on the case system
of the variety in question, case on the possessor may differ. In particular, Weiß (2008: 384)
claims that D in this construction bears the genitive in those varieties where this case category
is preserved, that it bears the dative, where the genitive is lost, that it bears the accusative,
where both genitive and dative are lost, and that it bears the common case, where all case
distinctions are lost. This can be summarized in (29):
(29) case of the possessor in the “adnominal possessive dative” construction (cf. Weiß 2008:
384)
gen > dat > acc > common
However, Weiß’ (2008) generalization does not account for the absence of the whole
construction in a variety. So Walser German lacks the adnominal Ddat Poss H construction.
Revising (29), we can state that it holds true, if a variety has this construction. Another
possible problem is that Henzen (1933: 101) states that in predicative constructions, the dative
is an option besides the genitive, yielding dits iš dm luikxas ‘this is the.DAT Lukas; Lukas
owns this’. Before we cannot exclude that this is change in progress, Weiß’ generalization
need not be dismissed.9
The resulting picture for those varieties lacking the genitive is that given in Figure 4.
The picture for varieties spoken in the Swiss canton of Valais as an example for a variety with
a preserved genitive is given in Figure 5.
H TO D / H FROM Ddat / H AT D / H Dgen / Dgen H / Poss H / Ddat Poss H
‘allative’ ‘ablative’ ‘locative’ ‘meronymic/partitive’ ‘possessive
(ownership)’
8 This observation needs to be distinguished from another one: In some West Middle German varieties, feminine
proper nouns get the neuter gender: das Gertrud ‘the.NEUT Gertrud’. In the PRdat Poss H construction the Poss
element agrees with the neuter of PR, taking a form of sein ‘its’.
9 In fact, Heinzen (1932: 101) mentions the optional dative in predicative possessive constructions in the context
of his discussion of “changes that alter the traditional status of the genitive” (“[…] daß der gen. auf dem wege
ist, in seiner althergebrachten geltung veränderungen zu erfahren.”)
16
/ H von Ddat / / Ø / Ø / / Ddat Poss H
(common: aus) (apposition, Det/Quant. systems)
Figure 4: Expressive strategies for adnominal possession in contemporary German non-
standard varieties (genitive lost)
The dotted lines leading from the H von Ddat construction to the ablative meaning
indicates that the primary function of this construction is no longer the expression of ablative
relations, but that of meronymic/partitive and possessive (ownership) relations. The loss of
the genitive constructions for expressing partitivity/meronymy is further compensated for by
appositional constructions and the determiner/quantifier systems. The second expressive
strategy for possessive (ownership) relations is the Ddat Poss H construction. The data from
(22) and (24) are accounted for by the dotted lines leaving the Ddat Poss H construction and
encompassing also meronymic/partitive relations.
H TO D / H FROM Ddat / H AT D / H Dgen / Dgen H / Poss H / Ddat Poss H
‘allative’ ‘ablative’ ‘locative’ ‘meronymic/partitive’ ‘possessive
(ownership)’
/ H von Ddat / / H Dgen / Dgen H / / Ddat Poss H
Figure 5: Expressive strategies for adnominal possession in a contemporary German non-
standard variety (genitive preserved)
The pattern in Figure 5 is mainly based on the data of Bart (2006), collected in the
context of the “Syntaktischer Atlas der deutschen Schweiz (SADS)” ‘Syntactic Atlas of Swiss
German Dialects’ (see http://www.ds.uzh.ch/dialektsyntax/ and references there). In the
region in question the Ddat Poss H is only marginally used. This may be due to the fact that the
prenominal genitive is still used for many originally genitival functions. However, the
observation that the H von Ddat construction is used very frequently for both
meronymic/partitive and possessive relations suggests that there is no simple relationship
between the presence or absence of the genitive and the presence or absence of alternative
strategies.10
10 I cannot give any data for postnominal genitives.
17
3.4 Expressive strategies for adnominal possession in contemporary Standard German
The structure-function connections in Standard German can be discussed rather briefly. The H
von D construction is used for meronymic/partitive and possessive (ownership) relations.
Ablative relations are primarily expressed by a construction with aus ‘out of, from’. The
prenominal and postnominal genitives can be connected to earlier developments: Only proper
nouns and some kin expressions occur in prenominal position, otherwise this construction
seems rather archaic (cf. Demske 2001: 251). Due to the necessary animate status of the
possessor this construction tends to be interpreted as a possessive (ownership) one (see sec. 5
below on this point) but partitive/meronymic interpretations are also possible. All other
genitives occur in the postnominal position. For this construction no restriction to either
partitive/meronymic or possessive (ownership) relations can be postulated. The Ddat Poss H
construction is absent in Standard German. Note that in both Wallis German and Standard
German the presence of the genitive correlates with the absence or marginal status of this
otherwise highly prominent construction. These considerations are summarized in Figure 6.
The dotted lines leaving the prenominal genitive construction and embracing the
meronymic/partitive meaning indicates that with this construction a possessive (ownership)
relation is dominant.
H TO D / H FROM Ddat / H AT D / H Dgen / Dgen H / Poss H / Ddat Poss H
‘allative’ ‘ablative’ ‘locative’ ‘meronymic/partitive’ ‘possessive
(ownership)’
/ H von Ddat / / H Dgen / Dgen H / / Ø
(abl. mit aus) (D: proper N)
Figure 6: Expressive strategies for adnominal possession in contemporary Standard German
After this rather global sketch of the states and processes in the structural-functional
spectrum of German adpositional possessive constructions, the next section takes a closer
look at rather subtle structural changes within the German NP – changes that must be part of
an explanation of the changes reported on above.
4. Structural changes in the German noun phrase: an RRG perspective
18
The developments sketched in section 3 seem to follow an inherent “logic”: Three
developments are observable that extend over several periods in the history of the German
language:
i. The genitive as a case category is gradually substituted by means of other
constructions.
ii. Prenominal attributive genitives gradually move into the postnominal position,
having left only proper nouns and some kin expressions in the prenominal position.
iii. The almost complete decay of the genitive case category and with it the original
expressive strategy for partitivity/meronymy and possession (ownership) is
accompanied by the expansion of the H von D construction from an original ablative
meaning to the original genitive meanings and by the coming about of the Ddat Poss H
construction.
The causes of (i) seem numerous and are not entirely clear. It has been mentioned
above that adverbal genitives were gradually replaced by concurrent constructions with
accusatives and prepositional objects. This development could be associated with the loss of
an aspectual category in German, but without providing a sufficient explanation, since it does
not cover ditransitive constructions involving a genitive (cf. Fleischer & Schallert 2011: 83–
101 for an overview). Another factor seems to be the decay of distinct inflectional genitive
forms causing syncretistic forms with other case categories (cf. Behaghel 1923: 479–483).
Whether or not, or in which way this is causally connected to the aforementioned
considerations is not entirely clear.
The causes of (ii) can probably be connected to structural changes in the German NP.
The general pattern seems to be this: Prenominal elements in German NPs are gradually
reanalyzed as determiners (for details on this and the following, see Demske 2001). Part of
this is a) the development of the definite article out of the demonstrative pronoun in OHG (cf.
Oubouzar 1992), b) the reanalysis of the possessive pronoun as adjectives (OHG or pre-OHG)
and then as determiners (NHG), c) the reanalysis of prenominal attributes as determiners
(NHG) (cf. Weiß 2008).
a) According to Demske (2001), the demonstrative pronoun in OHG marks pragmatic
definiteness, its absence unmarked definiteness. Where it stands, the demonstrative occurs
with adjectival attributes of the weak declension. Strong attributes occur in the absence of the
demonstrative. Unique referents occur without a demonstrative in early OHG. “In OHG the
definite article word accompanies nominals representing sortal concepts. The noun provides
those sortal information about the referent which are applicable to precisely one entity in the
given context.” (Demske 2001: 116, my translation). Thus, the distribution of strong and weak
attributive adjectives is semantically motivated in terms of pragmatic definiteness. However,
already in late OHG (with Notker) demonstratives begin to be used with semantically definite
expressions (cf. Oubouzar 1992): The referents of these expressions are identifiable
independently of the context, e.g., diu sunna ‘the sun’, so the demonstrative is redundant. And
in NHG die Tochter eines armen Bauern ‘the daughter of a poor farmer’ “the definite article
does not express definite reference, but the definiteness of the connection between the head
noun and the article.” (Demske 2001: 109). The association of adjectival declension
19
(strong/weak) and definiteness that worked in OHG may have become blurred in ENHG,
where the adjectival inflection was partially lost. The relation between article and noun had to
be reanalyzed as being morphologically licensed, not semantically: The determiner governs
the declension of the attributive adjective morphologically, not semantically.
b) Related to this development is that of the possessive pronoun (cf. Demske 2001:
132–163). In OHG many possessive pronouns already show the inflection of the strong
adjective and agree with the head noun in case, number, and gender. However, some of them
behave like personal pronouns in the genitive. In MHG the possessive pronoun is established
as an adjective. It co-occurs with articles and does not determine the strong or weak forms of
attributive adjectives. It may occur prenominally or postnominally, like other adjectives. In
NHG adnominal possessive pronouns are not adjectives any more. They show the inflectional
pattern of the indefinite article and show complementary distribution with following
adjectives as regards strong or weak declension. And in contrast to OHG and MHG, they
show complementary distribution with the definite and indefinite articles and cannot occur
postnominally. NPs with possessive articles are definite in the sense that their referents are
identifiable. Possessive articles are the product of the same processes as the definite article.
The result are two inflectional paradigms for possessive pronouns in NHG: one for adnominal
possessive articles (sein schönes Buch ‘his nice book’), and one for possessive pronouns (as
in: Seins ist ein schönes Buch ‘His one is a nice book’).
c) The changes associated with attributive genitives discussed in section 3 can also be
connected to the aforementioned structural changes in the German NP (cf. Demske 2001:
208–230). Prenominal genitives in NHG are reanalyzed as determiners just like possessive
pronouns and as a consequence of the coming about of morphologically motivated definite
articles. They show complementary distribution with respect to definite articles (*das Peters
Buch ‘the Peter’s book’), unlike in earlier times. Postnominal, but not prenominal genitives
may be modified to the right (*Peters, der aus Bremen kommt, Buch ‘Peter’s who is from
Bremen book’ vs. das Buch Peters, der aus Bremen kommt ‘The book of Peter who is from
Bremen’) and to the left (*des armen Peters Buch ‘poor Peter’s book’ vs. das Buch des armen
Peters ‘poor Peter’s book’). This was possible until ENHG time. Just like definite articles and
possessive articles, prenominal genitives determine the declension of following adjectives
(Pauls erster Roman ‘Pauls first novel’). The strong declension indicates that -s does not bear
grammatical information except that indicating possession (ownership) (cf. Weiß 2008).
How can the changes in (iii) be reconciled with this? Prepositional attributes including
the H von D construction seem not to be affected by the reformation of the prenominal
position. Instead, this construction takes over the semantic range of the originally prenominal
genitives. When we look at the Ddat Poss H construction, we have another prenominal
attribute. Were the changes reported on above valid without exception, we would expect
similar diagnoses as with the possessive pronoun and the prenominal genitive. The
prenominal cluster indeed seems to determine the declension of following adjectives (dem
Paul sein erster Roman, lit. ‘the Paul his first novel’) and is in complementary distribution
with the definite article (*der dem Peter sein erster Roman ‘the.NOM the.DAT Peter his first
novel’). However, it may be modified to the left and to the right (dem armen Peter von
20
Gegenüber sein Buch ‘the.DAT poor Peter from across (from here) his book’) and it may be
multiplied, as in (30), taken from the SyHD data:
(30) em Peere seiner Freundin ihr nei Audo
Det.DAT Peter his.DAT girlfriend her.NOM new car
‘the new car of Peter’s girlfriend’
Rhine-Franc., Pt_E4D, 10, Ober-Kinzig_3
The latter observations do not fit in the pattern. The dative in the Ddat Poss H
construction cannot be considered a pure article expression but seems to have the status of a
constituent. Poss alone could be considered an article making H definite, in accordance to the
reanalysis of possessive pronouns to articles. But the data in (26) to (28) above seem to point
in another direction. Poss in these sentences is no longer an anaphor, since it does not agree
with D in gender. If it determines the strong declension in following adjectives in these
varieties, then it also lacks grammatical features and must be considered an element indicating
nothing but a possessive (ownership) relationship. Poss would then no longer be an article in
varieties where (28) is possible ((26) and (27) being predicative constructions). I would
therefore propose the following: Poss in Ddat Poss H constructions is in different states of
grammaticalization in contemporary German varieties. Where it develops into a mere marker
of a possessive (ownership) relationship, the construction as a whole including Ddat is on its
way to being reanalyzed as a definite article expression, just like NHG possessive articles and
prenominal genitives. In varieties where it maintains its grammatical features, it may work as
a definite article, the Ddat component being an argument expression in terms of Role and
Reference Grammar (RRG).
In what follows I propose structural representations of OHG, MHG/ENHG, and
standard NHG possessive NPs as well as those from non-standard varieties in the context of
RRG (cf. van Valin & LaPolla 1997, van Valin 2005). These structural representations are
supposed to illustrate the aforementioned structural changes within the German (possessive)
NP throughout its history.
For those who are not familiar with the framework: Role and Reference Grammar
stands in the functional tradition of theories of grammar and views syntax as standing in the
service of semantics and pragmatics. Syntactic units are therefore considered to have semantic
correlates or to be pragmatically motivated. The syntactic representation is organized into
layers that correspond to bits of predicate-argument structures, i.e., semantic representation.
The idea is that this allows a representation of syntax that is valid for all languages. The
correlations are given below (cf. van Valin & LaPolla 1997: 27).
(31) Correlations between semantic elements and syntactic units in RRG
semantic elements syntactic units
predicate/reference nucleus
argument core argument
non-argument periphery
predicate + arguments core
21
predicate + arguments + non-arguments clause/NP (= core +
periphery)
Inside the clause, but outside the core there is another, non-universal position called
the precore slot (PrCS) where wh-expressions or topicalized elements can stand, for instance.
Left-detached adverbs, NPs, or PPs that are set off the rest of the clause by an intonational
break etc. can stand before precore elements (e.g., Yesterday, what did you show to Peter?)
and their position is called the left-detached position (LDP), a position outside the clause, but
inside the sentence. These positions are mainly pragmatically motivated. Categories like
definiteness, aspect, negation, modality, tense, evidentiality, illocutionary force and so on are
treated as operators, since they are qualitatively different from the semantic notions in
predicate-argument structures that underlie syntactic categories. Operators get a
representation separate from the syntactic structure, unlike functional categories in the
Principles & Parameters or Minimalist frameworks. Different operators in RRG apply to
different layers of the clause/NP, where higher layer operators have scope over lower layer
operators. RRG acknowledges the considerable parallelism between clauses and NPs and
grants NPs a layered structure, too. One difference is that the analogue of a clausal nucleus
dominates a PRED ‘predicate’, while the nominal nucleus dominates a REF ‘reference’.
There is nothing in an NP that exactly fits the clausal PrCS, LDP, or the core-internal subject
position, since prenominal NPs behave like PrCS, LDP or core internal subject elements in
some respects. Therefore, the NP structure includes a unique NP-initial position (NPIP) for
these elements. NPIP is outside the core but inside the NP. The general schema of the layered
structure of an NP is given in Figure 711:
11 Operators: ADJ/N: adjectival/nominal modification; NASP: nominal aspect; NUM: number; QUNT: quantity;
NEG: negation; DEF: definiteness; DEIC: deictics. On details see van Valin & LaPolla 1997: 56).
22
NP
CORE
NUC
REF
N
(NPIP)
(ARG)(ARG) (ARG)
(PERIPHERY)
(PERIPHERY)
NUC
NUC
CORE
CORE
NP
NP
ADJ/N
NASP
NUM
QUNT
NEG CORE
DEF
DEIC
Figure 7: Layered structure of the NP
The projection below “N” is the operator projection, above N resides the syntactic
representation with its different layers.
Applying this to OHG Dgen H construction expressing possessive (ownership) relation
yields Figures 8 and 9. What these structures illustrate is the change from an argument
personal pronoun (ira.GEN.F ‘her/of her’) to a possessive adjective standing in NPIP
(unseres.GEN ‘of our’) taking place partially in OHG and in the transition to MHG.
NP
CORE
NUC
REF
N
(PERIPHERY)
Ø ira Ø hús (Otfrid, 6, 3)
NP
gen
PRO
semantics:
POSS (3FM, house) ARG
Figure 8: OHG Dgen H construction expressing possessive (ownership) relation;
(she.GEN.F house.NOM; ‘her house/house of hers’)
23
NP
CORE
NUC
REF
N
NPIP
(PERIPHERY)
NPDEF
in unseres Ø druhtines Ø nemin
PP
CORE
NUC
PRED
P
NP
gen
CORE
NUC
REF
N
AP
gen
semantics:
POSS (lord, name)
ARG
A
PRON
Figure 9: OHG Dgen H construction expressing possessive (ownership) relation;
(in our.GEN lord.GEN name.NOM ‘in the name of our lord’)
Note that in OHG, genitival attributes can be expanded to the left and to the right by
modification (periphery).
As discussed in section 3, all kinds of D occur prenominally in OHG, whereas
partitive/meronymic genitives occur postnominally. The fact that in late OHG non-partitive
(and, trivially, non-possessive) inanimate and abstract entities tend to be placed post-
nominally, is captured in Figure 10, which can therefore be viewed contrast to the prenominal
genitives in Figures 8 and 9.
24
NP
CORE
NUC
REF
N
(PERIPHERY)
NP
díu geskáft téro Ø dingo Ø
NPgen
CORE
NUC
REF
N
DEF NPDEF
ARG
semantics:
[do (x, )] CAUSE[BECOME exist (y)]
→ creation (x, y)
Figure 10: OHG H Dgen construction expressing non-possessive relation
In OHG possession (ownership) is expressed by genitival attributes. The expansion of
other strategies does not yet happen. PPs with fon(n)(e/a) ‘from’ following H in adjacent
position or animate NPdat preceding H in adjacent position must be interpreted adverbally.
The latter case is illustrated in Figure 11. It stands in contrast to MHG and younger adnominal
PPs headed by von ‘from/of’ (Figure 12). However, the adjacent position of the OHG PP
headed by fon(n)(e/a) to H of NP is a probable source of reanalysis towards the H fon(n)(e/a)
D construction.
PP
gibót iz […] ther kéisor fona Rúmu
SENTENCE
CORE
NUC
PRED
V
NP
acc
PRO
semantics:
DO (emperor [command (emperor, 3sg)])
ARG
CLAUSE
NP
nom
N
ARG
PERIPHERY
CORE
NUC
PRED
P
NP
dat
N
PROP
ARG
DEF/
DEIC NP
Figure 11: OHG adverbal PP headed by fon(n)(e/a); (command.3.PST it.ACC Det.NOM
emperor.3 from Rome.DAT ‘the emperor commanded it from Rome)
25
Figure 12 illustrates an example from MHG where the abovementioned reanalysis
from an adverbal ablative PP to an adnominal ablative or possessive PP has already taken
place. The complex NP below seems to be ambiguous between an ablative and a possessive
(ownership) interpretation. In the ablative interpretation (COME-FROM (reeve, saxons)) the
preposition von ‘from’ is a predicating element and has a core and a nucleus. In the possessive
interpretation von ‘of’ is no predicate and lacks a nucleus and a core. It is a mere construction
marker as indicated by its constructional schema. These schemas contain “idiosyncratic,
language-specific features of constructions […].” (van Valin 2005: 132). In Figure 12 I have
specified von as a construction marker in the possessive reading in the corresponding
construction schema.
NP
CORE
NUC
REF
N
Dem vogte von den Sahsen
PP
(CORE)
(NUC)
(PRED)
P
NP
dat
CORE
NUC
REF
N
NP
DEF
CORE
NP
NUM
DEF
semantics:
COME-FROM (reeve, saxons)
POSS (saxons, reeve)
ARG
constructional schema:
semantics:
POSS (saxons, reeve)
…
morphology:
construction marker
von following
possessum
Figure 12: Ambiguous MHG H von(e) D construction expressing either ablative relation or
possessive (ownership) relation; (Det.DAT reeve.DAT from/of Det.DAT saxons.DAT ‘the
reeve from/of the Saxons’)
In MHG the Ddat Poss H construction too begins to become an expressive strategy for
possessive (ownership) relations. This strategy becomes increasingly prominent in ENHG and
is the primary means for expressing possession (ownership) in most contemporary non-
standard varieties. In some of these varieties it can already be used for partitive/meronymic
relations as well, pointing to a grammaticalization of the construction as a whole. Figure 13
illustrates a possessive example from MHG and a partitive/meronymic relation from a
contemporary Rhine-Franconian dialect.
26
NP
CORE
NUC
REF
N
MHG dem Ø Bernaer Ø sîniu ougen
NP
dat
CORE
NUC
REF
N
NP
DEF
CORE
NP
NUM
DEF
semantics:
POSS (one from Berne, eyes)
BE-PART-OF (foot, puppet)
NPIP
Rhine-Franc. de Ø Bopp Ø ehrn Fuß
(PERIPHERY) ?
constructional schema:
…
…
morphology:
construction marker
poss.pron. following
possessor
AP
gen
A
PRON
Figure 13: MHG and contemporary Rhine-Franconian Ddat Poss H construction expressing
possessive and partitive/meronymic relations, respectively; (Det.DAT one-from-Berne.DAT
his.NOM ougen.NOM ‘the eyes of the one from Berne’; Det.DAT puppet.DAT her.NOM
foot.NOM ‘the foot of the puppet’)
Following my proposal above, Poss in these constructions behaves like an adjective
and contributes definiteness to H (ougen ‘eyes’, Fuß ‘foot’), making the respective referents
identifiable in discourse. At the same time, it functions as the construction marker in this
construction, as indicated in the constructional schema. The dative NPs, coding the semantic
possessors, function as arguments to the respective cores/Hs. It is expected that NPdat can be
expanded to the left and to the right by means of modification. More data are necessary at this
point. In this state, the Ddat Poss part of the construction is not yet reanalyzed as an article
expression, but subsequent changes follow exactly this path. The structural reflexes of this
change are that Ddat is incorporated into NPIP which was formerly occupied by Poss alone.
This is a distinct process from the grammaticalization of the whole construction and is
restricted to its pre-core part, i.e., Ddat Poss. Figure 14 demonstrates this:
27
NP
CORE
NUC
REF
N
de Ø Gertrud Ø soi Brell
NPdat
NP
DEF
semantics:
POSS (gertrud, glasses) NPIP
NP
DEF
(PERIPHERY) ?
NPROP
constructional schema:
…
…
morphology:
construction marker
Poss following possessor
Figure 14: Contemporary Central-Hessian Ddat Poss H construction expressing possessive
(ownership) relation; Poss grammaticalized; (Det.DAT Gertrud.DAT his/its.NOM
glasses.NOM ‘Gertrud’s glasses’)
Poss does no longer behave like an adjective, can no longer function as an anaphor to
the possessor and constitutes the article expression for H/the core together with NPdat.
Whether or not NPdat can be expanded to the left and right by modification cannot be decided
without additional data, but if Ddat Poss functions as an article expression, I would predict that
it does not work.
Turning to Standard NHG we find a preserved genitive. As a possessor, only proper
nouns and some kin expression may stand prenominally, i.e., preceding H. An example is
given in Figure 15.
28
NP
CORE
NUC
REF
N
Ø Paul -s Ø erster Roman
NPgen
NPROP
semantics:
POSS (paul, novel)
NPIP
NP
DEF
NUCADJ
(PERIPHERY) X
constructional schema:
…
…
morphology:
construction marker
-s on possessor
Figure 15: Standard German Dgen H construction expressing possessive (ownership) relation;
(Paul.GEN-s.POSS first.NOM novel.NOM ‘Paul’s first novel’)
In Figure 15, Paul and the element -s together function as an article expression to
H/the core (Roman ‘novel’), making it definite. The categorial status of -s is difficult to
assess, because it is not clear if it is a genitive marker (cf. Demske 2001). In any case it is part
of NPIP. I consider it a construction marker, analogously to Poss and forms of von ‘of’ in the
above cases. Because of the reanalysis of Dgen towards an article expression, it cannot be
modified any more. Only if it had maintained its status as an argument, this would be
possible. Dgen has exactly this status when in postnominal position. This is given in Figure 16:
NP
CORE
NUC
REF
N
das Buch Ø Paul -s Ø
semantics:
POSS (paul, novel)
DEF
(PERIPHERY)
N
PROP
NP
gen
ARG
NP
constructional schema:
…
…
morphology:
construction marker
-s on possessor
29
Figure 16: Standard German H Dgen construction expressing possessive (ownership) relation;
(Det.NOM book.NOM Paul.GEN-s.POSS ‘Paul’s book’)
As discussed further above in this section, Dgen can indeed be modified in this position.
The similarity between the expressions in Figures 15 and 16 is only superficial. There is more
to these expressions except the positions of the genitive NP. The structural status of the
prenominal position makes the difference.
5. A tentative unification of structural and functional changes in German
adnominal possession
In sections 3 and 4 of this paper I have tried to trace changes in German adnominal NPs
revolving around the concept of possession (in the broad sense). These changes concerned a)
the relationship between adnominal expressive strategies in different periods of the German
language in dependence on supposedly constant conceptual-semantic differentiations that
require expression. The observed changes also concerned b) changes in the structure of the
German noun phrase that cannot easily be brought in connection with functional (conceptual-
semantic) considerations. Is there a unifying pattern in these developments?
Looking first at the changes occurring in the context of the prenominal and
postnominal genitive constructions, the following generalization in (32) seems to hold with
respect to the major developments:
(32) Expressions “moving” into postnominal position in the history of the German complex
NP adhere to the following hierarchy:
partitive expressions > non-person non-partitive expressions > appellative person expressions
(Germ. > OHG) (OHG > MHG) (MHG > ENHG)
Switching the perspective to the prenominal position yields the generalization in (33):
(33) expressions “staying” in prenominal position in the history of the German complex NP:
non-partitive expressions > person expressions > proper names(/kin names)
(Germ >OHG) (OHG > MHG) (MHG > ENHG)
On a closer look, these can be related to a sub-scale of the animacy hierarchy, namely
that pertaining to types of referential expressions. This scale is given in (34):
(34) proper name > kin expression > appellative expression (animate) > appellative
expression (inanimate) > abstract/mass noun
30
The rationale seems to be this: The concepts in (34) decrease from left to right with
respect to their inherent definiteness. That means proper names designate referents that are
unambiguously identifiable in a given discourse, kin expression are easily identifiable in
relation to someone in a given discourse, animate common nouns designate an indefinite
number of animate entities that are similar in a particular respect, etc. The connection to (32)
and (33) can be captured in the following way: Expressions moving into the postnominal
position in the periods of German (32) “climb up” the referential expressions scale (from right
to left). Expressions staying in prenominal position in the periods of German (33) reduce to
those at the top (left pole) of the referential expressions scale. That means those referents
whose identity is determined easiest in discourse remain in prenominal position. Referents
whose identity is less easy determined occur in postnominal position. At the same time the
exodus of expression types into the postnominal position means a tendency towards a
core>periphery or head>dependent order in the complex NP. The exception to this order is
constituted then by highly referential possessor expressions in the Dgen H construction in
contemporary Standard German and those non-standard varieties having preserved the
genitive.
This cannot be the whole story, since there is no plausible reason I am aware of why
appellatives designating inanimate entities should be less identifiable than those designating
animate entities. Because of this, a closer look at the relationship between
partitive/meronymic relations and possessive (ownership) relations seems necessary. In
particular, I propose the following principle:
(35) Pragmatic implicature:
Animate D expressions in a complex NP expressing conceptual sources of possession
are interpreted as possessors (ownership).
Why should this be the case and how does it contribute to the relationship between
(32)/(33) on the one hand and (34) on the other hand? Concerning the first question, there is a
widely acknowledged analogous principle at work with respect to the syntax-semantics
relationship. Van Valin & Wilkins (1996) on the basis of Holisky (1987) claim that in clauses,
animate causers (or effectors, in their terminology) are interpreted as intentional agents. This
leaves many lexical entries of verbs unspecified with respect to agentivity. Agentivity is
pragmatically inferred on the basis of a lexical feature of an argument, namely its animacy.
This is the reason why killers, hitters, and the like are interpreted as agents, if animate. Now
remember how possession (ownership) was characterized above in section 2: as a spatial
relationship plus what I have dubbed “control”. Control is an important ingredient of
agentivity (cf. Comrie 1989, Primus 1999). Now, I am not aware of any reason why Holisky’s
observation should hold for animate effectors but not for animate D expressions in the
constructions in question. This diagnosis also explains why the H von Ddat construction does
not only extend to partitive/meronymic meanings in the history of German but also to
possessive (ownership) ones, namely if Ddat is animate, thereby triggering (35) to become
active.
31
With respect to the second question, the relevance of (35) for the relationship between
(32)/(33) on the one hand and (34) on the other hand, we find that (34) is also in part a scale
of potential agentivity, and therefore of potential “real” possessors, i.e., those executing
control. From a neurolinguistic and cognitive perspective, language users and “the person on
the street” strive to identify the causer/controller of a sentence (cf. Bornkessel-Schlesewsky &
Schlesewsky 2009) and that of a real-life event (Kasper 2013) as fast as possible, respectively.
A similar process seems to be at work in the complex NPs in question. The processes within
adnominal genitive phrases have left potential agents in the prenominal position. (Note that in
virtually all possessive (ownership) relations the possessor is higher in animacy than the
possessum.)
However, language users’ cognitive strive for identifying the initiator/controller of any
event as soon as possible obviously competes with the structural tendency to develop a H>D
order within the noun phrase, corresponding to the order possessum>possessor. The only H,
i.e., possessor expressions that withstand this tendency, are highly referential, highly potent
(with respect to agentivity) entities, namely in the Dgen H in Standard German and non-
Standard varieties having preserved the genitive and the Ddat Poss H constructions in many
non-standard varieties. These constructions involving prenominal possessors in these varieties
are grammatical means for expressing possession (ownership) so that the pragmatic
implicature in (35) is no more necessary. The pragmatic implicature remains active in
postnominal possessor expressions, since these, though often animate or human, may be part
of “mere” partitive/meronymic relations.
Both principles – the pragmatic implicature in (35) as well as the search for the
agent/controller in relations – do not explain data where we find the Ddat Poss H construction
with an inanimate Ddat referent (as in (24) above from Rhine-Franconian involving a puppet;
(22) from a Low German variety involves a house).12 In these constructions inanimate D
expressions occur prenominally, although they are no possible agents/controllers. They can
only be interpreted as meronymic/partitive relations. One could argue that in the case of D
referring to a puppet empathy plays a role (cf. Kuno & Kaburaki 1977), causing human-like
entities to be treated as humans; this does not work for the Low German data. At present I
have no explanation for these data and can only speculate that they have to with an
idiosyncratic development of the adnominal possessive dative or the NP as a whole in the
history of Low German varieties.13
12 Remember that in Rhine-Franconian, a crane as “possessor” did not work (see (25)). It must at least be human-
like.
13 Lübben (1882: 108/109) and Lasch (1914), following him, cite Middle Low German data for the origins of the
Ddat Poss D construction that differ from those that are discussed for Middle and High German. Concerning the
latter, the usual explanations allude to the reanalysis of sympathetic adverbal datives (Er hat [dem Vater] [sein
Haus] angezündet (lit. ‘He set the father his house on fire/He set the father’s house on fire’ → [dem Vater sein
Haus] ist abgebrannt (lit. ‘The father his house is burned down/Father’s house burned down’) or the replacement
of the prenominal genitive by the dative plus an added emphasizing Poss (des Vaters Haus ‘Det.GEN
father.GEN house’ → des Vaters sein Haus ‘Det.GEN father.GEN his house’ → dem Vater sein Haus ‘Det.DAT
father.DAT his house’). The first hypothesis was put forward by Behaghel (1923), among others, the latter by
Grimm (1837). In contrast, Lübben finds the origins of the Low German variant of Ddat Poss D in examples like
dat wîf des he eren sone levendich makede (The.NOM woman whose.GEN he.NOM its.ACC son alive made;
‘the woman whose son he brought/woke to life’). The genitival relative pronoun was then used in the nominative
or accusative (deme, de dat pant sîn is (the-one.DAT who.NOM/ACC the.NOM deposit his is; ‘the one whose
32
6. Conclusion
This article constitutes the attempt to unify structural and functional observations concerning
adnominal NPs from the realm of possession in the diachrony and in contemporary varieties
of German which possibly reflect different developmental states in a continuum of changes
within the adnominal expression of possession. It has been demonstrated that different lines of
change are at work in the complex German NP, for instance that concerning the prenominal
position, that concerning the possessive element, that concerning the Ddat Poss H construction
as a whole, and that concerning the relationships between the range of expressive strategies
German varieties exhibits and the range of conceptual-semantic differentiations that need to
be expressed. The inclusion of data from Standard German and very different non-standard
German varieties (e.g., Low German vs. Walser German) forces one to the conclusion that
there are no simple causal relationships to be found. In fact, Walser German itself illustrates
this: While the preservation of the genitive and its functions could be argued to “block” the
frequent use of the adnominal possessive dative, the highly frequent use of the von periphrasis
in originally genitival functions points to a peaceful coexistence.
Primary sources
2. Merseburger Zauberspruch = Eichner, Heiner/Nedoma, Robert (2001): Die Merseburger
Zaubersprüche. Philosophische und sprachwissenschaftliche Probleme aus heutiger Sicht. Die
Sprache, 42, 1–196.
Dietrichs Flucht = Martin, Ernst (ed.) (1866): Deutsches Heldenbuch. Zweiter Teil. Alpharts
Tod, Dietrichs Flucht, Rabenschlacht. Berlin: Weidmannsche Buchhandlung.
Isidor = Eggers, Hans (ed.) (1964): Der althochdeutsche Isidor. Nach der Pariser Handschrift
und den Mondseer Fragmenten neu herausgegeben. Tübingen: Niemeyer.
Nibelungenlied = Batts, Michael S. (ed.) (1971): Das Nibelungenlied. Paralleldruck der
Handschriften A, B und C nebst Lesarten der übrigen Handschriften. Tübingen: Niemeyer.
Notker, Consolatio = Tax, Petrus (ed.) (1988): Notker der Deutsche: Boethius, De
consolatione Philosophiae, Buch I-II-III-IV/V. Tübingen: Niemeyer.
Notker, Psalter = Tax, Petrus (ed.) (1979): Notker der Deutsche: Der Psalter, Psalm 1–50.
Tübingen: Niemeyer.
Otfrid = Erdmann, Oskar (ed.) (61973): Otfrids Evangelienbuch. Tübingen: Niemeyer.
Notker, Psalmen, p. 172, l. 5
Parzival = Leitzmann, Albert (ed.) (1959–1961): [51959]: Wolfram von Eschenbach. 2.
Parzival: Buch VII bis XI. [51960]: Wolfram von Eschenbach. 3. Parzival: Buch XII bis XVI.
[71961]: Wolfram von Eschenbach. 1. Parzival: Buch I bis VI. Tübingen: Niemeyer.
deposit this is’) and later all genitival relations, not only relative constructions, were replaced by constructions
like mîn vader sîn hûs (my.ACC father.ACC his house; my father’s house). Such a course of events differs
considerably from those for Middle and High German and if it took place in that way, it becomes probable that
what we call “adnominal possessive dative” in German is not as uniform as it seems and might have different
historical origins and motivations for different areas, at least with respect to Low German vs. High/Middle
German varieties.
33
Willehalm = Lachmann, Karl (ed.) (1891): Wolfram von Eschenbach. Berlin: G. Reimer.
References
Bart, Gabriela (2006): Ds Grossvatersch Brilla oder di Brilla vam Grossvater. Zu den
Possessivkonstruktionen im Schweizerdeutschen. Unpublished Thesis, University of Zurich.
Behaghel, Otto (1923): Deutsche Syntax. Vol. 1: Die Wortklassen und Wortformen. A.
Nomen. Pronomen. Heidelberg: Carl Winter’s Universitätsbuchhandlung.
Behaghel, Otto (1924): Deutsche Syntax. Vol. 2: Die Wortklassen und Wortformen. B.
Adverbium. C. Verbum. Heidelberg: Carl Winter’s Universitätsbuchhandlung.
Behaghel, Otto (1932): Deutsche Syntax. Vol. 4: Wortstellung. Periodenbau. Heidelberg: Carl
Winter’s Universitätsbuchhandlung.
Bohnenberger, Karl (1913): Die Mundart der deutschen Walliser im Heimattal und in den
Außenorten. Frauenfeld: Huber & Co.
Bornkessel-Schlesewsky, Ina/Schlesewsky, Matthias (2009): The Role of Prominence
Information in the Real-Time Comprehension of Transitive Constructions: A Cross-Linguistic
Approach. Language and Linguistics Compass, 3/1, 19–58.
Carr, Charles T. (1933): The Position of the Genitive in German. The Modern Language
Review, 28/4, 465–479.
Chappell, Hilary/McGregor, William (eds) (1995): The grammar of inalienability. A
typological perspective on body part terms and the part-whole relation. Berlin/New York: de
Gruyter.
Comrie, Bernard (1989): Language universals and language typology. Oxford: Blackwell.
Demske, Ulrike (2001): Merkmale und Relationen. Diachrone Studien zur Nominalphrase im
Deutschen. Berlin/New York: de Gruyter.
Ebert, Robert Peter (1986): Deutsche Syntax 1300–1750. Bern et al.: Peter Lang.
Eroms, Werner (2000): Syntax der deutsche Sprache. Berlin/New York: de Gruyter.
Fleischer, Jürg/Kasper, Simon/Lenz, Alexandra (2012): Die Erhebung syntaktischer
Phänomene durch die indirekte Methode: Ergebnisse und Erfahrungen aus dem
Forschungsprojekt „Syntax hessischer Dialekte“ (SyHD). Zeitschrift für Dialektologie und
Linguistik, 79/1, 1–42.
Fleischer, Jürg/Schallert, Oliver (2011): Historische Syntax des Deutschen. Eine Einführung.
Tübingen: Narr
Heine, Bernd (1997): Possession. Cognitive sources, forces, and grammaticalization.
Cambridge: Cambridge University Press.
Henn-Memmesheimer, Beate (1986): Nonstandardmuster. Ihre Beschreibung in der Syntax
und das Problem ihrer Arealität. Tübingen: Niemeyer.
Henzen, Walter (1932): Der Genitiv im heutigen Wallis. Beiträge zur Geschichte der
deutschen Sprache und Literatur, 56, 91–138.
Holisky, D. A. (1987): The case of the intransitive subject in Tsova-Tush (Batsbi). Lingua,
71, 103–32 .
34
Kasper, Simon (2013): Grounding the Linking Competence in Culture and Nature. How
Action and Perception Shape the Syntax-Semantics Relationship. University of Marburg
Dissertation.
Kiefer, Heinrich (1910): Der Ersatz des adnominalen Genitivs im Deutschen. University of
Giessen Dissertation. Leipzig: Hoffmann.
Kluge, Friedrich (252011): Etymologisches Wörterbuch der deutschen Sprache. Ed. by E.
Seebold. Berlin/New York: de Gruyter.
Koptjevskaja-Tamm, Maria (2003): Possessive noun phrases in the languages of Europe. In:
Frans Plank (ed.): Noun Phrase Structure in the Languages of Europe, 621–722. Berlin:
Mouton de Gruyter.
Koß, Gerhard (1983): Realisierung von Kasusrelationen in den deutschen Dialekten. In:
Besch, W./Knoop, U./Putschke, W./Wiegand, H. E. (eds): Dialektologie: Ein Handbuch zur
deutschen und allgemeinen Dialektforschung. Vol. 2, 1242–1250.. Berlin/New York: de
Gruyter.
Kuno, Susumu/Kaburaki, Etsuko (1977): Empathy and syntax. Linguistic Inquiry, 8/4, 627–
672.
Lasch, Agathe (1914): Mittelniederdeutsche Grammatik. Tübingen: Niemeyer.
Lehmann, Christian (1998): Possession in Yucatec Maya. Structures – functions – typology.
München: Lincom Europa.
Lindow, W./Möhn, D./Niebaum, H./Stellmacher, D./Taubken, H./Wirrer, J. (1998):
Niederdeutsche Grammatik. Leer: Schuster.
Lübben, August (1882): Mittelniederdeutsche Grammatik. Nebst Chrestomathie und Glossar.
Leipzig: Weigel.
Mironow, Sergeij A. (1957): Zur vergleichenden Formenlehre der deutschen Mundarten.
Beiträge zur Geschichte der deutschen Sprache und Literatur (Halle), 79 (Special issue), 388–
414.
Oubouzar, Erika (1992): Zur Ausbildung des bestimmten Artikels im Ahd. In.: Desportes, Y.
(ed.): Althochdeutsch: Syntax und Semantik: Akten des Lyonner Kolloquiums zur Syntax und
Semantik des Althochdeutschen. Akten des Lyonner Kolloquiums zur Syntax und Semantik
des Althochdeutschen, March 1–3, 1990, 69–87. Lyon: Université Lyon III Jean Moulin.
Prell, Heinz-Peter (2000): Die Stellung des attributiven Genitivs im Mittelhochdeutschen. Zur
Notwendigkeit einer Syntax mittelhochdeutscher Prosa. Beiträge zur Geschichte der
deutschen Sprache und Literatur, 122, 23–39.
Primus, Beatrice (1999): Cases and Thematic Roles. Tübingen: Niemeyer.
Schmidt, Jürgen Erich/Herrgen, Joachim (2011): Sprachdynamik. Berlin: Erich Schmidt
Verlag.
Schirmunski, Viktor M. (2010): Deutsche Mundartkunde. Vergleichende Laut- und
Formenlehre der deutschen Mundarten. Ed. and comm. by Larissa Naiditsch. Frankfurt a. M.:
Lang.
Seiler, Hansjakob (1983): Possession as an operational dimension in language. Tübingen:
Narr.
Seiler, Hansjakob (2009): Subjectivity and objectivity in the domain of possession. Semiotica,
173/1–4, 417–429.
35
Shrier, Martha (1965): Case systems in German dialects. Language, 41, 420–438.
Stolz, Thomas/Kettler, Sonja/Stroh, Cornelia/Urdze,Aina (2008): Split possession. An areal-
linguistic study of the alienability correlation and related phenomena in the languages of
Europe. Amsterdam/Philadelphia: John Benjamins.
van Valin, Robert D. jr. (2005): Exploring the syntax-semantics interface. Cambridge: MIT
Press.
van Valin, Robert D. jr./LaPolla, Randy J. (1997): Syntax. Structure, meaning, and function.
Cambridge: Cambridge University Press.
van Valin, Robert D. jr./Wilkins, Wendy (1996): Thecase for “effector”. Case roles, agents
and agency revisited. In: Shibatani, M./Thompson, S. (eds): Grammtical constructions, 289–
322. Oxford: Oxford University Press.
Wegener, Heide (1985): Der Dativ im heutigen Deutsch. Tübingen: Narr.
Weise, Otto (1898): Dem Vater sein Haus. Zeitschrift für den deutschen Unterricht, 12, 287–
291.
Weiß, Helmut (2008): The Possessor that Appears Twice. Variation, Structure and function of
Possessive Doubling in German. In: Barbiers, S./Lekakou, M./van der Ham, M./Koeneman,
O. (eds): Microvariation in Syntactic Doubling, 381–401. Bingley: Emerald Group Publishing
Limited.
Weiß, Helmut (2012): The rise of DP-internal possessors. On the relationship of dialectal
synchrony to diachrony. In: De Vogelaer, G./Seiler, G. (eds): The dialect laboratory. Dialects
as a testing ground for theories of language change, 271–293. Amsterdam/Philadelphia: John
Benjamins.
Wilmanns, W. (1909): Deutsche Grammatik. Gotisch, Alt-, Mittel- und Neuhochdeutsch.
Dritte Abteilung: Flexion. 2. Hälfte: Nomen und Pronomen. Strassburg: Karl J. Trübner.
Wipf, Elisa (1910): Die Mundart von Visperterminen im Wallis. University of Zurich
Dissertation. Frauenfeld: Huber & Co.
Zifonun, Gisela (2003): Dem Vater sein Hut. Der Charme des Substandards und wie wir ihm
gerecht werden. Deutsche Sprache, 1/2003, 97–126.
Online sources:
www.syhd.info
