DE GRUYTER MOUTON
Zeitschrift für Sprachwissenschaft 2015; 34(2): 213–246
Svetlana Petrova*
Synchronic variation and diachronic
change in the expression of indefinite
reference: evidence from historical German
DOI 10.1515/zfs-2015-0011
Abstract: The present article discusses the semantic and discourse-pragmatic
properties of different competing types of indefinite noun phrases in Old High
German, the earliest attested period of German. In particular, it investigates
the behavior of indefinites marked by sum, ein and various interrogative-based
determiners, with respect to properties considered constitutive of specific indefinites from a theoretic and cross-linguistic perspective. Upon analyzing newly
retrieved corpus data, the paper shows that already at the beginning of the
attestation, all marked types of indefinites in historical German violate basic
conditions of specificity, understood in terms of any of the relevant notions
distinguished in the literature. This result rejects previous scenarios according
to which marked indefinites in historical German are correlates of specific reference and challenges the explanation of the diachronic development of ein from
a numeral towards an indefinite determiner via an assumed separate, intermediate stage during which it assigns referential-specific interpretation to the
noun phrase.
Keywords: indefinite noun phrases, indefinite article, specificity, Old and Middle High German
1 Introduction
Old High German (henceforth OHG), the earliest recorded period of German (c.
800–1050), displays variation between different types of indefinite noun phrases which is represented in (1a)–(1d). As a rule, such expressions are realized as
bare nouns (1a), but in addition, phrases modified by various markers, traditionally classified as indefinite pronouns (Braune 2004: 253), also occur in the
data. One of these is sum (1b), also attested in its compound form sumilîh, the
*Corresponding Author: Svetlana Petrova, Universität Wuppertal,
E-Mail: [email protected]
Unauthenticated
Download Date | 6/14/17 4:02 PM
214
Svetlana Petrova
DE GRUYTER MOUTON
cognate of the present-day English indefinite pronoun some, which is missing
in present-day German, with the notable exception of some Alemannic dialects
(Elvira Glaser, p.c.). Another one is ein (1c), which is etymologically related to
the numeral ein ‘one’ and a cognate of the modern German indefinite article
ein. A third type is represented by the interrogatives uuer ‘who’, uuaz ‘what’
and (h)uuelih ‘of what kind’ and various compounds thereof, such as giuuelih,
ethesuuer, ethesuuaz, etheslîh, sihuuer, sihuuelîh, etc., (1d):1
(1) a.
man uuas hiuuiskes
fater
man was family.gen.sg father
‘There was a certain householder’
homo erat pater familias
(T 203, 6; around 830 AD)2
b.
uuas sum man dar
was indef man there
‘A certain man lived there’
erat autem quidam homo ibi
(T 135, 6; around 830 AD)
c.
gieng zi imo ein
centenary
went to Him indef centurion
‘A centurion went to Him’
accessit ad eum centurio
(T 83, 9; around 830 AD)
1 The ability of interrogatives to function as indefinites as well is a common Indo-European
phenomenon (Behaghel 1923: § 239), also present in modern German (see Gallmann 1997: Ch.
5, and the literature therein). It is unclear how the compound forms emerged and how they
acquired indefinite meaning. The etymology of the first component is not completely transparent either. Forms with gi- are considered West Germanic substitutes of Gothic enclitic -uh,
equivalent to Lat. -que ‘and, also’, see Behaghel (1923: 387). OHG ethe(s) might be related to
the disjunctive connective OHG eththo ‘or’, Gothic aiþþau, see Braune (2004: § 167, note 10).
Finally, the origin of sihuuer is disputed. Behaghel (1923: 393) refers the previous proposals
trying to derive this form from the generalizing relative sohuuer ‘whoever’, or from the imperative of sih to sehan ‘see’.
2 The spelling of the examples is as in the editions listed in the references. “&” in Tatian
represents “et”, a slash “/” represents a line break. Spacing within words in MF is as in the
respective edition.
Unauthenticated
Download Date | 6/14/17 4:02 PM
DE GRUYTER MOUTON
d.
Variation and change of indefinite expressions
215
zuene sculdigon uuarun / sih uuelihemo Inlihere
two debtors were
indef.dat.sg creditor
‘a creditor had two debtors’
duo debitores erant / cuidam foeneratori
(T 238, 19–20; around 830 AD)
This variation has received attention in the traditional literature (Behaghel 1917,
Behaghel 1923: 31–138 and 360–407; Erben 1950; Braune 2004: 252–256; Oubouzar 2000; Szczepaniak 2009) but has remained unexplored from a more formal,
theoretic point of view.3 At the same time, similar variation between different
types of indefinite expressions, incl. different types of indefinite determiners,
has been presented in typological work since the 1970s4 and has advanced to
one of the central issues in the theoretical treatment of indefinites.
Very generally, the cross-linguistic variation in the domain of indefinite
reference manifests itself in the competition between one canonical and one or
more non-canonical types of indefinite expressions in the system of one and
the same language. What unifies the treatment of the special, or non-canonical
forms is their relatedness to specificity, i.e. to the notion that the speaker has a
particular referent in mind when using a type of expression, in contrast to
canonical types of indefinites which are ambiguous, or indifferent regarding
the existence of a referent or the speaker’s intention to predicate about that
referent.
Looking at the traditional treatment of indefinites in the history of German,
we discover major parallels between the functions assumed to apply to the
marked types of expressions, and the properties considered constitutive of specific indefinites in the linguistic discussion. The present article will argue, however, that the semantic contribution of the marked indefinite expressions in
OHG cannot be restricted to specificity. Rather, these expressions are used to
cover a broad range of meanings already at the earliest phase of their attestation, when they still occur at a very low rate, compared to canonical bare
nouns.
The article is structured as follows. Section 2 presents the corpus of the
study and deals with some empirical issues related to the native status, the
frequency and the distribution of different types of indefinites in OHG. Section 3
3 An exception is work by Jäger (2007, 2010) whose main interest is the type and presence of
indefinites in negative sentences. This work will be referred to in Section 3.2.1.
4 See Chung and Ladusaw (2004: Ch. 2, and the references therein), Givón (1973, 1981), Ionin
(2006, 2013), Jaggar (1985, 1988), Matthewson (1998), Wright and Givón (1987), von Heusinger
and Klein (2013), among others.
Unauthenticated
Download Date | 6/14/17 4:02 PM
216
Svetlana Petrova
DE GRUYTER MOUTON
introduces specificity as a notion associated with non-canonical indefinites
across languages. This section also presents the basic semantic, syntactic and
discourse-pragmatic properties of specific indefinites and tests how different
types of expressions in OHG behave with respect to these properties. Section 4
summarizes the results and discusses some implications of the study for describing the diachronic development of the indefinite article in German.
2 Corpus
The present investigation is based on a complete sample of marked indefinite
expressions found in texts from the entire OHG period, augmented by a sample
of indefinite bare nouns retrieved from these texts. The major documents of
the OHG tradition, Isidor, Tatian, Otfrid, which provide the basis of previous
investigations, were re-examined in complete length. In addition, texts which
were disregarded in the previous research were taken into consideration as
well, e.g. the documents included in Monseer Fragmente as well as the socalled Minor OHG documents, a collection of texts differing in genre and time
of composition and compiled in Steinmeyer (1971 [1916]).
Marked indefinites were retrieved electronically, by using the available resources.5 Bare indefinites are not searchable electronically, so they were collected manually. I collected a full sample of bare indefinites from the texts that
yielded examples of marked expressions. For Tatian and Otfrid’s Gospel Book,
the retrieval of bare nouns was restricted to selected parts only, due to the large
size of these documents.6
Table 1 lists the texts which yielded marked indefinite expressions in chronological order, grouped together into smaller successive sub-periods of around
30 years of duration each. Table 1 also provides the total number of words in
5 I searched through the OHG texts included in the TITUS database (Thesaurus Indogermanischer Text- und Sprachmaterialien), http://titus.fkidg1.uni-frankfurt.de, by using the word
searching tool. Additionally, I searched through the texts included in Referenzkorpus Altdeutsch, a newly launched annotated corpus of OHG and Old Saxon compiled by the Research
Project DDD – Referenzkorpus Altdeutsch (750–1050), http://www.deutschdiachrondigital.
de/. The corpus is searchable via ANNIS, http://annis-tools.org/.
6 For Tatian, I examined pp. 25–95 and pp. 200–230 of the manuscript edited by Masser
(1994). From Otfrid’s Gospel Book, I selected all bare indefinites found in Book 1 and 2, including the prefaces (Ad Ludowicum and Ad Salomonem). The percentages for Tatian and Otfrid’s
Gospel book in Table 1 are calculated for these text sections. The total of marked indefinites
in these sections is given before the slash. The total number of instances of marked indefinites
in the entire documents is given after the slash.
Unauthenticated
Download Date | 6/14/17 4:02 PM
Phase
HM
I
MF
Phase
T
Phase
O
Phase
L
G
Ch
Phase
APh
DD
MM
HiH
BBS
APB
I (beginning to 810 AD)
Hamelburger Marktbeschreibung (777 AD)
Pariser Codex (around 800 AD)
Monseer Fragmente (around 810 AD)
II (810 AD to 840 AD)
Tatian (around 830 AD)
III (840 AD to 870 AD)
Otfrid (around 870 AD)
IV (870 AD to 910 AD)
Ludwigslied (882 AD)
Georgslied (896 AD)
Christius und die Samariterin (908 AD)
V (after 910 AD)
Der Ältere Physiologus (11th cent.)
De Definitone (11th cent.)
Memento Mori (11th/12th cent.)
Himmel und Hölle (12th cent.)
Bamberger Blutsegen (12th cent.)
Ahd. Predigtsammlung B (12th/13th cent.)
Text and time of composition
Table 1: Corpus.
2
27
40
153
177
7
7
6
31
1
13
5
1
3
54.109
79.341
392
471
233
1.912
120
802
616
93
1.730
total
100
5.892
8.795
tokens
–
2 (15 %)
3 (60 %)
(50 %)
2 (67 %)
9 (29 %)
6 (86 %)
3 (43 %)
4 (67 %)
97 (76 %)
115 (87 %)
–
26 (96 %)
36 (90 %)
bare nouns
–
–
–
–
–
–
–
–
–
–
10/22 (8 %)
–
1 (4 %)
4 (10 %)
sum N
1/6 (1 %)
–
–
–
uuelih N
22
1
11
2
1
1
(71 %)
(100 %)
(85 %)
(40 %)
(50 %)
(33 %)
1 (14 %)
4 (57 %)
2 (33 %)
–
–
–
–
–
–
–
–
–
30/80 (24 %) –
5/10 (4 %)
2 (100 %)
–
–
ein N
DE GRUYTER MOUTON
Variation and change of indefinite expressions
Unauthenticated
Download Date | 6/14/17 4:02 PM
217
218
DE GRUYTER MOUTON
Svetlana Petrova
each document, to give a rough orientation regarding the size of the respective
text. Most importantly, it contains the sum of indefinite noun phrases of all
types, incl. the proportion of each individual type per text (section).
Table 1 reveals some important points regarding the frequency and distribution of marked indefinites in OHG. First, it shows that these expressions are
infrequent during a long period of time. Note that a frequency of 100 % in
Hamelburger Marktbeschreibung, the oldest document in the corpus, should be
treated with caution, given the small size of this document. It cannot be taken
to indicate an overwhelming preference for modified indefinites already by the
end of the eighth century, because a series of texts afterwards still shows a
comparatively low percentage of marked indefinites.
Second, note that the different markers of indefiniteness do not co-occur
within one and the same text, with the notable exception of Tatian, which is a
unique representative of a typologically rare but theoretically intriguing scenario involving a multiple distinction in the formal expression of indefiniteness
(see von Heusinger and Klein 2013 for an overview of languages displaying
similar systems).
Third, we observe that sum occurs as an adnominal marker of indefiniteness only during the earliest attestation. Already in Otfrid’s work (around 870
AD), it is restricted to its independent use as in (2). By contrast, ein as an
adnominal marker of indefiniteness is present during the entire attestation. In
early OHG,7 it is predominantly used to express cardinality (3a) and related
functions, e.g. the uniqueness or the exceptionality of a referent (3b)–(3c).8
This type of examples was not included in the corpus of the present investigation because ein was not treated as an indefinite marker in these cases. But the
hitherto disregarded text of Hamelburger Marktbeschreibung yields the earliest
conclusive examples in which ein clearly departs from its cardinal function and
acts as an adnominal marker of indefiniteness (3d), which is conclusive for the
purpose of the investigation:
(2) Sum quad er dáti
widar gót
indef said he did.3sg.pret.subj against God
‘someone claimed that he had acted against God’
(O III 20, 61; around 870 AD)
7 Similar observations on one in Old English are reported by Rissanen (1997).
8 In examples of the type of (3c), ein has a function similar to that of the exclusive focussensitive operator only, also indicated by the Latin equivalent solus ‘a single one’.
Unauthenticated
Download Date | 6/14/17 4:02 PM
DE GRUYTER MOUTON
(3) a.
Variation and change of indefinite expressions
219
context: a lord pays different sums of money to his servants (Mt 25,24)
therde eina / talenta intfieng
rel.prt one
talent received
‘he who had received one talent’
qui unum / talentum acceperat
(T 262, 22–23; around 830 AD)
b.
in dhem
dhrim
heidim
scal
man ziuuaare
in the.dat.pl three.dat.pl entities.dat.pl should indef truly
eina gotnissa beodan
one divinity signify
‘by these three entities one should signify one single divinity’
in eis personis una diuinitas praedicanda est
(I 21, 10; around 800 AD)
c.
In themo
einen
bróte
nileb&
ther man
in the.dat.sg one.dat.sg bread.dat.sg neg.lives the man
non In solo pane uiuit homo
‘Man shall not live by bread alone’
(T 50, 1; around 830 AD)
d.
inde
in ein
sol
thence in indef wallow
‘thence to a wallow’
(HM 21; 777 AD)
Given the large number of texts that are translations from Latin, the role of
potential lexical correspondences and the native status of the indefinites in
OHG should be considered.
Latin has an elaborate system of indefinite markers which Haspelmath
(1997) classifies in the way given in (4a)–(4e):
(4) a. the -dam series (e.g. quidam), expressing specific-known reference
b. the ali-series (aliquis), expressing specific-unknown and irrealis nonspecific reference9
c. the -quam series, used in questions, indirect negation, comparatives
and conditionals
d. the n-series, used in direct negation
9 But note that aliquis is ambiguous, e.g. it may also be used as a simple existential some,
see Gianollo (2013) for a discussions and references.
Unauthenticated
Download Date | 6/14/17 4:02 PM
220
Svetlana Petrova
e.
DE GRUYTER MOUTON
the series -vis/-libet and -cumque as well as the reduplicated form quisquis, representing free choice items.
The role of the original in the selection of indefinite markers in OHG has raised
some discussion in the literature. Grimm (1837), e.g., claims that sum is driven
by Latin influence because it is the proper lexical equivalent of Lat. quidam
‘certain’, while ein is the genuine OHG marker of indefiniteness, occurring in
native style.10 Consider (5a)–(5b) from the Tatian translation. In (5a), where the
vernacular uses sum, the Latin original contains the indefinite pronoun quidam,
while in (5b), where the original provides no explicit marker of indefiniteness,
the OHG translation opts for ein:
(5) a.
b.
sum uúib
martha ginemnit
indef woman Martha called
‘a certain woman called Martha’
mulier quędam martha nomine
(T 99, 20; around 830 AD)
Inti slíumo findet ir / eina esilin gibuntana
and soon find you indef ass
tied.acc
‘and straight you shall find an ass tied’
& statim Inueni&is / asinam alligatam
(T 189, 2–3; around 830 AD)
But there is counterevidence suggesting that OHG sum is native, rather than
triggered by Lat. quidam. Oubouzar (2000) refers to a single example, given in
(6a), in which OHG sum N represents a Latin bare noun (see also Fobbe 2004:
107). The corpus of the present investigation confirms this picture for Tatian
and yields additional examples from other translations, see (6b):
(6) a.
sum man hab&a zuene suni
indef man had
two sons
‘A man had two sons’
homo habebat duos filios
(T 202, 15; around 830 AD)
10 Grimm (1837) suggests that ein functions as an indefinite article already in OHG times. But
note that next to its low frequency, historical ein displays some properties that are untypical
of the modern German indefinite article. E.g., it is used with abstract and mass nouns, it
displays plural forms (also contrary to the numeral ein) and gives rise to a peculiar definite
use, i.e. it refers to antecedents in the previous discourse. For examples, discussions and
references, see Donhauser (1995), Donhauser and Petrova (2012).
Unauthenticated
Download Date | 6/14/17 4:02 PM
DE GRUYTER MOUTON
b.
Variation and change of indefinite expressions
221
uuirdit
imo
gataan sum pina
pass.aux.3sg.pres him.dat done indef sorrow
‘if tribulation arises’
facta autem tribulatione et persecution
(MF 9, 14; around 810 AD)
Similarly, the selection of the interrogative-based forms should be treated as
independent of the corresponding lexical pattern in the original. These forms
render Latin indefinites of various series, and in addition, they are selected to
express Latin bare nouns, see (7):
(7) furliez in
ther diuual / zi sihuueliheru zîti
left
him.acc det devil
to indef
time
‘the devil departed from him for a while’
reliquit eum diabolus / usque ad tempus
(T 50, 27–28; around 830 AD)
Finally, the lexical correspondence between OHG ein and Lat. unus ‘one’ outside its cardinal interpretation should be considered. E.g., in (8), unus refers to
the existence of a referent, but not necessarily to its uniqueness. But note that
there are examples in which indefinite ein N translates a Latin bare noun, as
in (5b) above.
(8) gisah einan
figboum
saw indef.acc.sg fig tree
‘[he] saw a fig tree’
uidens fici arborem unam
(T 199, 23; around 830 AD)
These observations suggest that the various types of indefinite markers in OHG
are part of the genuine system of the language of the respective period and
that their use and distribution follow the rules of native grammar rather than
those of Latin style.
3 Properties of the marked indefinites in OHG
3.1 The notion of specificity
A central property underlying the treatment of the special, or non-canonical
types of indefinites across languages is their relatedness to specificity, i.e. to
Unauthenticated
Download Date | 6/14/17 4:02 PM
222
Svetlana Petrova
DE GRUYTER MOUTON
the notion that the speaker has a particular referent in mind when using a
special type of expression, in contrast to existential indefinites which assert the
presence of some referent, or to generic indefinites which refer to classes of
individuals. In some languages, specific reference is associated with special
marking, e.g. particular, certain or specific in English, bestimmt and gewiss in
German, special morphemes or grammaticalized articles (see von Heusinger
2011 for an overview). At the same time, in German and English, indefinites
without modifying adjectives of the type particular, bestimmt, etc., are ambiguous regarding the existence of a particular referent in the context. Consider the
interpretation of the indefinite noun phrase a book about butterflies in (9a)–
(9b), adapted from Ionin (2010: 450, 451):
(9) Mary wants to read a book about butterflies.
a. Mary wants to read a particular/special book about butterflies
b. Mary wants to read some book about butterflies, no matter which
On the one hand, a book about butterflies may establish reference to a specific
book about butterflies which Mary intends to read. In this case, the indefinite
expression is compatible with adjectives triggering specific reference, such as
particular or specific in English (9a). But on the other hand, the same indefinite
expression may refer to any book about butterflies, made explicit by a continuation of the type no matter which in (9b). In terms of Givón (1981: 36), in the first
case, the indefinite article acts as a referential-indefinite marker establishing
reference to an individual of a type, whose exact identity matters in the context,
while in the second one, the noun phrase denotes type membership, and the
identity of the referent is incidental to the communication.
In the literature, the ability of certain classes of indefinites to establish
reference to individuals that the speaker has in mind has been related to a
series of semantic, syntactic and discourse-pragmatic properties. Some of the
most relevant and widely accepted ones are summarized in (10a)–(10d):
(10) a.
b.
c.
d.
wide scope of the indefinite expression in opaque environments
referentiality and presupposition of existence of a referent
speaker’s familiarity with the referent of the indefinite expression
implication of relevance, or discourse prominence of the referent.
It has been noticed that these features manifest themselves in peculiarly different ways across languages, making specificity a notoriously difficult notion,
which applies to a heterogeneous class of phenomena (von Heusinger 2011).
Unauthenticated
Download Date | 6/14/17 4:02 PM
DE GRUYTER MOUTON
Variation and change of indefinite expressions
223
In the next subsections, I will survey these features of specific indefinites,
placing special emphasis on the methods and diagnostics used to detect specificity in theoretic and cross-linguistic terms, and then will examine how various
types of indefinite expressions in OHG behave with respect to these features.
3.2 Properties of specific indefinites and the behavior of
OHG indefinites
3.2.1 Scopal properties
With Fodor and Sag (1982) and much subsequent literature, the ambiguity in
the referential properties of indefinites of the type a(n) N correlates with a
difference in the scopal behavior of such indefinites in opaque environments.
Consider (11), where the indefinite expression occurs in the presence of a quantifier:
(11) Every woman talked to a child in fifth grade.
The indefinite a child in fifth grade receives two11 different interpretations. According to the first one, a child in fifth grade refers to different individuals
which co-vary with the referents to which the quantificational expression every
woman applies. In this case, the indefinite is within the scope of the quantifier
and displays its so-called quantificational interpretation. Alternatively, a child
in fifth grade can refer to a unique individual such that every woman talked to
this individual. In this case, the indefinite takes scope over the quantifier and
displays its so-called referential interpretation.
It has been observed that the same differences in the scopal behavior of
indefinites also arise in the presence of negative elements (12), intensional operators like believe (13), or modal expressions like want (14). In the a-paraphrases, the indefinite remains within the scope of the respective operator, in the
way shown for the quantificational interpretation of indefinites in (11) above.
In contrast, indefinites bearing reference to a unique individual exhibit wide
scope, i.e. they are outside the scope of the respective operator, represented in
the b-paraphrases of the examples.
11 I omit cases of intermediate scope which are discussed in Ionin (2010).
Unauthenticated
Download Date | 6/14/17 4:02 PM
224
Svetlana Petrova
DE GRUYTER MOUTON
(12) Jack didn’t see a misprint.
a. it is not the case that Jack saw any misprint
b. there is a particular misprint, such that Jack did not see that misprint
(13) Paula believed that Bill talked to an important politician.
a. Paula believed that Bill talked to some important politician, no matter
whom
b. there is a particular important politician, such that Paula believed that
Bill talked to that politician
(14) Mary wants to marry a Norwegian.
a. Mary wants to marry a Norwegian, no matter whom
b. there is a particular Norwegian, such that Mary wants to marry this
Norwegian
The scopal behavior of the different types of indefinites in historical German
has not been addressed in the previous literature because the traditional research was unaware of this diagnostics on the behavior of indefinites in general. Note, however, that the literature cites some conclusive examples, such as
(15) from the Middle High German period (MHG, c. 1050–1350), in which an
indefinite occurs in the presence of a modal operator:
(15) ich muoz einer
frowen
rûmen diu lant
I may indef.dat.sg lady.dat.sg praise the land.pl
‘I have reason to praise the lands of a lady’
(von Kraus 1930: 3)
The only plausible interpretation is that there is a particular lady such that the
lyrical person has reason to praise the lands of this lady. In other words, the
indefinite takes scope over the modal operator. An alternative interpretation
according to which the indefinite in (15) refers to some lady whose lands should
be praised, is excluded given our knowledge of the availability of a unique
individual acting as the addressee of minnesong poetry. By virtue of examples
like (15), one might be inclined to conclude that at least ein N in historical
German takes scope over modal operators, i.e. shares the scopal behavior of
specific indefinites in opaque environments. This conclusion, however, is precipitate and definitely wrong. I checked the OHG corpus for the diagnostics
allowing insights into the scopal behavior of the various types of indefinites
included therein. I found no indefinites in the presence of quantificational ex-
Unauthenticated
Download Date | 6/14/17 4:02 PM
DE GRUYTER MOUTON
Variation and change of indefinite expressions
225
pressions of the type every N.12 With negative operators, as Oubouzar (2000)
and Fobbe (2004) also observe, only bare nouns are attested,13 see (16) for
illustration:
(16) uuir nihabemes cuning
we NEG.have king
‘we have no king’
non habemus regem (Io 19,15)
(T 310, 8; around 830 AD)
But luckily, there is data for marked indefinites in the presence of modal operators which supports a conclusive interpretation. Consider (17a)–(17c). The first
two sentences represent different translations of Mt 12,38. In both cases, the
indefinites refer to any sign that Jesus might perform, rather than to some
specific miracle that the speakers desire to see. Given this context, the indefi12 There is evidence for indefinites in relative clauses whose antecedent is a quantifier, e.g.
iogiuuelih ‘everyone’ (i), or relatives of the type anyone who, who(so)ever:
(i) iogiuuelih thiethar gisihit uúib / sie
zigeronne
every
rel.ptc looks woman her.acc to.lust.dat
‘Whosoever looks on a woman to lust after her’
omnis qui uiderit mulierem / ad concupiscendum eam
(T 63, 21–22; around 830 AD)
In all cases of the type, the indefinites are bare nouns, displaying narrow scope interpretation.
E.g., in (i), it is evident that adultery is meant as resulting from lusting after any woman, not
after a particular one.
13 Rissanen (1988) similarly remarks that Old English sum is restricted to affirmative contexts.
Jäger (2007: 149), who investigates indefinites under negation in OHG, accounts for a high
number of bare nouns in these contexts as well but mentions cases of ein N. I ignore these
cases because ein indicates cardinality or exclusion (see [3c] above) rather than indefiniteness,
as also suggested by the translation of ex. (16) in Jäger (2007: 148). See also (i), in which ein
N translates Lat. nullus ‘not a single one’ and co-occurs with the adjective einich ‘single’:
(i) so selp so im
noh
ein tempel ni bileiph noh einich altari noh einich offerunc
as
them neither one temple neg left
nor single altar nor single sacrificial
ghelstar, so sama ni bileiph
im
einich chuninc noh einich sacerdos
idol
as
neg remained them single king
nor single priest
‘they were left without a single king or priest in the same way in which they remained
without a single temple or altar or idol’
iam sicut nullum templum nullum altare nullum sacrificium ita nullus rex nullus sacerdos
remansit
(I 35, 17; around 800 AD)
Jäger (2007, 2010: 805) also accounts for NPIs in negative contexts such as einig and dehein(ig)
as well as for the negative indefinite nihein(ig) in negative sentences. These cases, however,
remain beyond the scope of my investigation.
Unauthenticated
Download Date | 6/14/17 4:02 PM
226
DE GRUYTER MOUTON
Svetlana Petrova
nites are clearly within the scope of the modal operator. The same applies for
(17c). Here, the interrogative-based indefinite uuelih, used for Lat. aliquid ‘some’
as a simple existential, refers to any sign that Pilate expects to see from Jesus,
and not to a particular one:
(17) a.
uuir uuollen / fon thir
zeichan
gisehan
we want.1pl from you.dat.sg sign.acc.sg see.inf
‘we would see a sign from thee’
uolumus / a te signum uidere
(T 92, 23–24; around 830 AD)
b.
uuellemes fona dir
sum zeihhan
ga seha
want.1pl from you.dat.sg indef sign.acc.sg see.inf
‘we would see a sign from thee’
uolumus a te signum uidere
(MF 6, 27; around 810 AD)
c.
uúanta
sih uuelih zeihan /
gisehan
hoped.3sg refl indef sign.acc.sg see.inf
‘he hoped to see a sign’
sperabat signum aliquod / uidere
(T 307, 21–22; around 830 AD)
Note that (17b)–(17c) are found in texts from the earliest stages of the attestation, i.e. from the first half of the ninth century. By virtue of this data, we have
to conclude that already the earliest instances of the various marked indefinites
in OHG displayed narrow scope in opaque environments, which is a property
untypical for specific indefinites in theoretic and cross-linguistic perspective.
3.2.2 Referentiality and the presupposition of existence of a referent
The ability of specific indefinites to refer to a particular individual that the
speaker has in mind implies that the speaker presupposes the existence of this
individual in the linguistic context. In other words, specificity depends on the
property of the indefinite noun phrase to act as a referring expression in the
sense of Karttunen (1976).
Karttunen (1976) specifies the conditions under which indefinite noun
phrases establish discourse referents. Very generally, these conditions pertain
to the presence of a referent or the truth of a proposition. Root vs. complement
clauses display well-known differences in this respect. Indefinite noun phrases
Unauthenticated
Download Date | 6/14/17 4:02 PM
DE GRUYTER MOUTON
Variation and change of indefinite expressions
227
in root clauses introduce discourse referents if the clause is affirmative, which
results in the possibility to take up the referent by anaphoric expressions in
the subsequent discourse, as demonstrated in (18a). In contrast, sentential negation in root clauses blocks the establishment of discourse referents because
it explicitly denies the existence of the respective entity (18b):14
(18) a. Bill has a cari. Iti /The cari /Bill’s cari is blue.
b. Bill doesn’t have a cari. *Iti /*The cari /*Bill’s cari is blue.
In complement clauses, the referential status of indefinite expressions depends
on the semantic properties of the embedding predicate, i.e. on the issue whether or not the licensing predicate implies the truth of the proposition. This interacts with the polarity of the matrix clause. Factive verbs like know, realize or
regret presuppose the truth of their complements independently of the presence
or absence of sentential negation. Thus, indefinites in such clauses always introduce discourse referents. But implicative verbs like manage, remember, see
or hear imply the truth of their complements only if the licensing predicate is
used in an affirmative context. With negative implicative verbs like forget, fail
or neglect, the reverse relation holds.
Finally, the property to introduce discourse referents does not apply to generic expressions and predicative nouns because they do not introduce individuals but rather denote classes of individuals or assign properties to individuals.
In the philological treatment of indefinites in historical German, we find
some relevant observations alluding to the referential status of the respective
marked types of expressions. Recall the discussion on the absence of marked
indefinites in negative contexts outlined in Section 3.2.1. Desportes (2000) reports that indefinites in negative sentences (19) and generic contexts (20) use
to occur as bare nouns as late as in MHG times. This might be taken to suggest
that noun phrases which fail to refer are resistant to markers of indefiniteness
until late in the history of German, implying that sum and ein are incompatible
with the non-existence of a referent:
14 Some contexts are special in that they do not imply the existence of an individual but
nevertheless allow anaphoric reference. Consider the sentences involving future reference or
modality in (i)–(ii). The existence of a referent cannot be presupposed at the time of the
utterance, but the entities introduced by indefinite noun phrases may obviously be resumed
by anaphors in the subsequent context. Karttunen (1976) assumes that indefinite expressions
in such contexts introduce so-called short term referents, whose anaphoric resumption is
possible as long as the condition of prospectivity or modality is maintained in the context:
(i) You must write a letteri to your parents and send iti right away.
(ii) Mary wants to marry a rich mani . Hei must be a banker.
Unauthenticated
Download Date | 6/14/17 4:02 PM
228
Svetlana Petrova
DE GRUYTER MOUTON
(19) Ez
enwárt
nie
bote
enpfangen […] baz
expl neg.pass.aux.3sg.pret never messenger received
better
‘Never was a messenger received a better way’ (NL 562; 12th/13th cent.)
(20) den
besten,
den
ie
vróuwé gewan
the.acc.sg best.acc.sg rel.acc.sg ever lady
received
‘the best [husband] whom a lady ever received’
(NL 1233; 12th/13th cent.)
The lack of marked indefinites in negative root clauses is confirmed for the
data in my corpus. Similarly, marked indefinites were not found in the relevant
classes of complement clauses.15 But there is another type of evidence suggesting that the use of marked indefinites in OHG does not necessarily presuppose
the existence of a referent. These are hypothetical contexts and yes/no questions, demonstrated in (21a)–(21c).16
(21) a.
b.
uuirdit
imo
gataan sum pina
pass.aux.3sg.pres him.dat done indef sorrow
‘if tribulation arises’
facta autem tribulatione et persecutione
(MF 9, 14; around 810 AD)
Yrhugis
thar thoh éines
man, ther
concider.2sg.ind.pres there then indef.gen.sg man rel
thir
si
irbólgan
you.dat.sg is.pres.subj annoyed
‘if you recall a man who might be angry at you’
(O II 18, 21; around 870 AD)
15 See (i)–(ii) involving bare nouns in the complements of the implicative verbs hear and see
under negation:
(i) wanta er
ni hórta man thaz, / Thaz io fon mágadburti
man
because earlier neg heared indef this
that ever from birth.by.a.virgin man
gibóran wurti
born
pass.aux.3sg.pret.subj
‘because never before has anyone heard about a man borne by a virgin mother’
(O I 17, 16–17; around 870 AD)
(ii) wanta ér
man súlih ni gisáh / Thaz mán io thes githáhti
because earlier indef such neg saw
that man ever this thought.subj
‘because no one has seen that anyone has ever considered [doing] such a thing’
(O III 8, 26–27; around 870 AD)
16 Note that sih uuelicheru in (21c) translates the Latin free choice marker of the -cumque
series.
Unauthenticated
Download Date | 6/14/17 4:02 PM
DE GRUYTER MOUTON
c.
Variation and change of indefinite expressions
229
ist arloubit mann /
zi uorlazzanna
sina quenun / fon
is allowed man.dat.sg to dismiss.inf.dat his wife
for
sih uuelicheru sachu
indef.dat.sg reason
‘is it allowed for a man to dismiss his wife for any reason’
Si lic& homini / dimittere uxorem suam / quacumque ex causa
(T 160, 15–17; around 830 AD)
Note that the examples in (21a)–(21c) illustrate downward entailing environments17 which are among those licensing negative polarity items (NPIs, see
Giannakidou 2011), i.e. elements that are regarded referentially deficient because they are “unable to introduce discourse referents on their own” (Giannakidou 2011: 1662). It can be shown that the OHG indefinites in these examples
share the non-referential status of NPIs. E.g., the indefinite in (21a) refers to
any sorrow that might challenge the faithfulness of Christians. Similarly, ein N
in (21b) refers to any potential adversary that a man might have. Finally, in
(21c) the speaker does not commit himself to the existence of a referent denoted
by the indefinite noun phrase. In sum, all three types of marked indefinites in
OHG may be detected in contexts in which the presence of a particular individual or entity is not presupposed by the speaker.
Let us look at generic indefinites. In the corpus, these take the form of bare
nouns in a great part of the data. E.g., in (22), the indefinite boto refers to
messengers in general, rather than to a specific messenger:
(22) Tho sprach er érlicho […] so bóto
scal
then spoke he respectful as messenger should
‘Then he spoke in a respectful manner, as is expected of messengers’
(O I 5, 13–14; around 870 AD)
But we also find sporadic examples of ein N in generic use already in OHG. The
earliest conclusive examples are provided as early as in Otfrid’s work, i.e.
around 870 AD. Consider (23):
(23) soso éin man sih scal
wérien
as a man refl should protect.inf
‘as a man is supposed to protect himself’
(O IV 17, 13; around 870 AD)
17 Thanks to one anonymous reviewer for drawing my attention to this property.
Unauthenticated
Download Date | 6/14/17 4:02 PM
230
Svetlana Petrova
DE GRUYTER MOUTON
Let us finally examine indefinites in predicative use. In texts from the earlier
phase of the attestation, predicative indefinites take the form of bare nouns
without any exception, as in (24a). But from 890 AD onwards, ein N starts to
be used in predicative function as well, which co-vary with bare nouns until
the end of the OHG period, as illustrated in (24b)–(24e):
(24) a.
bist thu uuîzago
are you prophet
‘are you a prophet’
proph&a es tú
(T 47, 14; around 830 AD)
b.
er quat, Gorio wâri
ein
goukelari
he said George be.3sg.pret.subj indef sorcerer
‘he said that St George was a sorcerer’
(G 5, 3; 896 AD)
c.
daz thu maht forasago sin
that you may prophet be.inf
‘that you may be a prophet’
(Ch 28; 908 AD)
d.
daz ist ein
hélfant
this is indef elephant
‘this is an elephant’
(APh 8, 128, 78; 11th cent.)
e.
Ist siu denne uuarhafto magit
is she then truthfully virgin
‘if she is a virgin indeed’
(APh 3, 126, 39–40; 11th cent.)
From this, we must conclude that against the claims in previous research, all
three markers of indefiniteness in OHG may occur in contexts in which the
respective noun phrase fails to express reference to a single individual present
in the context. In other words, these markers are possible with noun phrases
which lack properties of referring expressions in the sense of Karttunen (1976)
and consequently fail to satisfy a basic condition for specificity.
3.2.3 Identifiability and speaker’s knowledge of the referent
Fodor and Sag (1982) observe that the specific (in their terms referential) interpretation of indefinite noun phrases involves “a certain amount of knowledge
Unauthenticated
Download Date | 6/14/17 4:02 PM
DE GRUYTER MOUTON
Variation and change of indefinite expressions
231
on part of the speaker” (Fodor and Sag 1982: 359). But it has been recognized
that knowledge is distinct from presupposing the existence of a referent. In (25)
from Ionin (2006: 182), a referent exists, but the speaker is unaware of his/her
precise identity:
(25) Sarah wants to talk to a colleague of mine, but I don’t know which one.
Therefore, the literature has introduced a special type of specificity which depends on the condition of knowledge and identifiability, referred to as epistemic
specificity (Farkas 2002).18
Prototypically, the information that is necessary to detect epistemic specificity is provided by the context. E.g., in (26a), where the context suggests that
the speaker is able to identify the referent, the most natural interpretation of
the indefinite is the specific one. By contrast, the speaker’s ignorance of the
referent indicated by the context yields a non-specific interpretation of the indefinite (26b):
A student cheated on the exam. It was the guy who sits in the very
back.
b. A student cheated on the exam. I wonder which student it was.
(26) a.
Additionally, it has been recognized that expressions containing specificity
markers like certain (27) and related expressions are infelicitous in contexts
signaling the speaker’s ignorance or indifference regarding the identity of the
referent (see Ionin 2006 and Ionin 2010 for cross-linguistic parallels and references):
(27) James read a certain book. #I have no idea which book it is.
Let us look how the different types of indefinites in historical German behave
with respect to the conditions of epistemic specificity. The philological literature remarks that the different special types of indefinite expressions in OHG
are used to express familiarity with the referent of the indefinite expression
(von Kraus 1930; Erben 1950). E.g., in medieval poetry of the genre of minne-
18 See also the distinction between specific-known vs. specific-unknown reference in Haspelmath (1997). Von Heusinger (2011: 1032) uses as a diagnostic the possibility to replace the
indefinite noun phrase by an expression with identical reference, e.g. a proper name (substitution of referentially identical expressions).
Unauthenticated
Download Date | 6/14/17 4:02 PM
232
DE GRUYTER MOUTON
Svetlana Petrova
song, ein N is used to refer to the beloved lady, the addressee of the poem, who
exists but whose precise identity must not be disclosed, see (28):
(28) ich bin holt
einer
frouwen
I am devoted indef.dat.sg lady.dat.sg
‘I am devoted to a lady’
(von Kraus 1930: 3)
Judging on the identifiability of a referent and speaker’s knowledge is difficult
for texts from older periods. E.g., in the context of (29), no explicit information
about the referent of the marked indefinite expression in einan gárton ‘to a
garden’ is provided. But we may guess that it was obvious to people acquainted
with the contents of the New Testament that the respective indefinite refers to
the Garden of Gethsemane, where Jesus went with his disciples on the night of
his arrest. This information may be assumed to be part of common knowledge
of the time:
(29) gíang in einan
gárton
went in indef.acc.sg garden
‘[he] went to a garden’
(O IV 16, 1; around 870 AD)
Fobbe (2004) accounts for effects related to the epistemic state of the referent
of various indefinite markers in historical German, observing a noteworthy difference between the three types of indefinites in OHG. She classifies OHG sum
and ein as representatives of the stage of specific-known reference in the implicational map of indefinites developed by Haspelmath (1997), while the interrogative-based forms, according to her, are broader in meaning, taking more abstract functions, e.g. specific-unknown as well as various non-specific functions
(Fobbe 2004: 251 and 264).
But the data in (30a)–(30d) suggests that indifference regarding knowledge
and identity must be extended to all kinds of marked indefinite expressions
present in OHG:
(30) a.
uuirdit
imo
gataan sum pina
pass.aux.3sg.res him.dat done indef sorrow
‘if tribulation arises’
facta autem tribulatione et persecutione (Mt 13,22)
(MF 9, 14; around 810 AD)
Unauthenticated
Download Date | 6/14/17 4:02 PM
DE GRUYTER MOUTON
Variation and change of indefinite expressions
b.
arsteig In einan
murbóum
climbed into indef.acc.sg sycamore.tree
‘climbed up into a sycamore tree’
ascendit In arborem sicomorum
(T 186, 30; around 830 AD)
c.
Só thiu selben
Krístes
kráft eina géislun […]
then this same.gen.sg Christ.gen.sg power indef scourge
gifláht
bound
‘then, the power of the same Jesus made a scourge’
(O II 11, 9; around 870 AD)
d.
Thie fúriston
thaz gihórtun joh ein
giráti datun
the chief.priests this heard
and indef meeting did
‘The chief priests heard about that and held a meeting’
(O III 16, 73; around 870 AD)19
233
It is unnatural to assume the speaker’s familiarity with the referent of the indefinite in these examples. E.g., in (30a), the denotation of the indefinite sum
pina ‘some sorrow’ is open for any kind of distress, consequently, it is inadequate to assume that the speaker intends to refer to one particular kind of
trouble, which is expressed by the indefinite. Consider ein N in (30b). The context is that Zachhaeus, a man little of stature, climbed up into a sycamore tree
in order to be able to see Jesus entering Jericho. Whereas the speaker is convinced that a tree of the respective kind exists in the discourse situation, it is
doubtful if the identity of the tree can be specified more accurately.20 An analo19 An earlier version of this sentence from the Gospel of Matthew contains a bare noun, see
(i):
(i) uuorah tun ga rati
did
meeting
‘[they] held a meeting’
consilium faciebant
(MF 4, 30; around 810 AD)
Note, however, that this fact cannot be taken to suggest that the picture drawn by Fobbe
(2004) was valid for the very early phases of the attestation. Rather, examples in which a
marked indefinite rejects the condition of identifiability of a referent are present already in
texts prior to Otfrid (around 870 AD), see, e.g. (27a)–(27c).
20 Oubouzar (2000) argues that here ein emphasizes the specialty of the referent: “einan
boum drückt nicht nur aus, dass es sich um einen Vertreter der Gattung handelt, sondern,
dass es der spezifische Baum ist, von dem aus Jesus gesehen werden kann” [einan boum not
only denotes a single representative of the kind, but rather refers to the specific tree from
which Jesus can be seen] (Oubouzar 2000: 259; my translation). Szczepaniak (2009: 83) also
Unauthenticated
Download Date | 6/14/17 4:02 PM
234
Svetlana Petrova
DE GRUYTER MOUTON
gous case from Otfrid is presented in (30c). The indefinite eina géislun refers to
the scourge that Jesus made to expel the traders and the changers of money
from the Temple of Jerusalem (Io 2,15). While it is true that a scourge exists, it
is hardly plausible to assume that the speaker is able to specify its precise
identity or to provide special features of this referent. Finally, note that indefinite expressions of the type ein N are attested in idiomatic expressions of the
type ein girati duan ‘to hold a meeting’ (30d), eina klaga duan ‘to lament’, sih
einan duam duon ‘to benefit from performing an act’ or einen ruam duan ‘to do
a favour’ in which the nominal phrases can hardly be assigned reference to a
specific entity that the speaker can identify in any of the cases.
Most importantly, we find marked indefinite expressions in contexts which
explicitly express the speaker’s ignorance of the referent’s identity. E.g., in
(31a), the speaker commits himself to the mysterious, unknown nature of the
respective entity. Similarly, in (31b), the guilt of Barnabas is left unspecified:
(31) a.
b.
In dhrim
fingrum
chiuuisso dher heilego forasago
in three.dat.pl fingers.dat.pl certainly the holy
prophet
dhea dhrifaldun ebanchiliihnissa dhera
almahtigun gotliihhin
the threefold equality
the.gen.pl almighty divinity
mit sumes
chirunes
uuagu
uuac
with indef.gen.sg mystery.gen.sg scale.dat.sg weighed
‘with the three fingers the holy prophet certainly weighted the threefold equality of almighty divinity by use of scales of some mystery’
sub quadam mysterii lance librauit
(I 19, 15; around 800 AD)
thuruh
gistriti / uuelihaz
because.of sedition indef
‘because of a certain sedition’
seditionem / quandam (Lk 23,19)
(T 311, 18–19; around 830 AD)
To conclude, the OHG marked types of indefinites cannot be confined to epistemic specificity because already in the earliest texts, they are licensed in contexts in which the speaker is not able to identify the referent of the respective
noun phrase, or in which the precise identity of the referent is not definable.
adopts this interpretation. But in my opinion, it is unnatural. In the respective context, we
find no indications regarding the specialty of the tree.
Unauthenticated
Download Date | 6/14/17 4:02 PM
DE GRUYTER MOUTON
Variation and change of indefinite expressions
235
3.2.4 Implication of relevance, or discourse prominence
Another aspect of specificity, addressed in the general discussion on non-canonical types of expressions, concerns the pragmatic behavior of the respective
referent in the discourse. Very generally, the choice of a non-canonical form of
indefinite expression has been taken to imply that the denoted referent bears
special relevance in the context. However, the notion of relevance has been
pursued in two principled ways. On the one hand, Ionin (2006) has proposed
an approach to specificity in terms of noteworthiness, arguing that indefinite
this in English is felicitous in contexts in which the speaker states an unexpected, or noteworthy property of the referent. This condition is fulfilled in (32a)
where the context conveys information suggesting that the respective stamp is
worthy of note. By contrast, in (32b), there is nothing special about this referent, and consequently, the use of this is infelicitous:
(32) a.
He put on a / this 31 cent stamp on the envelope, and only realized later
that it was worth a fortune because it was unperforated.
b. He put on a / #this 31 cent stamp on the envelope, so he must want it
to go airmail.
On the other hand, relevance has been understood in terms of discourse prominence, i.e. as a strategy of marking a referent that will be resumed in the following context. It has been observed that the use of some special forms of noncanonical indefinites, such as xad in spoken Hebrew (Givón 1981; Wright and
Givón 1987: 8) or odinR in Russian (Ionin 2013), or various other languages
reported in Wright and Givón (1987), is infelicitous if the referent of the indefinite expression is dismissed in the subsequent discourse, see (33a) vs. (33b):
(33) a.
Maša pošla v magazin, kupila tam odnu
knigu i
Mary went to store
bought ther indef.acc.sg.fem book and
potom čitala ee ves’ večer
then read it all evening
‘Mary went to the store, bought a specific book there and read it all
evening’
Unauthenticated
Download Date | 6/14/17 4:02 PM
236
Svetlana Petrova
b.
DE GRUYTER MOUTON
Maša pošla v magazin, kupila tam #odnu
knigu a
Mary went to store
bought there indef.acc.sg.fem book but
potom prišla domoj i
legla spat’
then came home and went sleep
‘Mary went to the store, bought a specific book there and then came
home and went to bed’
(Ionin 2013: 77–78)
Several studies have compared the discourse prominence of referents of canonical and non-canonical expressions by measuring the frequency of anaphoric
resumption in the following context. Wright and Givón (1987) have examined
how often the referents of specially marked indefinites are resumed in the following context, compared to that of the canonical type (referential persistence)
and how often this anaphoric expression provides the sentence topic of the
utterance (topic persistence). The same method has been applied to investigating the contribution of dieser and so’n compared to ein N in German (Deichsel
2011; Deichsel and von Heusinger 2011; von Heusinger 2012).21 The overall result
of these studies is that in contrast to canonical indefinites, non-canonical ones
endow their referents with a special discourse potential regarding the following
context. This supports the hypothesis that referring to an entity by use of a
marked indefinite expression implies that the speaker intends to continue the
discourse by saying something about this referent.
Both aspects of the notion of relevance, which need to be kept apart (see
Deichsel 2011), are recognizable in the previous discussion of the role of marked
indefinite expressions in OHG. It has been argued that indefinites of the type
sum N and ein N in historical German introduce referents bearing a special,
noteworthy feature, but also that they are relevant as protagonists in the story.
It has also been observed that, as a consequence of their discourse pragmatic
relevance, entities denoted by indefinite expressions involving sum and ein are
likely to be resumed by means of anaphoric reference in the following context
(see von Kraus 1930; Oubouzar 2000; Szczepaniak 2009). Consider the follow21 Not much is known about the emergence of the novel indefinite markers dieser N and so’n
N in German (see von Heusinger 2012 who accounts for the young age of these types of
expressions). Maybe an OHG precursor of indefinite dieser N can be observed in the following
example from Otfrid, displaying the simple demonstrative this as a means of introducing a
referent to the discourse:
(i) Fluog er súnnun
pad […] zi theru /
ítis / frono
flew he sun.gen.sg path
to this.dat.sg lady blissful
‘[The angel of the Lord] went along the path of the sun to this blissful lady’
(O I, 5, 5–6; around 870)
Unauthenticated
Download Date | 6/14/17 4:02 PM
DE GRUYTER MOUTON
Variation and change of indefinite expressions
237
ing citations from Szczepaniak (2009), who resumes observations originally
made by Oubouzar (2000: 260):
“sum determiniert solche Nominalphrasen, die saliente, zentrale Diskursreferenten bezeichnen […], von denen im weiteren Text noch die Rede sein wird” = sum applies to
nominal phrases which denote salient, core discourse referents […] which will be resumed
in the following passage. (Szczepaniak 2009: 82)
“Mit ein wird der außergewöhnliche Charakter des eingerührten Referenten unterstrichen.
Er wird als etwas Besonderes dargestellt, worüber man noch mehr hören wird” = Ein
emphasizes that the referent just introduced is of extraordinary nature. He is characterized as something special which the reader will learn more about in the course of the
story. (Szczepaniak 2009: 83)
Note, however, that anaphoric resumption of the referent of a marked indefinite
expression is not obligatory in OHG. E.g., already the two earliest instances of
ein N found in Hamelburger Marktbeschreibung (777) are left unresumed in the
following context (34a). Neither is the referent introduced by ein N in (34b),
which is actually the example that has raised the discussions quoted above (cf.
also fn. 20). Consider that ein N is also possible in discourse-final position, e.g.
in the last sentence of a chapter, as in (34c), in the last line of Book III, Chapter
25 of Otfrid’s Gospel Book. This is incompatible with the idea that using this
type of indefinite expressions implies the speaker’s intention to continue the
discourse on the respective referent.
(34) a.
inde
in ein
sol,
inde
in ein
steininaz
hog,
thence in indef wallow thence in indef stony.acc.sg.neut hill
inde
in Steinfirst, inde
in Sala in then
elm [end of passage]
thence in Steinfirst thence in Saale to the.acc elm
‘thence to a wallow, thence to a stony hill, thence to Steinfirst, thence
to the river Saale, to the elm’
(HM 21–22; 777 AD)
b.
arsteig In einan
murbóum / thaz her inan
climbed into indef.acc.sg sycamore.tree that he him.acc.sg
gisahi. /
uuanta her thanan uuas farenti
saw.pret.subj while he there was going
‘climbed up into a sycamore tree in order to see Him [= Jesus] passing
by there’
ascendit In arborem sicomorum / ut uider& illum / quia Inde erat transiturus
(T 186, 30–187, 1; around 830 AD)
Unauthenticated
Download Date | 6/14/17 4:02 PM
238
Svetlana Petrova
c.
DE GRUYTER MOUTON
joh fuar mit sínen
thanana in eina wúastinna
and went with his.dat.pl from.there in indef desert
‘and went with His disciples from there to a desert place’
(O III 25, 40; around 870 AD)
But although referents of marked indefinite expressions do not have to be resumed categorically, referential persistence might still play a role in the choice
of indefinites in OHG. Donhauser and Petrova (2012) provide a pilot study examining a small corpus containing an equal number of bare nouns and indefinites
marked by sum and ein from Tatian and Otfrid, and determine the number of
anaphoric resumption of the respective referents in eight following clauses.
They find out that although marked indefinite expressions in OHG are not resumed obligatorily, their referents are more likely to be taken up by anaphors
in the following discourse than those of bare nouns.
I examined the correlation between the type of expression and discourse
potential of a referent for the corpus of the present study. I determined the
discourse persistence of referents of the types of indefinite noun phrases by
counting the frequency of anaphoric mention of these referents in seven subsequent clauses. I ignored the contexts in which the indefinite fails to refer (e.g.
negative root clauses, special complements clauses, generic and predicative
expressions), because non-referents do not license discourse anaphors. But I
retained indefinites in modalized and future contexts because they may give
rise to so-called short-term referents in the sense of Karttunen (1976), see fn. 14.
Table 2 shows how often referents of bare nouns and modified expression
are resumed within 7 clauses after introduction, from no resumption (0) to
seven resumptions (7). Table 2 also gives the absolute number of instances of
referential indefinite expressions licensing anaphoric resumption of their referent, and the sum of resumptions for each type.
On the basis of this data, I determined the ratio of resumed and unresumed
referents of bare nouns and modified indefinites in the various stages of the
OHG attestation, see Table 3.
Table 3 shows that the percentage of resumed referents of bare nouns is
higher than that of modified expressions in the beginning of the attestation
(Phase I) but decreases over time. By contrast, with modified indefinite expressions, the ratio of resumed referents increases. It exceeds the ratio of resumed
referents of bare nouns in Phase II and remains higher than this during the
remaining part of the investigated period.
This picture is confirmed if we calculate the average resumption of referents of various types of indefinite expressions in the different phases of the
attestation. For this purpose, I apply a method used by Wright and Givón (1987)
Unauthenticated
Download Date | 6/14/17 4:02 PM
DE GRUYTER MOUTON
239
Variation and change of indefinite expressions
Table 2: Anaphoric resumption of referents of bare nouns and modified expressions.
Phase I
bare nouns
modified nouns
Phase II bare nouns
modified nouns
Phase III bare nouns
modified nouns
Phase IV bare nouns
modified nouns
Phase V bare nouns
modified nouns
frequency of resumption
0
1
2
3
4
5
6
7
total of
examples
sum of
resumptions
15
6
5 4
2
–
1 –
47 20 10 5
5
3
3 6
23
4
4 1
7 23
9 9
2
–
1 –
1
2
– 1
3
1
– 1
13
3
2 2
1
–
2
3
5
6
–
1
–
3
–
–
5
1
1
8
–
–
–
3
40
4
101
35
49
71
3
6
5
29
70
6
146
115
103
201
2
15
4
71
9
1
8
7
4
4
–
1
–
1
–
–
4
7
7
5
–
–
–
3
Table 3: Anaphoric resumption of referents of different types of indefinite noun phrases.
bare nouns
total
Phase
Phase
Phase
Phase
Phase
I
40
II 101
III 49
IV
3
V
5
modified expressions
Resumed
N
%
unresumed
N
%
total
resumed
N
%
unresumed
N
%
25
54
26
1
2
15
47
23
2
3
4
35
71
6
29
2
30
64
5
16
2
5
7
1
13
62.5
53.5
53.1
33.3
40.0
37.5
46.5
46.9
66.6
60.0
50.0
85.7
90.1
83.3
55.2
50.0
14.3
8.9
26.7
44.8
to determine referential persistence as the product of the sum of resumptions
and the number of referents available for resumption. The results are given in
Table 4.
Table 4 shows that the average resumption of referents of bare nouns is
higher than that of marked expressions in the earliest attestation (Phase I). In
this phase, every referent introduced by a bare noun reappears 1,75 times on
average within the following seven clauses, while referents of marked expressions are resumed slightly less often, namely 1,5 times on average. But from
the time of composition of Tatian (around 830 AD) onwards, we observe an
increase in the average resumption of referents of marked indefinite expressions. It raises from 1,5 in Phase I to 3,3 in Phase II and exceeds the average
resumption of referents of bare nouns consistently in each of the subsequent
periods.
Unauthenticated
Download Date | 6/14/17 4:02 PM
240
DE GRUYTER MOUTON
Svetlana Petrova
Table 4: Average resumption of referents of various indefinite expressions in OHG.
Phase
Phase
Phase
Phase
Phase
bare nouns
modified expressions
total of
sum of
average
instances resumptions resumption
total of
instances
sum of
resumptions
average
resumption
4
35
71
6
29
6
115
201
15
71
1.5
3.3
2.8
2.6
2.4
I
40
II 101
III
49
IV
3
V
5
70
146
103
2
4
1.75
1.4
2.1
0.7
0.8
Table 5: The significance of resumption in the selection of noun phrase type.
bare nouns
Phase II
Phase III
modified expressions
resumed
unresumed
resumed
unresumed
54
26
47
23
30
34
5
7
result of χ2-test
χ2 = 11.447, p < .001
χ2 = 8.960, p = .003
Note, however, that the ratio of resumed referents of bare nouns still remains relatively high (e.g., it is still above 50 % in Phase II and III, see Table
3). From the figures presented above, we cannot tell if there is a statistically
significant correlation between the choice of a particular type of expression
and the intention of the speaker to continue the discourse on the referent of
that expression. This is revealed by the χ2-test. With low frequencies, as in
Phase I, IV and V, this test is not applicable. But for Phase II and III, we obtain
frequencies that allow statistical examination. The results are provided in Table
5. In both cases we obtain a statistically significant correlation between the
type of the indefinite expression and the discourse potential of the referent.
Let us summarize the findings of this examination. We observe that resumption is a factor that governs the choice of marked indefinites over bare
nouns from the middle of the ninth century onwards. But this does not necessarily indicate that marked indefinites are a genuine means of expressing specificity in terms of discourse prominence. First, they trigger resumption of the
referent less often at the beginning of the attestation, which amounts to the
conclusion that signaling prominence is not their inherent property at the time
when they enter the system. Second, the resumption of referents remains optional for marked indefinites during the entire OHG attestation, which is untypical of markers of prominence like those reported in Wright and Givón (1987).
Unauthenticated
Download Date | 6/14/17 4:02 PM
DE GRUYTER MOUTON
Variation and change of indefinite expressions
241
Third, because of its delay in time, the increase in the preference for marked
indefinites in cases of resumption contradicts our expectation that over time,
these expressions will become semantically more abstract and vague. In sum,
the picture that we encounter suggests that resumption starts to influence the
choice for one variable over the other during the OHG period, but it does not
enforce the conclusion that the core meaning of this variable is to mark specificity in terms of discourse prominence.
4 Conclusions and outlook
Relating to the current debate on the properties of different types of indefinites
in language, the present paper addresses formal variation in the domain of
indefinite reference in OHG, the earliest recorded period of German. It introduces the inventory of different competing types of indefinite noun phrases and
discusses the semantic and discourse-pragmatic behavior of the marked, or
non-canonical types of expressions with respect to properties considered constitutive of specific indefinites in well-known theoretic and typological work.
The present analysis shows that contrary to previous claims, none of the
marked types of indefinites in historical German satisfies necessary conditions
on specificity in terms of any of the notions distinguished in the semantic literature. By contrast, all types of marked indefinite expressions display semantic
ambiguities already at the earliest stage of their attestation, despite of the fact
that they occur at a low rate in the data, compared to the standard type of
indefinites.
This study not only shows that on closer inspection, all previous claims
about marked indefinite expressions in historical German prove wrong, but it
also challenges the application of well-known scenarios of the emergence of
indefinite articles to the German data. Givón (1981) accounts for a seemingly
universal process whereby indefinite determiners emerge from quantifiers,
more specifically from the numeral ‘one’. An important step in this process, he
claims, is that the numeral extends its original meaning beyond that of expressing cardinality and acquires the status of a referential-indefinite marker, assigning specific interpretation to a noun phrase.
Givón (1981) introduces the method of mapping the synchronic uses, or
categories of indefiniteness, as individual stages onto a diachronic scale of
development. The evolution of indefinite articles from numerals is accounted
for in terms of the three-stage model given in (35):
Unauthenticated
Download Date | 6/14/17 4:02 PM
242
DE GRUYTER MOUTON
Svetlana Petrova
(35) Stage I
numeral/
quantifier
Stage II
referential-specific
marker
Stage III
generic meaning/
indefinite article
In the history of German, the numeral ein gives rise to a grammaticalized indefinite article. But the results of the present study show that a separate stage
during which ein represents a genuine marker of referential-specific interpretation cannot be identified in the data. It might be speculated that such a stage
was present before the beginning of the recorded history of German. Note, however, that the situation revealed for OHG is not exceptional. Carlier (2012) on
Old French and Geist (2013) on Bulgarian similarly show that the original numerals uns and edin ‘one’ are broader in meaning than prototypical specificity
marker, without being equivalent to fully grammaticalized indefinite determiners.
Another claim, accounting for the emergence of marked indefinite expressions, turns out to be definitely wrong from the perspective of the German data.
According to Wright and Givón (1987: 28), special formal marking on indefinites
does not arise in order to distinguish them from other types of indefinites on
semantic grounds – note that the unmarked, or standard types of indefinites
may be semantically specific as well – but as a strategy of marking pragmatically specific indefinites, i.e. phrases whose referents will matter in the following
discourse. Von Heusinger (2012), discussing the properties of the modern German indefinite dieser and so’n, expresses doubts regarding the validity of this
claim. The present study provides an example showing that special forms of
indefinites do not emerge as markers of referential persistence, i.e. the German
development contradicts the assumed scenario that marking of pragmatic specificity necessarily comes first in diachronic plan.
One aspect of diachronic development might be revealed by the OHG data,
though. Following suggestions originally addressed by Givón (1981), Heine
(1997) proposes a more fine-grained classification of semantic types of indefiniteness, yielding a more elaborate diachronic scale of the emergence of indefinite determiners, represented in (36):
(36) Heine’s (1997) evolutionary
Stage I
Stage II
numeral presentational
marker
scale of indefinites markers
Stage III
Stage IV
Stage V
specificnon-specific
indefinite
unresumed non-referential determiner
In this model, generic indefinites are treated as representatives of stage IV of
non-specific/non-referential indefinites, while predicative use is constitutive of
Unauthenticated
Download Date | 6/14/17 4:02 PM
DE GRUYTER MOUTON
Variation and change of indefinite expressions
243
the final stage V of the implicational scale. Geist (2013) and von Heusinger and
Klein (2013) apply the implicational scale of Heine (1997) by identifying the
degree of spread of the indefinite markers in the languages they investigate.
Applied to the OHG data, Heine’s (1997) scale helps to recognize a gradual
change in the semantic development of ein, which is missing with the remaining markers, namely that ein enters the domain of generic indefinites in Phase
III of the OHG attestation, and later, in Phase IV, the domain of nominal predicates. However, the consolidation of the present-day situation, incl. some wellknown interpretational effects emerging from the variation between ein N vs.
bare nouns as nominal predicates, which modern German shares with other
languages (Geist 2014 for summary and references), has to be left for future
research.
Acknowledgements: Earlier versions were presented at RED14 – Indefinites in
Discourse Structure, June 27–28, 2014 in Cologne, and Auf- und Abbau des Definitartikels im Deutschen, September 4–5, 2014 in Hamburg. I am grateful to
the participants of these conferences for questions, comments and suggestions.
Furthermore, I would like to thank Renata Szczepaniak for sharing an unpublished manuscript with me. I also profited from discussions and cooperations
with Ines Fiedler, Chiara Gianollo, Ljudmila Geist, Klaus von Heusinger, Milena
Kühnast and Edgar Onea, and from valuable comments and criticism by two
anonymous reviewers as well as by the editors of Zeitschrift für Sprachwissenschaft. I also appreciate Carsten Dahlmann’s help with the proof-reading and
the styles.
Sources
APB
APh
BBS
Ch
DD
G
HiH
Althochdeutsche Predigtsammlung B. In Elias von Steinmeyer (ed.), Die kleineren
althochdeutschen Sprachdenkmäler, 168–172. Dublin/Zürich: Weidmann, 1971 [1916].
Althochdeutscher Phyiologus. In Elias von Steinmeyer (ed.), Die kleineren
althochdeutschen Sprachdenkmäler, 124–134. Dublin/Zürich: Weidmann, 1971 [1916].
Bamberger Blutsegen. In Elias von Steinmeyer (ed.), Die kleineren althochdeutschen
Sprachdenkmäler, 377–380. Dublin/Zürich: Weidmann, 1971 [1916].
Christus und die Samariterin. In Elias von Steinmeyer (ed.), Die kleineren
althochdeutschen Sprachdenkmäler, 89–91. Dublin/Zürich: Weidmann, 1971 [1916].
De Definitione. In Elias von Steinmeyer (ed.), Die kleineren althochdeutschen
Sprachdenkmäler, 118–120. Dublin/Zürich: Weidmann, 1971 [1916].
Georgslied. In Fritz Tschirch (ed.), Frühmittelalterliches Deutsch, 45–49. Halle:
Niemeyer, 1955.
Himmel und Hölle. In Elias von Steinmeyer (ed.), Die kleineren althochdeutschen
Sprachdenkmäler, 153–155. Dublin/Zürich: Weidmann, 1971 [1916].
Unauthenticated
Download Date | 6/14/17 4:02 PM
244
HM
I
L
MF
MM
NL
O
T
Svetlana Petrova
DE GRUYTER MOUTON
Hamelburger Marktbeschreibung. In Elias von Steinmeyer (ed.), Die kleineren
althochdeutschen Sprachdenkmäler, 62. Dublin/Zürich: Weidmann, 1971 [1916].
Der althochdeutsche Isidor. Facsimile-Ausgabe des Pariser Codex nebst christlichem
Texte der Pariser und Monseer Bruchstücke. Mit Einleitung, grammatischer Darstellung
und einem ausführlichem Glossar. George A. Hench (ed.). Straßburg: Karl J. Trübner,
1893.
Ludwigslied. In Elias von Steinmeyer (ed.), Die kleineren althochdeutschen
Sprachdenkmäler, 85–88. Dublin/Zürich: Weidmann, 1971 [1916].
The Monsee Fragments. Newly collated text with introduction, notes, grammatical
treatise and exhaustive glossary and a photo-lithographic fac-simile. George A. Hench
(ed.). Straßburg: Karl J. Trübner, 1890.
Memento Mori. In Wilhelm Braune (ed.), Althochdeutsches Lesebuch, 142–144. 17th
edn. Tübingen: Niemeyer, 1994.
Das Nibelungenlied. Helmut de Boor (ed.). 21th edn. Wiesbaden: Albert, 1979.
Otfrids Evangelienbuch. Oskar Erdmann (ed.). 6th edn. Tübingen: Niemeyer, 1973.
Die lateinisch-althochdeutsche Tatianbilingue Stiftsbibliothek St. Gallen Cod. 56.
Achim Masser (ed.). Göttingen: Vandenhoeck & Ruprecht, 1994.
References
Behaghel, Otto. 1917. Die Indefinitpronomina HWAS und SUMS. Beiträge zur Geschichte der
deutschen Sprache und Literatur 42. 158–161.
Behaghel, Otto. 1923. Deutsche Syntax. Eine geschichtliche Darstellung. Band I. Die
Wortformen. A. Nomen. Pronomen. Heidelberg: Winter.
Braune, Wilhelm. 2004. Althochdeutsche Grammatik I: Laut- und Formenlehre. 15th edn.
Tübingen: Niemeyer.
Carlier, Anne (2012). Grammaticalization in progress in Old French: Indefinite articles. In
Deborah L. Arteaga (ed.), Research on Old French: The state of the art, 45–60.
Dordrecht: Springer.
Chung, Sandra & William A. Ladusaw. 2004. Restriction and saturation. Cambridge, MA: MIT
Press.
Deichsel, Annika. 2011. The discourse effects of the indefinite demonstrative dieser in
German. In Proceedings of the 23rd European Summer School in Logic, Language, &
Information (ESSLLI) – ESSLI2011 Student Session, 70–78. http://web.stanford.edu/
~danlass/esslli2011stus/proceedings.html (accessed 16 June 2015).
Deichsel, Annika & Klaus von Heusinger. 2011. The cataphoric potential of indefinites in
German. In Iris Henrickx, Sobha Lalitha Devi, Antonio Branco & Ruslan Mitkov (eds.),
Anaphora and reference resolution. 8 th Discourse Anaphora and Anaphor Resolution
Colloquium, DAARC 2011, Faro, Portugal, 144–156. Berlin: Springer.
Desportes, Yvon. 2000. Artikel im Mittelhochdeutschen. Lässt sich Paul Valentins Modell
des Artikelsystems im heutigen Deutsch auf das Mittelhochdeutsche übertragen? In
Yvon Desportes (ed.), Zur Geschichte der Nominalgruppe im älteren Deutsch, 213–253.
Heidelberg: Winter.
Donhauser, Karin. 1995. Gibt es einen Indefinitartikel Plural im Althochdeutschen? –
Überlegungen zur Syn- und Diachronie des deutschen Artikelsystems. In Gisela
Unauthenticated
Download Date | 6/14/17 4:02 PM
DE GRUYTER MOUTON
Variation and change of indefinite expressions
245
Brandt & Reiner Hünecke (eds.), Wie redet der Deudsche man jnn solchem fall? Studien
zur deutschen Sprachgeschichte. Festschrift anlässlich des 65. Geburtstages von Erwin
Arndt, 61–72. Stuttgart: Akademischer Verlag Stuttgart.
Donhauser, Karin & Svetlana Petrova. 2012. Sprachliche Strategien zur Aktivierung und
Deaktivierung von Diskursreferenten in deutschsprachigen Texten des Mittelalters.
Paragrana. Internationale Zeitschrift für Historische Anthropologie 21(2). 159–176.
Erben, Johannes. 1950. Syntaktische Untersuchungen zu einer Grundlegung der Geschichte
der indefiniten Pronomina im Deutschen. Beiträge zur Geschichte der deutschen
Sprache und Literatur 72. 193–221.
Farkas, Donka. 2002. Specificity distinctions. Journal of Semantics 19(3). 213–243.
Fobbe, Eilika. 2004. Die Indefinitpronomina des Deutschen. Aspekte ihrer Verwendung und
historischen Entwicklung. Heidelberg: Winter.
Fodor, Janet & Ivan Sag. 1982. Referential and quantificational indefinites. Linguistics and
Philosophy 5(3). 355–398.
Gallmann, Peter. 1997. Zur Morphosyntax und Lexik der w-Wörter. In Arbeitspapiere des
Sonderforschungsbereichs 340 – Sprachtheoretische Grundlagen für die
Computerlinguistik, Bericht Nr. 107. Tübingen: Universität Tübingen.
Geist, Ljudmila. 2013. Bulgarian edin: The rise of an indefinite article. In Uwe Junghanns,
Dorothee Fehrmann, Denisa Lenertová & Hagen Pitsch (eds.), Formal description of
Slavic languages (FDSL) 9, 125–148. Frankfurt am Main: Peter Lang.
Geist, Ljudmila. 2014. Bare predicate nominals in German. In Ludmila Veselovská & Markéta
Janebová (eds.), Complex visibles out there. Proceedings of the Olomouc Linguistic
Colloquium 2014: Language use and linguistic structure, 83–106. Olomouc: Palacký
University.
Giannakidou, Anastasia. 2011. Negative and positive polarity items. In Klaus von Heusinger,
Claudia Maienborn & Paul Portner (eds.), Semantics. An international handbook of
natural language meaning, vol. 2, 1660–1712. Berlin & Boston: de Gruyter.
Gianollo, Chiara, 2013. Latin aliquis as an epistemic indefinite. In Sofiana Chiriacescu (ed.),
Proceedings of the VI Nereus International Workshop ‘Theoretical implications at the
syntax/semantics interface in Romance’ (Arbeitspapiere des Fachbereichs
Sprachwissenschaft 127), 55–81. Konstanz: Universität Konstanz.
Givón, Talmy. 1973. Opacity and reference in language: An inquiry into the role of
modalities. In John P. Kimball (ed.), Syntax and semantics, vol. 2, 95–122. New York &
London: Seminar Press.
Givón, Talmy. 1981. On the development of the numeral ‘one’ as an indefinite marker. Folia
Linguistica Historica 2. 35–53.
Grimm, Jacob. 1837. Deutsche Grammatik, Band 4. Göttingen: Dieterich.
Haspelmath, Martin. 1997. Indefinite Pronouns. Oxford: Claredon Press.
Heine, Bernd. 1997. Cognitive Foundations of grammar. Oxford: Oxford University Press.
von Heusinger, Klaus. 2011. Specificity. In Klaus von Heusinger, Claudia Maienborn & Paul
Portner (eds.), Semantics. An international handbook of natural language meaning, vol.
2, 1025–1058. Berlin & Boston: de Gruyter.
von Heusinger, Klaus. 2012. Referentialität, Spezifizität, und Diskursprominenz im
Sprachvergleich. In Lutz Gunkel & Gisela Zifonun (eds.), Deutsch im Sprachvergleich –
Grammatische Kontraste und Konvergenzen, 417–455. Berlin & Boston: de Gruyter.
von Heusinger, Klaus & Udo Klein. 2013. The distribution of two indefinite articles: The case
of Uzbek. In Cornelia Ebert & Stefan Hinterwimmer (eds.), Different kinds of specificity
across languages, 155–176. Dordrecht: Springer.
Unauthenticated
Download Date | 6/14/17 4:02 PM
246
Svetlana Petrova
DE GRUYTER MOUTON
Ionin, Tania. 2006. This is definitely specific. Natural Language Semantics 14. 175–234.
Ionin, Tania. 2010. Specificity. In Louise Cummings (ed.), The pragmatics encyclopedia, 449–
452. London: Routledge.
Ionin, Tania. 2013. Pragmatic variation among specificity markers. In Cornelia Ebert & Stefan
Hinterwimmer (eds.), Different kinds of specificity across languages, 75–103. Dordrecht:
Springer.
Jaggar, Philip. 1985. Factors governing morphological coding of referents in Hausa narrative
discourse. Los Angeles, CA: University of California dissertation.
Jaggar, Philip. 1988. Discourse-deployability and indefinite NP-marking in Hausa. In Graham
Furniss & Philip Jaggar (eds.), Studies in Hausa language and linguistics, 45–61.
London: Keegan Paul.
Jäger, Agnes. 2007. No changes: On the history of German indefinite determiners in the
scope of negation In Elisabeth Stark, Elisabeth Leiss & Werner Abraham (eds.), Nominal
determination: Typology, context constraints, and historical emergence, 141–170.
Amsterdam: Benjamins.
Jäger, Agnes. 2010. Anything is nothing is something. On the diachrony of polarity types of
indefinites. Natural Language and Linguistic Theory 28(4). 787–822.
Karttunen, Lauri. 1976. Discourse referents. In James D. McCawley (ed.), Syntax and
semantics 7: Notes from the linguistic underground, 363–385. New York: Academic
Press.
von Kraus, Carl. 1930. Das sog. demonstrative ein im Mittelhochdeutschen. Zeitschrift für
deutsches Altertum und deutsche Literatur 67. 1–22.
Matthewson, Lisa. 1998. On the interpretation of wide-scope indefinites. Natural Language
Semantics 7(1). 79–134.
Oubouzar, Erika. 2000. Zur Entwicklung von ein in der Nominalgruppe des
Althochdeutschen. In Yvon Desportes (ed.), Zur Geschichte der Nominalgruppe im
älteren Deutsch, 255–268. Heidelberg: Winter.
Rissanen, Matti. 1988. Where philology and linguistics meet: Reference, (in)definiteness and
Old English sum. In Dieter Kastovsky & Gero Bauer (eds.), Luick revisited. Papers read
at the Luick-Symposium at Schloss Liechtenstein 1985, 295–310. Tübingen: Narr.
Rissanen, Matti. 1997. The pronominalization of one. In Matti Rissanen, Merja Kytö & Kirsi
Heikkonen (eds.), Grammaticalization at work: Studies of long-term developments in
English, 87–143. Berlin & New York: de Gruyter.
Szczepaniak, Renata. 2009. Grammatikalisierung im Deutschen. Tübingen: Narr.
Wright, Sue Ellen & Talmy Givón. 1987. The pragmatics of indefinite reference: Quantified
text-based studies. Studies in Language 11(1). 1–33.
Unauthenticated
Download Date | 6/14/17 4:02 PM