ArticlePDF Available

Grammaticalization and related changes in contemporary German



This is an overview of ongoing or fashionable grammaticalization processes in different grammatical domains of German at the end of the 20th century, with a glance at some related grammatical changes.
Christiani Lehmanni inedita, publicanda, publicata
Grammaticalization and related changes in
contemporary German
huius textus situs retis mundialis
dies manuscripti postremum modificati
occasio orationis habitae
Symposium on 'Grammaticalization', Eugene, Or., 12.-
volumen publicationem continens
Traugott, Elizabeth C. & Heine, Bernd (eds.) 1991,
Approaches to grammaticalization. Vol. II: Focus on types
of grammatical markers. Amsterdam & Philadelphia: J.
Benjamins (Typological Studies in Language, 19:2); 2
annus publicationis
1This paper has greatly benefitted from comments by Ulrike Claudi, Hartmut Haberland, Bernd Heine,
Ekkehard König, John McWhorter, Joachim Mugdan, Thomas Stolz, Elizabeth Traugott and Joseph
Voyles. Their contributions are acknowledged here with sincere thanks, but not singled out in the
following text.
2Cf. Lehmann 1982 for a systematic treatment.
Grammaticalization and related changes
in contemporary German
Christian Lehmann
University of Bielefeld
1. Introduction1
1.1. In a sense, every grammatical change leads to a new grammar. The term
grammaticalization could be interpreted as ‘creation of new grammar’ (where grammar
is taken as a mass noun). Then grammatical change would equal grammaticalization.
Extension is a fate that traditional terms commonly undergo when they are taken up in
modern linguistics. Here, ‘grammaticalization’ will be used in a narrower sense. A
variation affecting a linguistic unit will be considered a case of grammaticalization if it
is directed towards subjecting that unit more to rules of grammar. This proposition is not
meant to carry any implications as to whether rules pre-exist units or vice versa. Among
the implications that it does carry are the following 2:
- The element in question turns from a less grammatical (initially, lexical) element into
a more grammatical one.
- The element loses substance both on the phonological and the semantic sides; its
selection restrictions are loosened.
- The freedom to manipulate the element decreases. It is integrated into a paradigm, it
becomes increasingly obligatory in certain constructions, and it occupies a fixed
(ultimately, morphological) position.
A paradigm case which illustrates most of these features is the grammaticalization of the
Latin preposition ad ‘to’ to the Spanish case marker a ‘accusative’. Phonological attrition
goes hand in hand with the loss of concrete semantic features. The original sense of
direction towards an entity or a place is extended to mark a being that is reached by the
verbal action. This yields an indirect object marker. As such, its locus is with human
complements. Then it is extended towards a direct object marker, first only of human, then
increasingly of other animate objects. At the same time, it regularly fuses with the definite
Christian Lehmann
article el; and it enters a minimal paradigm with de (which itself is grammaticalized from
a Latin local preposition).
Analogical change is the other principal kind of grammatical change. The weak
conjugation of erstwhile strong verbs in several Germanic languages is a well-known
example. Grammaticalization and analogy have some points in common, among them the
expansion of the distribution of an element and the introduction of a subregularity into the
grammar. Also, grammaticalization may involve reanalysis of an underlying construction.
Reanalysis, in turn, presupposes a pattern which can serve as its model. For instance, the
grammaticalization of German haben ‘have’ to a perfect auxiliary involves, as an early
stage, the construction [hat NPacc Vpart. pass.], where the NPacc is the direct object of hat and
the Vpart.pass. is a predicative complement of the direct object. This is reanalyzed as a
periphrastic form of V, which takes NPacc as its direct object. This reanalysis, however,
is rendered possible only because German already possessed the periphrastic verb form
[ist Vpart.pass.], which thus could serve as an analogical model. For such reasons, it is often
difficult to distinguish grammaticalization from analogical change.
Finally, grammatical change is to be distinguished from lexical change. The latter often
also involves reduction and coalescence, as when Anglo-Saxon hl
‘bread-warden’ becomes Modern English lord. Changes like terrific ‘horrifying’ >
‘remarkable’ have both the generalization of meaning and use and the codification in
certain grammatical positions (here, the position of an evaluator/intensifier) in common
with grammaticalization. Therefore, there will also be borderline cases between
grammatical and lexical change.
1.2. The purpose of this contribution is to review a variety of grammatical constructions
in contemporary non-literary German which are currently in vogue but are either not
considered or at least not (yet) accepted by standard grammars. All of them share at least
some of the features that are characteristic of grammaticalization. Some of them can be
classed as cases of grammaticalization beyond doubt. Others are not so clear and may
show more aspects of analogical grammatical change or of lexical change. They will be
presented nonetheless, because they occasion some discussion on the limits of
grammaticalization. There will also be the question as to whether the phenomena in their
totality indicate a drift in the contemporary language.
Not being a Germanist, I am not in the best position to write on grammaticalization in
German. The following survey is not intended to be exhaustive or balanced. Some changes
will be highlighted, others mentioned only in passing. For the past three years or so, I have
been gathering data on the present topic. They are instances of ongoing changes or current
fashions, occasionally encountered and unsystematically noted down. Some of them are
tied up with my own dialect environment, which is Northwest Germany. For several of the
phenomena to be mentioned below, I have no spontaneous data at all. Thus, my data
cannot claim to be representative or to show the extent to which a phenomenon is or has
been gaining ground. Furthermore, no arguments ex silencio are allowed: it is quite
Grammaticalization and related changes in contemporary German
3As usual, the ongoing change creates uneasiness, manifesting itself, e.g., in hypercorrection. Thus,
a club president addresses the ‘dear members’, in a circular of 1987, by Liebe Mitglieder/-innen, as if
Mitglied were masculine or Mitglied(er)innen could be a grammatical form of German; and a report on
women's fight sports features Der Kung-Fu-Lehrer ... schätzt die Zahl seiner weiblichen
Kursteilnehmerinnen auf ungefähr zwanzig Prozent. [Zeit-Magazin, 1988] ‘The Kung-Fu instructor
estimates the number of his female course participantesses at about 20%.’
possible that the most interesting ongoing changes have escaped my notice. To the extent,
however, that there are data, they prove that the phenomenon exists at a supra-idiolectal
There is also the methodological difficulty of chronology. Historical linguists and
philologists are used to working through the relevant corpus in order to ascertain the
earliest occurrence of a given phenomenon and to trace its tradition. I have not been able
to do this. All I can say is that, except where stated otherwise, the phenomena treated have
aroused my attention during the past ten years or so as things that were not formerly in the
language. Many of them are not in the grammars. Nevertheless, should anyone be inclined
to conclude that just because something is currently in the colloquial language but
condemned by normative grammar, it must be a recent development or indicative of an
ongoing change, it is beneficial to read Sandig 1973. There it is shown that some of the
nonstandard features of contemporary German have been in the language since the Old
High German stage. Even a phenomenon that has been spreading only since very recently
does not need to have been invented very recently. It is quite possible that it can be found
in earlier literature, and such earlier occurrences may or may not be the source of the
recent development.
2. Nominal categories
2.1. Gender and number
2.1.1. There are currently changes in all of the traditional nominal categories, i.e. gender,
number and case. The change in gender concerns the nature of the opposition between the
masculine and feminine genders. As in many other languages, it used to be a privative
opposition, with masculine as the unmarked member. For about fifteen years, this
relationship has been gradually shifted in favor of an equipollent opposition between the
two. This manifests itself clearly in the use of the derivational suffix -in ‘-ess’. The most
obvious examples occur in job announcements. For a position admitting both women and
men, formerly a Lehrer ‘teacher’ would have been sought; now invariably a post for a
Lehrer/in (i.e. Lehrer/Lehrerin) is announced. This is a clear case of a language change
resulting from a change in society.
The phenomenon has some of the characteristics of grammaticalization, above all the
quantitative expansion of the feminine suffix to the detriment of the zero form.3 So far,
however, the suffix remains a derivational one; it has not become a gender marker. Hence,
Christian Lehmann
we should probably not class this change as a case of grammaticalization, but instead learn
from it that expansion should not be used too naively as a criterion of grammaticalization.
2.1.2. One of the German plural allomorphs is -s. While this has a stronger basis in the
Low German dialect, it used to be one of the rarest in standard German. It has occurred
for a long time in certain forms ending in posttonic -el/n (Jungens ‘boys’, Mädels ‘girls’,
Kumpels ‘miners, buddies’), which may be said to belong to the colloquial register of
standard German. Apart from such marginal exceptions, the suffix originated in loans with
one of two non-German phonological traits: nouns ending in a vowel other than
(Mamas, Tees ‘teas’, Nazis, Autos, Zulus), and certain non-feminine nouns stressed on the
last syllable (Kartons, Parks).
It is apparently during the course of this century that -s has been continually gaining
ground (see Wurzel 1984, ch. 4 for details and explanations). It is penetrating into nominal
morphology by a relaxation of the above compound condition: it now suffices that the
noun in question be either a loan or have a stress or final full vowel in the last syllable.
First, while most loan nouns still form their plural in one of the genuinely German
allomorphs (Visionen, Passagiere, Computer(0)), -s is increasingly used on
non-Germanized loans with various phonological structures (Sputniks, nonstandard
Lexikons). Second, it is gradually becoming the usual plural allmorph for words ending
in a long vowel or in a stressed syllable. Native German nouns with such a phonological
structure are rare. Uhus ‘eagle-owls’ is the only plural form for this noun. If wo ‘where’
were substantivized, it would form an -s-plural: die wo's. However, the relevant
phonological classes are fed by abbreviations and acronyms (PCs, Profs), which are very
frequent and where -s is currently most productive both in spoken and in written language.
Other words are changing their plural marking (Tablette > Tabletts ‘trays’, Konten >
Kontos ‘accounts’). Yet others take the -s on top of an irregular plural which is probably
not felt as a morphological plural: nonstandard Spaghettis, Brokkolis.
Paradigm instances of the development of plural morphemes by grammaticalization
include Mandarin men ‘class’ and Bengali jana ‘host’, both of which have become plural
suffixes. The case of German -s, instead, is another example of analogical expansion. It
shows how problematic the distinction between the latter and grammaticalization can be.
Clear cases of grammaticalization, such as the two mentioned, show a loosening of
selection restrictions. This is parallel to the loosening of the restriction on plural -s that
the noun so marked should be foreign and end either in a long vowel or on a stressed
syllable. The reason why we nevertheless distinguish between grammaticalization and
analogical extension is that the latter fails on the central criterion of grammaticalization,
which is that the unit in question should become more subject to rules of grammar.
2.2. Case relations
The development of case markers is one of the standard examples of grammaticalization.
Its reverse side is the loss of old inflectional case markers. Reduction to zero is the logical
Grammaticalization and related changes in contemporary German
4I even have one isolated example without article: Die Botschaft Jesus ‘The Message of Jesus’ (book
title of 1972).
end phase of grammaticalization. Let us first look at the relevant phenomena and then
discuss the problems behind them.
2.2.1. The possessive attribute relation
Case endings have been reduced and lost in German from Proto-Indo-European times.
Feminine nouns, except proper names, no longer inflect for case. Neuter nouns form the
genitive in -s. Masculine nouns fall into two declension classes and accordingly form the
genitive either in -s or in -en. These markers have proved astonishingly resistant.
Nevertheless, those nouns that take the -s allomorph lack it in a variety of cases, all of
which have in common an article form - definite des or indefinite eines -which shows the
genitive. Two of these cases are provided for by standard grammar:
a) Foreign nouns, especially if they end in s, often lack genitive marking: des
Atlas/Numerus, but also des Kazike.
b) Those proper names which are generally articleless (chiefly, person and place names)
do not take their genitive -s after an article. Such names normally take the definite article
only if they are expanded by an attribute, as in des jungen Goethe ‘of young Goethe’, des
alten Berlin ‘of old Berlin’. Names which are by themselves complex allow the definite
article more easily: des Ludwig van Beethoven. However, this construction is not
colloquial and connotes distance between speaker and referent.
The genitive attribute with article, but no mark on the noun(s) is currently extended in two
directions: First, to other composite attributes, especially those including appositions.
Thus, while des Konjunktiv is nonstandard, des Konjunktiv Imperfekt is all right. And
examples like E1 and E2, although not (yet) standard, are increasingly found.
E1. ... ist eine Wanderausstellung des Koblenzer Bundesarchiv geplant. [radio, 1988]
‘... a flying exhibition of the Koblenz Federal Archive is planned.’
E2. Anstelle des traditionell gewählten Mann des Jahres prangt ein Computer auf dem
Titel. [computer magazine, 1988]
‘Instead of the traditionally elected man of the year, the title page boasts a
Second, the postnominal possessive attribute without genitive marking extends to such
simple proper names as are generally accompanied by the definite article: des
Saturn/Mai/Hunsrück (mountain range).4
Inside possessive attribution, there are currently two ways of renewing the genitive. We
first turn to the preposition von ‘from, of’, which has been slowly making its way into the
possessive attribute relation for a long time. Most genitive attributes are paraphrasable by
a prepositional attribute involving von (die Oper von Ludwig van Beethoven ‘Ludwig van
Christian Lehmann
5Cf. Plank 1980, esp. 296-298, on this matter. Hartmut Haberland draws my attention to the fact that
the non-distinctness explanation is not sufficient to generalize to the genitive singular, since while das
Wahrzeichen Oslos ‘Oslo's landmark’ is fine, *die Förderung Erdöls ‘the hauling of petrol’ is not, despite
its distinctness.
Beethoven's opera’), although this is often stylistically clumsy and not high style. Thus,
die CPU von dem Computer will be heard, but not read.
However, there are at least two cases where von must be used to signal the possessive
relation. The first involves a personal possessive relation in an indefinite NP. The
indefinite counterpart to the definite NP mein Schwager (= nonstandard der Schwager von
mir) ‘my brother-in-law’ is ein Schwager von mir, which, keeping Schwager as the head
constant, has no alternative.
Second, no noun has a distinctive genitive plural morpheme. If the genitive noun is
modified by determiners or adjectives, these will have different patterns of case
syncretism. For example, the masculine noun form Effekte can be nominative, genitive or
accusative plural. The adjective form stilistischer can be nominative singular masculine,
genitive and dative singular feminine and genitive plural of any gender. If the two form
a phrase, as in Erzielung stilistischer Effekte ‘attainment of stylistic effects’, the case of
this phrase is calculated as the intersection set of the case meanings of the two constituent
forms, reducing to the genitive plural in this case (see Werner 1979). When the noun
stands alone, case ambiguity is not resolved by morphological means. Now, while the
other cases do not require morphological disambiguation, so that Effekte may constitute
an NP in the nominative or accusative, it may not constitute an NP in the genitive. The
modifier in Erzielung stilistischer Effekte cannot be dropped. Therefore, if Effekte alone
is to constitute an attribute, von has to be used: Erzielung von Effekten.5
The other possibility of formally renewing the possessive attribute is by means of the
possessive pronoun. There is a cross-linguistically widespread construction which
combines the possessed noun with a possessive pronoun (referring to the possessor) and
an (optional) noun phrase in some appropriate syntactic relation (representing the
possessor). In German, this construction is nonstandard and dialectal. There are variants
of it which differ as to case inflection of the possessive NP and possessor gender marking
in the possessive pronoun. A mapping of this construction into Standard German would
look as in E3.a.
E3. a. der Lehrerin ihr Hut
‘the (female) teacher's hat’
b. die Lehrerin ihr Hut
c. die Lehrerin sein Hut
In E3.a, the case relation of the possessive NP is the dative (just as in the corresponding
Hungarian construction), and the possessive pronoun duly reflects the possessor's gender.
Grammaticalization and related changes in contemporary German
However, the dialects and sociolects which actually use this construction (for instance,
Ruhrdeutsch, a variety of Lower Frankish) often also reduce inflection in various ways.
Thus, the possessive NP may appear in the nominative, as in E3.b. Finally, the possessive
pronoun is limited to the masculine-possessor variant, as in c. This is, incidentally, also
what happened in the development of the English genitive 's.
Beside its function in attribution, the genitive has another one in verbal valency. There,
however, it has retreated to a very few verbs, most of which have alternative valencies or
enough synonyms to allow their avoidance. Typical examples include sich erinnern
‘remember’, which can take a genitive object, but more commonly takes a prepositional
object with an, or bedürfen ‘need’, which has an almost perfect synonym in the transitive
2.2.2. The dative relation
In the plural, the dative is the only case with a morphological mark of its own (-(e)n). In
the singular, neuter nouns have -e/Ø, while masculine nouns have either -eor -(e)n,
according to declension class. The two allomorphs differ markedly in their stability. The
-en allomorph is fairly stable. Dem Herr ‘to the sir’ is very substandard and characteristic
of child language. Yet, as in the instance of the loss of the genitive -s, omission of the
dative mark is encouraged if the noun has postnominal attributes; cf. E4.
E4. Flucht vom Planet der Affen [German movie title, 1971]
‘Escape from the Planet of the Apes’
The -e allomorph has all but disappeared from the colloquial and even non-literary written
language. Most nouns do not admit it at all (see the list in Heidolph et al. 1981:598), and
those that do admit it normally do not carry it: dem Manne/Hause/Gewichte ‘to the
man/house/weight’ are stilted. The suffix survives only in some phraseologisms such as
am Tage ‘in the daytime’, zu Hause ‘at home’. Fossilization is a common symptom of
grammaticalization in its terminal phase.
Unlike the English and Latin-Romance dative cases, and unlike the German, English and
Latin-Romance genitive cases, there is no single preposition taking the place of the dative
inflection. The following verbs can all take simple indirect objects, but alternatively allow
prepositional objects as indicated: senden/schicken an ‘send’, weiterleiten an ‘forward’,
genügen für ‘suffice’. Other verbs taking indirect object have derived counterparts taking
direct object: (be-)drohen ‘menace’, (be-)dienen ‘serve’, (be-)raten ‘advise’,
(an-)schreiben ‘write’. Earlier yet, experiencer verbs which used to take the perceived
entity as the subject and the experiencer in the dative changed their valence, taking the
experiencer as the subject and the perceived entity as the direct object: mir ahnt > ich ahne
‘I forebode’, mir träumt > ich träume ‘I dream’. On the other hand, many verbs such as
geben ‘give’, fehlen ‘be absent’, gehorchen ‘obey’, have a stable indirect object
construction without an alternative. Outside of verbal valency, the dative is mostly
Christian Lehmann
substitutable by prepositions, especially für ‘for’. Thus, ich kaufe ihm ein Buch ‘I buy him
a book’ = ich kaufe ein Buch für ihn, das klingt mir merkwürdig ‘that sounds weird to me’
= das klingt merkwürdig für mich. It would be a rewarding study to disentangle and then
recombine change in inflection paradigms, in syntactic relations and in semantic relations
in this area.
2.2.3. Loss by grammaticalization
As I have argued in earlier publications (e.g., Lehmann 1985), it is futile to ask whether
the grammaticalized prepositions oust the older case suffixes or whether, instead, the
attrition of the case inflection attracts the prepositions to fill the gap. It seems more
appropriate to view adpositions and case affixes as being at two stages of the same scale
and both being grammaticalized simultaneously and in parallel.
It has long been clear that one of the most palpable symptoms of grammaticalization is
reduction on both the semantic and the phonological sides of the language sign. The
logical endpoint of this is complete loss. However, not every case of loss of a grammatical
element can be regarded as purely a matter of grammaticalization. This is also true of the
reduction of case inflection. Both the examples of the genitive and of the dative show that
factors at very different levels contribute to this reduction. At the level of the
grammaticalization scale which includes adpositions and case affixes, it is, loosely
speaking, a question of fulfilling a given function with an alternative device. If, however,
case forms disappear because there are no longer any verbs which require or even admit
them, it is rather a question of no longer fulfilling a given function. To the extent that this
is true of German genitive and dative inflections, their fate is not exclusively one of
2.3. Complex prepositions
The formation of new prepositions through the combination of a (mostly relational) noun
with either a preposition or a case suffix is one of the most common grammaticalization
processes in the world. It has been done in German, as in other Indo-European languages,
for millennia. Since about 1975, there has been a new wave of complex prepositions. The
most common of these are the following:
im Zuge ‘by, during, in’,
im Wege ‘by (way/means of)’,
im Vorfeld ‘on the eve (of), before’,
im Anschluß ‘following, after’,
im Gefolge ‘in the sequel (of), after’,
im Verfolg ‘in the course (of)’.
They are mostly promoted in bureaucratic and journalistic jargon; none of them are really
colloquial. Here are some illustrative examples:
Grammaticalization and related changes in contemporary German
6One that comes to mind is auf der Basis ‘on the base’, which is parallel to aufgrund. However, this
is not as recent, and, moreover, it does not yet seem to be a preposition.
E5. Im Zuge der Regierungsumbildung wurden zwei Ministerien zusammengelegt.
[example constructed on a common pattern]
‘In the course of the reshaping of the government, two ministeries were conflated.’
E6. wenn die Bußgeldstelle im Wege der automatischen Datenverarbeitung arbeitet
[radio interview, 1987]
‘if the fines department works by means of automatic data processing’
E7. Im Vorfeld seiner Reise hatte der pakistanische Außenminister vor einigen Tagen
erklärt, ... [radio news, 1987]
‘On the eve of his journey, the Pakistan Secretary of State had declared some days
ago ...’
E8. Im Anschluß an die Nachrichten der Wetterbericht. [daily news]
‘After the news, the weather forecast.’
E9. Die vorliegende Untersuchung ist die revidierte Fassung einer Arbeit, die im
Gefolge eines Seminars über ‘Sprachliche Variation und Diversität’ entstand ...
[linguistic paper, 1981]
‘The present study is the revised version of a paper that originated in consequence
of a seminar on "Linguistic variation and diversity"’
E10. Im Verfolg der Untersuchungen ergab sich, ... [scientific article, 1989]
‘In pursuance of the investigations, it turned out ...’
The following features of this set are noteworthy:
1. All the cases involve a metaphor; the respective literal meanings are as follows:
im Zuge ‘in the train/procession (of)’,
im Wege ‘in the way (of)’,
im Vorfeld ‘in the fore-field (of)’,
im Anschluß (an) ‘in connection (with)’,
im Gefolge ‘in the suite (of)’,
im Verfolg ‘in pursuance (of)’.
2. There is a common formal pattern here: The basic nouns are bi- or trisyllabic, with final
or penultimate stress. They combine with the form im (elementary locative preposition in
plus definite article) to yield a complex preposition which (except for im Anschluß)
requires the genitive of its complement. The pattern has been productive for a certain
period (since about 1975, I guess). Earlier, such complex prepositions as aufgrund ‘on the
basis’, anhand ‘with the aid’, and even the postposition zufolge ‘according to’ came into
usage. I am not aware of any formation parallel to the latter in the last years.6 That the
above is a currently productive pattern is confirmed by an additional consideration: What
is meant by im Wege would normally be expressed by auf dem Wege. Weg after the
Christian Lehmann
7All of my examples stem from independent discourses. They do not reflect any idiolectal
idiosyncrasies or patterns evolved during a particular discourse.
8In this paper, I am using ‘phraseology’ and related terms in the broad sense defended in, e.g., Pilz
(1983). This includes relatively fixed collocations without regard to their semantic regularity, and also
collocations of non-lexical elements.
preposition in occurs exclusively in this complex preposition. There appears to be
analogical pressure at work.
3. An examination of my data reveals a feature that so far seems to have gone unnoticed
in research on grammaticalization: the complex prepositions arise in a narrowly
circumscribed set of contexts. The complement of im Zuge and im Verfolg is a verbal
noun, normally an action noun. The complement of im Wege is a new technology in all
three recorded examples; in two of them it is automatic data processing.7 The complement
of im Vorfeld is a social activity, often some sort of meeting. The complement of im
Gefolge is usually a trend-setting, preferably scientific activity. This is a preliminary hint
that phraseology 8 is at work in grammaticalization: When a newly coined periphrastic
expression is received by other members of the speech community, it will not be in
isolation, but in the context in which it was originally coined. It will not then spread at
once to all kinds of contexts which, given the rules of grammar, would admit it, but will
initially be restricted to certain collocations which come close to being phraseologisms.
The interplay of grammaticalization and lexicalization can also be seen inside each of the
complex prepositions. First of all, im Zuge and im Wege show the dative ending -e which
is falling out of use in free syntax (cf. §2.2.2), but survives in phraseological contexts.
Second, the typical symptom of lexicalization, namely a disturbance of compositionality
by either semantic or formal changes, is again observable in im Wege, since its internal
structure does not obey the respective subrule of syntax. This is why complex prepositions
are termed ‘prepositional phraseologisms’ in Burger et al. 1982:23. If they remain in the
language, one day each of them will become an entry in the German lexicon, just as the
earlier auf Grund has now been entered as aufgrund.
4. According to their external syntactic behavior, the six complex prepositions fall into
two groups: those that are used exclusively as prepositions (i.e., they require a
complement), and those which allow adverbial use. Im Zuge and im Wege belong in the
former group, im Vorfeld and im Anschluß belong in the latter. E11 and E12 illustrate the
E11. ... ist die Verfahrensweise der LABK auch darauf gerichtet, im Außenverhältnis
fachaufsichtliche Interventionen grundsätzlich im Vorfeld abzuwehren.
[administrative circular, 1987]
‘the procedure of the LABK is also directed, in its external relations, principally
Grammaticalization and related changes in contemporary German
9The style of the translation mirrors the original one.
10 Seiler's (1973) interpretation of the periphrasis involving ‘have’ and ‘be’ auxiliaries as expressing
the configuration ‘possessor of an act’ would seem to be a case in point.
11 apart from the fact that adverbial im Wege would interfere with the idiom im Wege sein ‘to be an
towards precluding in advance interventions of disciplinary supervision.’9
E12. Mehr darüber gleich im Anschluß. [radio lead-in, 1989]
‘More on this directly hereafter.’
E12 should be compared with E8. The rhetorical purpose of E11 is apparently to avoid
being precise on what exactly might provoke an intervention. It may be remarked at this
point that, in general, periphrases do not seem to be used for the sake of greater precision.
The law of iconicity according to which more expression generally corresponds to more
content here appears to be suspended: the periphrases add vagueness by being intercalated
into, and thus dilating, relations that would otherwise be direct.10
E13 shows im Gefolge without a complement.
E13. Sie haben also auf Edelmann und in dém Gefolge auf Dieckmann hingewiesen.
‘you have referred to Edelmann and, in consequence of that, to Dieckmann.’ [radio
interview, 1987; a research tradition is being referred to.]
In dém Gefolge must mean either ‘thereafter’ or ‘in the succession of Edelmann’. To the
extent that grammaticality judgements are possible in the case of "trendy" phrases, E13
is probably ungrammatical in regard to the variety to which it belongs. I have no example
of im Gefolge or im Verfolg in simple adverbial use.
Taken by themselves, the lexical nouns Weg and Zug would probably be classed as
non-relational, while Vorfeld, Anschluß, Gefolge and Verfolg appear to be relational. The
first guess would therefore be that the former set could be used adverbially, the latter not.
The data point in the opposite direction. A possible explanation might be that only a
relational noun gives the information that reference to something else is implied, whereas
adverbial im Zuge, im Wege would not link up with the semantic context in this way (cf.
Seiler 1983).11
2.4. von ... her and other limitative constructions
2.4.1. By far the most fashionable of those grammatical constructions which are unknown
to normative grammar is the von-her construction. It has been in everybody's mouth since
about 1980, and its new function seems so natural to speakers that most of the people
whose attention I drew to its "trendy" character were completely unaware of it.
Before the change, the relevant situation was roughly as follows: The discontinuous
preposition von ... her (preposition von ‘from’ plus deictic particle her ‘hither’) had the
Christian Lehmann
strictly concrete-local meaning ‘(coming) from’, as in E14.
E14. Von draußen vom Walde komme ich her. [folk poem]
‘I am coming from outside, from the woods.’
E15 shows a slight metaphorical extension in an adverbial phrase at the sentence level.
E15. Er war vermögend, schon von den Eltern her, besaß zudem ein Herrenmodegeschäft
... [Schöne Welt, 1987]
‘He was already affluent from his parents, and in addition owned a gentlemen's
fashion shop.’
The discontinuous preposition was also commonly used in a set of metaphoric
phraseologisms centering around the notion eine Sache von einem Standpunkt her
betrachten ‘to view a matter from a standpoint’, such as von X her
gesehen/betrachtet/beurteilt, etc. ‘seen/viewed/judged from X’. Such phrases were and
are used as sentence-level adverbials to express such notions as ‘with reference to, in/with
regard to/of, with respect to, as regards, considering X’. Other devices to express this have
been such clumsy circumlocutions as was X betrifft, im Hinblick auf X, hinsichtlich X, but
also simpler constructions with in ‘in’ and nach ‘after’. We will return to these in §2.4.2.
The phrase von X her gesehen/betrachtet is already fairly fixed and depleted. Otherwise
it would not be possible to express the same semantic relation to X either in an ablative
(von X her gesehen) or in an allative (im Hinblick auf X, hinsichtlich X) metaphor. The
first and decisive step in the recent development of von-her is the omission of the past
participle of a verb of viewing which formerly closed the adverbial. E16 is an example
which might as well contain such a participle (e.g., von heute her gesehen) and would then
illustrate the typical use of the latter.
E16. Solche Sätze wirken von heute her nichts weniger denn kabarettistisch überzogen.
[Die Zeit, 1986]
‘Such sentences, viewed from today's standpoint, seem anything but cabaretishly
The omission of the participle is a remarkable feat, since it is the head of the phrase. It
seems natural to assume that constructions such as the one in E15 served as a formal
model. At this stage, we have a discontinuous preposition which expresses a limitative
relation between a proposition and a factor which restricts the scope for which the
proposition is valid. E17 is a very early example.
E17. Auch ist die Zahl der bisher in Linear A aufgefundenen Inschriften nicht groß (etwa
300) und ihr Textumfang zumeist recht gering, desgleichen lassen sie von ihrem
Inhalt her (Abrechnungsquittungen, kurze Weihaufschriften) keine breitere
Erfassung der Sprache erwarten. [linguistic book, 1970]
‘Also, the number of the inscriptions discovered so far in Linear A is not large
(around 300), and their text length is, for the most part, rather small; likewise,
considering their content (clearing receipts, short votive inscriptions), they do not
Grammaticalization and related changes in contemporary German
make one expect any broader conception of the language.’
E18 - E23 show how much the initial meaning of von ... her has broadened.
E18. Der ist nicht der allerschönste, den ich vorführen kann, aber von der Tendenz her
wird sicher klar, was gemeint ist. [free speech, 1987]
‘That is certainly not the most beautiful one that I can demonstrate; but the tendency
of what is meant will surely get clear.’
E19. Wenn der Minister in dieser Sache Wort hält, kann ich nur sagen, daß wir von der
Zielrichtung her einer Auffassung sind. [radio interview, 1987]
‘If the minister keeps his word in this matter, I can only say that we are of one
opinion as regards the goal.’
E20. Also von den Bildern her, mir pers'önlich sagt der nich so zú, der Macke.
[conversation at a Macke exhibition, 1987]
‘These pictures, me personally Macke does not appeal to that much.’
E21. [Unter dem Fuß des Rechners] duckt sich der vorderste Teil der Mutterplatine. Von
der Breite her schließt er schön ordentlich mit der Querleiste der Tastatur. [computer
magazine, 1986]
‘Under the foot of the computer is located the foremost part of the motherboard. In
its width, it is exactly flush with the cross-bar of the keyboard.’
E22. [Ad ‘move alpha’:] Nun, dieser Ausdruck ist ein zeitlicher und suggeriert sozusagen
vom Ausdruck her, daß da etwas passiert. [free speech, 1987]
‘Well, this expression is a temporal one and implies, you know, by this expression
that something is happening here.’
E23. A: Treffen die hauptsächlich auf Substantive zu? - B: Ja, [C: Auf Prädikatoren.] -
also von den Wortarten her auf Prädikatoren. [discussion, 1987]
‘A: Do those apply principally to nouns? - B: Yes [C, interrupting: to predicators]
- well, in regard to word classes, to predicators.’
Some of these examples show that the ablative meaning of the metaphorical base is
completely lost. This is particularly clear in E19, which features von der Zielrichtung her
‘from the goal-direction’. Others show that von ... her is used as a convenient device for
introducing concepts which are somehow felt as relevant but whose exact relation to the
point is not (to be made) clear. This can be seen in E23. Most important, however, are the
roles of von-her phrases as contrastive topic (in the sense of Chafe 1976), as simple topic
and as shadow actant. These will now be taken up in turn.
A contrast between the NP in the von-her phrase and a concept in the context can already
be seen in E17. E24 and E25 are unmistakable examples of von-her phrases functioning
as contrastive topics.
E24. Es ist diese und nur diese Dimension, die seine Kunst als subversiv vorkommen läßt
und indizierende Saubermänner auf den Plan ruft. Vom bloßen Wortlaut her sind die
Christian Lehmann
Formulierungen nämlich eher von der verbindlichen Art. [Die Zeit, 1986]
‘It is this and only this dimension [of interactive speech] that makes his art appear
subversive and that evokes indexing expurgators. For, as regards the wording alone,
the formulations are rather of the diplomatic kind.’
E25. Herbert Hupka ist auf Ceylon geboren. Die ersten drei Lebensjahre verbrachte er in
Australien, aber von Herkunft, Elternhaus und Schule her sind Ratibor und
Oberschlesien seine Heimat. [book, 1980]
‘Herbert Hupka was born in Ceylon. The first three years of his life he spent in
Australia, but as for origin, home and school, Ratibor and Upper Silesia are his
The use of von ... her to mark simple topic can be regarded as a further step in its
depletion. Relevant examples are E21 and E23. By far the most common phrase of this
type, and the most common use of von ... her in general, is the phraseologism von daher,
as in E26.
E26. ... Also von dáher weiß ich nicht, was das sóll. [from a discussion, 1986]
‘In view of this, I wonder what that is supposed to be (good for).’
Von daher is a sentence introducer that links the sentence with the preceding context. The
nature of the suggested relation is as vague as in the other cases.
Loss of communicative force in the sense of functional sentence perspective has often
been regarded as a factor in grammaticalization, above all in the various hypotheses about
a development from topic to subject. In the present connection, it suffices to realize that
when the communicative function of an element fades, its syntactic cohesion with the
context tends to tighten. The limitative force of the von-her phrase then relates directly to
a particular sentence actant. In a sense, it represents that actant at a different syntactic
level. A shadow actant is, then, a syntagm which bears a limitative semantic relation to a
certain actant. Semantically, it is part of that actant or on equal footing with it as regards
the predication. It does not bear itself the syntactic function of that actant, but a small
transformation, with no change in voice and little change in meaning, could put it into that
actant's position. E18 and E20 show what is meant more effectively than further
Given both the general prominence of the subject over the other actant functions and the
natural affinity between subjects and topics, it becomes intelligible that the shadow subject
is by far the most common shadow actant function of von ... her. Typically, the syntactic
subject is either empty, as in E18, or it bears a relation of factual connection to the NP in
the von-her phrase, e.g. a possessive relation in E17, E21 and E24, a source relation in
E20. An extreme example of this kind is E22, in which the syntactic subject is repeated
in a von-her phrase. An example of a shadow object will be found below.
Concomitant with the development from contrastive topic via simple topic to shadow
actant, we find an increasing integration of the von-her phrase into the clause. In some of
Grammaticalization and related changes in contemporary German
12 Given that her is not a postposition, the logical alternative of dropping the first part is unavailable.
the examples, especially E17, E19 and E22, it is no longer a sentence-level adverbial, but
well integrated into the clause. It may be recalled that decrease of structural scope is a
symptom of grammaticalization (cf. Lehmann 1982, ch. IV.3.1).
Although German allows a variety of discontinuous constructions, it does not exactly
favor them. One may therefore wonder whether discontinuity would not be an obstacle in
the grammaticalization of an element. Actually, the second part of von ... her is dropped
more and more often,12 so that the result is the simple preposition von, which is in the
center of the system of grammatical relations, anyway. Here are some examples:
E27. Lisp ist nach FORTRAN die zweitälteste Programmiersprache, von der
Sprachkonzeption aber völlig anders. [computer magazine, 1987]
‘Lisp is, after FORTRAN, the second-oldest programming language, but completely
different in its language-conception.’
E28. weil sie eben in Relation zu den anderen Kandidaten von den Vorleistungen nicht
das aufzuweisen hatte [from a discussion, 1986]
‘because, in relation to the other candidates, she did not have that much to offer from
her previous performance’
E29. Die Gesellschaft ist inzwischen auf über 800 Mitglieder angewachsen, und das ist
natürlich von der Zielsetzung der DGfS äußerst erfreulich. [free speech, 1987]
‘The society has meanwhile grown to over 800 members, and this is, of course,
highly gratifying, in view of the goal of the DGfS.’
E30. besonders raffiniert von der Farbe [sticker at a dress in a show-window, 1987]
‘particularly sophisticated in its color’
E31. Man müßte nun für jeden dieser zusammengesetzten Datentypen ... Prozeduren
schreiben, was allein vom Schreibaufwand sehr aufwendig wäre. [programming
handbook, 1984]
‘One would now have to write procedures for each of these complex data types,
which would be very expensive considering the writing expenditure alone.’
E27 exemplifies the contrastive topic, E30 the shadow subject (the color of the dress is
sophisticated), E31 a limitative relation which is close to the shadow subject. E28 shows
a particularly high degree of the integration of the von (- her) phrase, which in this case
yields a shadow object; an adequate paraphrase would be weil sie eben in Relation zu den
anderen Kandidaten nicht die Vorleistungen aufzuweisen hatte ‘because ... she did not
have that previous performance to offer’. Incidentally, a comparison of E29 with E19
confirms the point made in section 2.3 about the phraseological locus of many newly
grammaticalized elements.
2.4.2. The examples are from all kinds of non-literary texts, from informal talk via journal
articles to scientific books. There can be no doubt that (quite independently from the use
Christian Lehmann
13 Moreover, E30 shows the extent to which the new function of von tends to oust older ones. Ten years
ago, this expression would have been interpreted exclusively, with stress on raffiníert instead of on Fárbe,
as implying that the color has achieved something sophisticated; cf. Wie raffiniert von dir! ‘How clever
of you!’.
of a "trendy" expression) several of them evidence unpremeditated, negligent talk or bad
style, e.g. E18, E20, E21, E22, E23, E30,13 E31. To complete the picture, we should
ascertain that the alternatives to von (- her) are alive, although they are not used as
spontaneously as their competitor. Those that I have data for are hinsichtlich ‘in regard to’
(E32), gemessen an ‘measured against’ (E33), in ‘in’ (E34) and nach ‘after’ (E35).
E32. Der mit 12 MHz getaktete ARC Turbo 12 kann sich hinsichtlich seiner Leistung mit
dem PS/2 Modell 60 von IBM messen. [computer magazine, 1987]
‘The ARC Turbo 12, tacted with 12 mhz, can compete with the PS/2 model 60 of
IBM in performance.’
E33. Der neue Tischcomputer aus Texas stellt nämlich, gemessen an Preis und Leistung,
alles bislang Dagewesene in den Schatten. [Capital, 1986]
‘The new desk computer from Texas indeed overshadows, in regard to price and
performance, everything that has appeared up to now.’
E34. Wir sind ja doch eben gerade in der Téchnik, von der Téchnik her, wesentlich
weiter. [radio interview, 1987]
‘For we are substantially more advanced particularly in technology, viewed
E35. Noch ein Wort dazu, was Beispiele ihrer Funktión nach sein können. [free speech,
‘Another word on what examples can be as to their function.’
In each of these examples, the phrase in question could easily be substituted by a von-her
phrase, with no change in meaning. E34 is particularly intriguing as the speaker offers the
paraphrase himself, apparently thinking that the received construction does not sufficiently
bring out his point. Compare, moreover, E35 with E17 and E32/E33 with E28. If von
(-her) should be further grammaticalized, then the above alternatives will probably
survive, as they allow the expression of more specific limitative relations.
It is very hard to prognosticate where this development can lead. On the one hand, it is
probably no accident that it is just the preposition von which acquires the function of
shadow subject. It is well established in the function of marking the passive agent, which
probably has semantic affinities with the shadow subject as illustrated above. In view of
this, there is little to prevent German from getting double subject constructions just as
Japanese. On the other hand, it might be an obstacle for such a development that von
already has so many functions, in particular the ablative and possessive functions, so that
it might prove undesirable to freight it with an additional one.
Grammaticalization and related changes in contemporary German
14 Thus, it is not possible to say wenn nicht so viel Salz drin wäre, möchte ich das wohl ‘if there were
not so much salt in it, I would like it’.
2.4.3. Finally, yet another limitative construction should be mentioned, which I have
singled out because it is a word formation process. It combines the topic (or shadow
actant) noun with the complex derivational suffix -mäßig ([[x]N-mäßig]Adj ‘by the measure
of x’) and thus converts it into an adjective which, for the present purpose, is used
adverbially. This construction alternates to a large extent with the von-her construction
and is insofar synonymous with it. Examples can be derived from E28 by replacing von
den Vorleistungen by vorleistungsmäßig, or from E32 by replacing hinsichtlich seiner
Leistung by leistungsmäßig.
This process is fashionable, too. However, as is typical in word-formation, it is restricted
to single-noun topics which, in one of the alternative constructions, would be
accompanied at most by a definite determiner. The complex suffix -mäßig is in a paradigm
with a few other adjective-forming suffixes such as (N)-förmig ‘of the form (of N)’,
(N)-haltig ‘containing (N)’ and (N)-artig ‘of the kind (of N)’. These are currently
productive and illustrate the development of new derivational suffixes from formerly free
lexical morphemes. The phenomenon has been much investigated in German linguistics
(cf., e.g., Inghult 1975). It illustrates the transition from compounding to derivation. It
might be subsumed under grammaticalization. This would entail that derivation is more
grammatical than composition. The latter problems are discussed in Lehmann 1989. Both
these and the question of the function of all the devices mentioned in this section need
much more investigation.
3. Verbal categories
3.1. Modal verbs
There is a long tradition in German of creating verbal categories by embedding a
non-finite form of the full verb in a periphrasis with another verb which then degenerates
to an auxiliary. The most recent of these which is well established as an auxiliary, but still
expanding, is würde ‘would’, which forms an analytic subjunctive which can substitute
for the synthetic past subjunctive. The relationship between the two is very similar to the
one between the synthetic genitive and the analytic von construction (cf. §2.2.1).
Normative grammar recommends the use of the synthetic form wherever possible, but the
lower the level of formality, the more the analytic form prevails. Also, there are cases
where the synthetic form is insufficient and thus impossible. For example, originally
möchte was the past subjunctive of mag ‘like’. Now, however, this category must be
paraphrased by würde mögen, because möchte has lexicalized into a polite form of will
‘want’.14 I hypothesize that the grammatical necessity of using an analytic construction
instead of a competing synthetic one is a safe hint as to the direction in which the
development goes. However, on the whole there have been no revolutionary developments
Christian Lehmann
15 I am told that it has existed in Pennsylvania Dutch since the 17. century. There it seems possible that
a syntactic loan has met with support by a predisposition in certain German dialects.
in our century as regards the expansion of würde.
More recent is the development of brauchen ‘need, require’ into a modal verb. As this is
well studied by now (cf., e.g., Wurzel 1984:149), a few remarks will suffice here. The
established modal verbs, such as wollen ‘will’, können ‘can’, müssen ‘must’ and a couple
of others, form a very homogeneous group. They differ from lexical verbs in such respects
as being defective as to verbal categories, having different personal endings and taking
their infinitive complement without zu ‘to’. All of these are common symptoms of
In standard German, the verb brauchen belongs in the class of verbs which take an
infinitive complement with zu, as in E36.
E36. Das beabsichtigt/braucht er nicht zweimal zu sagen.
‘He does not intend/need to say that twice.’
In colloquial German and various dialects, it is making its way into the group of modal
verbs by adapting itself semantically and formally to its constraints (see Lehmann 1982,
ch. IV.4.2.2 on paradigmaticization). Semantically, it functions as the counterpart of
müssen in negative contexts, as illustrated in E37.
E37. a. Das muß er zweimal sagen.
‘He must say that twice.’
b. Das brauch er nicht zweimal sagen.
‘He need not say that twice.
Incidentally, it is not quite clear why this should happen, since das muß er nicht zweimal
sagen means the same as E37.b. Anyway, if the former alternative did not exist, brauchen
would be, as part of the paradigm of modal verbs, in a suppletive relation with müssen.
The formal adjustment of brauchen involves loss of the final -t in the third person singular
present and omission of the zu before the infinitive; cf. E37.b with E36. It is a safe bet that
brauchen will become a fully integrated modal verb if it survives.
3.2. Progressive aspect
Other European and Non-European languages have a progressive aspect of the type I am
working, estoy trabajando or even je suis en train de travailler. Standard German, which
is, perhaps, one of the poorest languages as regards the category of aspect, has nothing of
the sort. However, in a couple of West German dialects, in particular Lower Frankish
(spoken, e.g., in Cologne), a progressive aspect has come into use.15 This is expressed by
a periphrasis which involves sein ‘be’ plus the infinitive, which is governed via the
preposition an ‘at’ fused with the definite article of the infinitive to yield am. E38 is a
Grammaticalization and related changes in contemporary German
16 I did not gather data on the progressive aspect because it is too familiar in my dialect environment.
17 This point has been clarified to me by Ulrike Claudi.
typical and, at the same time, the earliest example I know of.16
E38. Ich bin am Arbeiten. [letter by August Macke, 1908]
‘I am working.’
There are a couple of possible models for this construction. On the one hand, there is the
gerund of the type ich habe mich beim Arbeiten verletzt ‘I have hurt myself at work/while
working’, which also features the infinitive depending on a preposition fused with the
definite article. This can also occur as the predicate of a copula sentence: ich bin beim
Arbeiten ‘I am at work/working’. On the other hand, there is the modal expression ich
habe zu arbeiten ‘I have to work’, which features the infinitive depending, via the
preposition zu, on an auxiliary verb and thus forming a periphrastic verb form. Except for
the preposition an itself, which is not otherwise used in standard German for the
embedding of non-finite verbs, all the structural features of the progressive construction
are well established in standard German. It is currently spreading into the colloquial
language of Northwest Germany. The interesting feature about the relative chronology of
the spread is that intransitive verbs allow the progressive aspect much more easily than
transitive or other plurivalent ones. Thus, while many speakers would accept E38 at the
colloquial level, examples like E39 are still rare outside of the original dialect area.
E39. Ich bin jetzt dieses Orientierungs-Papier am überarbeiten. [Colonese speaker, 1988]
‘I am currently reworking this orientation paper.’
As E39 shows, the internal syntax of the verbal group in the progressive follows the
pattern of other periphrastic verbal categories: auxiliary verb in second position, then all
the complements and adjuncts (in partly variable order), finally the non-finite form of the
full verb. The preposition-plus-article portmanteau morph am forms a constituent with the
non-finite verb form.
There is an alternative construction, exemplified in E40.b.17
E40. a. Martha ist Flaschen am sortieren.
‘Martha is sorting out bottles.’
b. Martha ist am Flaschen-Sortieren.
‘Martha is bottle-sorting.’
Here the complement comes between am and the non-finite verb form. This construction
is obligatory if the direct object is indefinite and non-specific; cf. B41.
E41. Jupp ist am Radfahren / *Rad am fahren.
‘Joe is bicycle-riding.’
Again, examples like E40.b and E41 are more widely accepted than those like E40.a. They
have to be analyzed as instances of incorporation. This is in consonance with the above
Christian Lehmann
18 The surface structure would have to bring out the fact that the preposition and the article form a
observation that the progressive is much more common with intransitive than with
transitive verb constructions: the incorporation construction is treated like an intransitive
If one wants an explanation for the order in the spread of the progressive, there is a
structural and a semantic factor that can be adduced. First, the relevant shallow 18 syntactic
structure of constructions such as those in E39 and E40.a is as in F1.
F1. Progressive of syntactically complex verbal
| |
| ________|_________
| | |
| | _______|_______
| | | |
| | | |
[Complements and Adjuncts] Preposition Definite Article Non-finite Verb
This means that the material depending on the non-finite verb form is outside the scope
of the definite article which determines it. This is non-iconic and unprecedented in
German syntax.
Second, the semantic locus of the progressive aspect is in intransitive verbs. Progressivity
and intransitivity have a natural affinity (see, e.g., Hopper & Thompson 1980), as the
progressive aspect centers the attention on the situation of the subject. Given such
conditions, one would predict that if a grammatical category is introduced into a language,
it will start from its natural locus and only later penetrate into less prototypical domains.
This is borne out by the diffusion of the German progressive aspect. In the meantime, it
is intriguing to observe the conspiracy of the structural and the semantic factors, which
are completely independent.
3.3. Andative and venitive
Motion deixis is well established in German in the form of the particles hin/her ‘away
from/towards the deictic center’. These combine freely with motion verbs; cf.
hinschwimmen ‘swim towards a point outside the deictic center’ vs. herschwimmen ‘swim
towards the deictic center’. Their combination with other verbs, however, essentially leads
to lexicalization, with the deictic feature being lost. Cf., from sagen ‘say’, hinsagen ‘say
carelessly’ and hersagen ‘rehearse’, or, from machen ‘make’, hinmachen ‘hurry up’ and
hermachen ‘produce an impression’. The idea of the grammatical categories andative vs.
venitive, however, is that any verbal action can be characterized as having its origin in the
deictic center (and, thus, being directed away from it) vs. having its origin outside the
deictic center (and, thus, preferably being directed towards it). Cf. the Italian periphrastic
Grammaticalization and related changes in contemporary German
19 Leaving the direct object unpromoted is not standard and uncommon at the nonstandard level.
20 Its narrowest locus is, of course, in constructions which are close to its etymological origin, such as
sie bekam ein Buch geschenkt ‘she was presented a book’, etymologically ‘she got a book presented’.
passive with andare vs. venire as a paradigm example.
Again in the colloquial language of Northwest Germany, there has been an andative for
quite some time. It is formed by coordinating the full verb with hingehen ‘go away from
the deictic center’, as in E42.
E42. Nun könnte man ja hingehen und das Ganze nochmal schreiben.
‘Now one [including the deictic center] could go right ahead and write the whole
thing again.’
The range of functions of this construction is not yet understood. I realized its andative
function only when I first heard its venitive counterpart, as in E43.
E43. Was wir nun nicht verstehen, ist, warum man nicht herkommt und das
schutzgeimpfte Tier dem antibakteriell positiven Tier gleichstellt. [radio interview,
‘What we now do not understand is why one [i.e. persons outside the deictic center]
does not put the protectively inoculated animal on a par with the antibacterially
positive animal.’
Herkommen und, as in E43, is much rarer than hingehen und. Its venitive function,
however, is unmistakable. Lacking sufficient data, I cannot say more about this.
3.4. The addressee passive
German passives demote the active subject and signal that whatever is the subject is not
the agent. The standard passive involves the use of werden ‘become’ and sein ‘beas
auxiliaries. If applied to transitive verbs, it involves the promotion of the active direct
object to subject.19 The colloquial language has, in addition, an addressee passive (see
Heidolph et al. 1981: 556f). This involves bekommen/kriegen ‘get’ as auxiliaries. Its locus
of application is ditransitive verbs with an indirect object, as in E44.20
E44. a. Die Arbeiter zeigten ihr den ganzen Betrieb.
‘The workers showed her the whole operation.’
b. Sie kriegte (von den Arbeitern) den ganzen Betrieb gezeigt.
‘She was shown the whole operation (by the workers).’
The active indirect object is promoted to subject. Other complements remain unaffected.
Just as in the standard passive, the active subject can be adjoined in a von-phrase.
The construction is currently expanding in the following directions:
Christian Lehmann
21 Kriegen and bekommen, in- or outside the periphrastic passive, are basically register variants of each
1. The variant employing bekommen penetrates into the standard language.21 Thus, a
version of E44.b with bekam instead of kriegte appearing in the radio news is not
2. The addressee need not be an indirect object. NPs in benefactive or possessive
functions, as long as they are datives in the active version, can also be promoted, as in
E45. Sie kriegte den Wagen repariert.
‘She got the car repaired.’
3. The verb does not need to have a non-dative complement. The addressee passive is
becoming possible with verbs that take only an indirect object.
E46. Sonja bekommt immer von euch geholfen. [conversation, 1987]
‘Sonja always gets helped by you.’
4. The complement to be promoted does not even need to be in the dative. Other syntactic
functions are allowable, as long as they express something like an addressee semantic
function, as in E47.
E47. Dann kriege ich immer geschimpft. [conversation, 1987]
‘Then I always get scolded.’
E46 and E47 are observed data. Speaker judgements vary on such examples; they are
outside my own idiolect. Steps 2 - 4 of the expansion nicely illustrate the dropping of
selection restrictions which stem from the etymology of the construction.
Restricted to the colloquial language, there is yet another periphrastic construction
involving bekommen/kriegen, with a meaning that Seiler (1973) calls ‘possessor of an act’,
as exemplified in E48.
E48. Kriegst du das bis heute abend geschrieben?
‘Can you (manage to) get that written by this evening?’
The construction is formally identical to the addressee passive. Here, however, there is no
implication that the subject cannot be agent. Instead, there is an attribution of the act to
the subject as the one who is responsible for it, so that he will either have to do it himself
or else see to it that it is done. Probably, this auxiliary function of bekommen/ kriegen is
based on (interference with) a different meaning of kriegen, namely ‘manage, succeed’.
3.5. Function verbs
A function verb construction (Funktionsverbgefüge, also [verbale] Streckform ‘stretched
[verbal] form’) is a syntagm consisting of a moderately grammaticalized verb and an
abstract verbal noun as its complement such that the syntagm bears a paradigmatic relation
Grammaticalization and related changes in contemporary German
to the verb underlying the verbal noun. The paradigmatic relation may consist of a valency
change or a (sometimes very slight) change in the aktionsart of the full verb. The
grammaticalized verb takes on those grammatical properties of the corresponding full verb
which can only be expressed by a finite (as opposed to nominalized) verb, including
subject and direct object government, while the verbal noun retains the lexical properties
of the full verb, including that part of its valency which is not affected by nominalization.
The grammaticalized verb is called a function verb (Funktionsverb). Function verb
constructions have been a favorite topic in German linguistics (cf. Helbig 1979, Heidolph
et al. 1981:431-442); thus I can be brief here.
Function verb constructions have their stylistic locus in bureaucratic, scientific and
journalistic jargon. Some function verbs are more grammaticalized and form highly
regular constructions; also, several function verbs are synonyms and thus mutually
exchangeable with a given verbal noun. Other function verb constructions are more
lexicalized and have the character of phraseologisms (cf. Burger et al. 1982:37); they may
not even have a simple verb corresponding to them. There are transitions between the two
groups. Thus, some of the highly regular function verb constructions such as Anwendung
finden ‘find application’, as in E52, are fairly common and to that extent approach the
status of phraseologisms. The whole strategy, thus, is productive more or less like a
derivative pattern. I will here concentrate on the grammaticalized cases. Some of the most
common function verbs are kommen ‘come’, gelangen ‘arrive, reach’, bringen ‘bring’,
finden ‘find’, erfahren ‘experience, undergo’, unterziehen ‘subject’, zuführen ‘bring to’,
vornehmen ‘execute’. Here are some recorded examples:
E49. eine Notationsvereinfachung, bei der nur mehr die Endeinheiten D und 1/2 zur
Darstellung kommen. [linguistic book, 1984]
‘a notational simplification in which only the terminal units D and 1/2 are
E50. Wir haben dieses Konzert aufgenommen und bringen es heute zur Sendung. [radio,
‘We have recorded this concert and are broadcasting it today.
E51. bis ein Stapelprogramm geladen und zur Ausführung gebracht wird [computer
magazine, 1987]
‘until a stack program is loaded and executed’
E52. Für Sendungen nach den österreichischen Zollausschlußgebieten ... finden die für
den Verkehr innerhalb des Bereichs der Deutschen Bundespost geltenden Gebühren
Anwendung. [mail information booklet, 1982]
‘For shipments to the Austrian Customs-Exclusion Territories, the fees which are
valid for traffic within the domain of the Deutsche Bundespost apply.’
E53. Einige Bilder erfahren auch eine neue Interpretation ... [book commercial, 1987]
‘Also, some pictures get a new interpretation’
Christian Lehmann
E54. Diese Regelungen sollen noch einer vollständigen Überarbeitung unterzogen
werden. [radio interview, 1987]
‘These regulations are yet to be subjected to a complete revision.’
E55. daß sie einer geregelten Bejagung zugeführt werden können. [hunting journal, 1987]
‘that they [the kinds of game] can be subjected to a controlled hunting’
E56. Herr N.N., der seinerzeit die Aufstellung des Glasbruchschadens vorgenommen hat
[business letter, 1988]
‘Mr. N.N., who at that time made the statement of the glass damage’
The most productive function verbs form morpho-syntactic paradigms. F2 presents the
commonest pattern (VN = verbal noun).
F2. Terminative function verb constructions
version active passive
simple full verb xsbj V yobj ysbj wird (von x) Vpass
function verb
xsbj bringt yobj zu VNdat ysbj kommt/gelangt (durch x) zu VNdat
xsbj unterzieht yobj VNdat
ysbj findet/erfährt (durch x) VNacc
These constructions all involve terminative aktionsart. As is apparent from F2, the
periphrasis of an active simple verb generally involves transitive function verbs, while the
periphrasis of a passive simple verb generally involves intransitive function verbs. Here
as elsewhere in the grammar of German and other languages, bringen is the
causative-transitive counterpart of kommen. The same lexical relationship obtains between
unterziehen and erfahren. The choice between the two function verb patterns in F2
depends on the degree of affectedness of the object of the simple verb;
finden/erfahren/unterziehen are appropriate with more strongly affected objects.
4. Intensification
Intensification is probably an area which continuously undergoes change in most
languages; people are never content with available means to express the importance of
what they have to say. Very much goes on here at the nonstandard level, where what was
a strong expression yesterday is antiquated today. However, these are chiefly lexical
changes. We will look at two subareas in which other languages have grammatical means,
so that there is a chance of the German phenomena being cases of grammaticalization.
Moreover, the two phenomena described in this section are typical of the sometimes stilted
register of journalistic prose.
4.1. mehr als
Grammaticalization and related changes in contemporary German
The history of modern mehr als ‘more than’ remains to be written. Probably the first and
decisive step in its development into an intensifier was a syntactic reanalysis. Cf. E57.a
and b.
E57. a. Das ist mehr als das Doppelte.
‘That is more than double.’
b. Das ist mehr als ausreichend. [computer magazine, 1987]
‘That is more than sufficient.’
In E57.a, mehr is the head of the phrase, als das Doppelte is some kind of complement.
E57.b can be analyzed analogously. But current usage presupposes its reanalysis with
ausreichend as the head of the phrase, mehr als as a modifier. Cf. Matthews 1981:156f
for similar reanalyses.
Phraseology again plays a major role. The collocation in E57.b was found in two computer
journals of 1987. Other common collocations are mehr als zweifelhaft/fraglich ‘more than
doubtful/questionable’. These cases all admit a literal interpretation. That is, mehr als
ausreichend could signify, on a scale reaching from wanting over insufficient, sufficient,
adequate to abundant, a point above ‘sufficient’; mehr als zweifelhaft could imply
‘improbable’, assuming that this is the next step on the scale of plausibility; and similarly
for all the other constructions. However, there are reasons for assuming that this is not the
intended interpretation. First, how do we know that mehr als ausreichend means ‘above
"sufficient" on the appropriate scale’, whereas mehr als zweifelhaft means ‘below
"doubtful" on the appropriate scale’? Second, there are collocations with adjectives such
that the literal interpretation does not make sense, as in E58 and E59.
E58. Ungeachtet solcher Erfolge sind aber die Aussichten mehr als gering, daß die
Resolution den angestrebten Erfolg ... erreicht. [written radio commentary, 1987]
‘Despite such successes, the prospects are more than dim that the resolution will
attain the desired success.’
E59. Mehr als deutlich führt dieses etwas ausführlich beschriebene Beispielprogramm vor
Augen, wie systemnah ein Software-Designer mit Modula-2 programmieren kann
... [computer magazine, 1987]
‘This example program, described in some detail, demonstrates more than clearly
how closely to the system a software designer can in fact program with Modula-2'
In E58, a literal interpretation would require a level above gering ‘low’, whereas the
intended interpretation is, if anything, a level below ‘low’. In E59, there is simply no
appropriate scale, so that mehr als deutlich cannot refer to a level above ‘clear’; it can only
mean ‘with particular clarity’.
This extension to adjectives for which only an intensifier interpretation is available marks
a step in the expansion of mehr als. Another step is its transfer to nouns, as in E60 and
Christian Lehmann
22 This is the headline of an extended commercial of an insurance company which also features a
trustworthy-looking representative. The headline is not taken up in the commercial text, so that I cannot
tell what it means. I take this to be a symptom of the depletion and the extremely phraseological character
of mehr als.
E60. Sein Versprechen ist mehr als eine Versicherung. [commercial, 1987]22
‘His promise is more than an insurance.’
E61. Wenn es aber gelänge, die "Verweildauer auf vier Tage auszudehnen", dann wäre
das mehr als ein Erfolg. [Die Zeit, 1988]
‘But if it were possible "to extend the period of stay [of tourists] to four days," that
would be more than a success.’
Before 1980 or so, mehr als ein Erfolg could only have meant, with the appropriate
intonation, ‘two or more successes’. In E61, it apparently means something like ‘a good
success’. Similarly, mehr als eine Versicherung in E60 is not supposed to mean ‘two or
more insurances’, but probably ‘a good insurance’.
The last step in the expansion of mehr als so far is its transfer to verbs, as in E62 - E64.
E62. Der Nachteil, daß das Fehlen eines Linkers die Entwicklung größerer Programm-
projekte erschwert, wird durch die hohe Übersetzungsgeschwindigkeit von
Turbo-Pascal mehr als aufgehoben. [computer magazine, 1987]
‘The disadvantage that the development of program packets is impeded by the lack
of a linker is more than compensated for by the high compilation speed of
E63. [der Nachteil] läßt sich durch die hohe Druckgeschwindigkeit, also ein
Mehrfach-Ausdrucken eines Dokuments, mehr als ausgleichen. [computer magazine,
‘[the disadvantage] can more than be compensated for by the high printing speed -
that is, a repeated printing of a document.’
E64. ... Vertrauensbeweis, den er mit seinen Forschungen mehr als rechtfertigte. [written
speech, 1987]
‘... proof of trust which he more than justified by his research’
Mehr als V means something like ‘V and do something in addition’, or even simply ‘fully
V’. Note again the phraseological character of the collocations in E62 and E63. The
development is so recent that one cannot yet say which lexical groups of verbs allow mehr
als as a modifier. What is certain is that only verb forms which are final in the verbal
syntagm allow it. This includes non-finite verb forms, as in E62 and E63, and verbs in
subordinate clauses, as in E64. This restriction is a consequence of the almighty
verb-second principle in independent declarative sentences, which allows no modifiers
before the finite verb. Thus, he more than justified that proof of trust has no direct
translation into German.
Grammaticalization and related changes in contemporary German
The historiography of mehr als as an intensifier would probably reveal that it is a loan
translation from English, in which more than is further advanced in its development. It is
not subject to ordering constraints; and it has a negative counterpart, less than, as in the
other co-editors have been less than responsive [letter, 1987], which does not have a
German equivalent. The latter point may be interpreted as showing that the English phrase
is paradigmaticized and, thus, grammaticalized at least to some extent. In the German
case, this is not so clear. Mehr als forms a phraseologism, but it is doubtful whether this
phraseologism is making its way into the grammar.
4.2. zunehmend
Continuous intensification of a property which is expressed by an adjective, as in ‘faster
and faster’, used to be expressed by immer ‘ever’, as in immer schneller. For concepts
expressed by verbs undergoing a change of continuous increase, the usual expression was
immer mehr ‘ever more’. Recently (estimate: since 1980), immer (mehr) has tended to be
replaced by zunehmend increasingly’. E65 - E67 show this for the intensification of
adjectives, E68 shows it for verbs.
E65. Die Nachfrage wird zunehmend schwächer. [radio commentary, 1987]
‘The demand becomes increasingly weaker.’
E66. Diese Schattenwirtschaft wird zunehmend weniger von der Partei kontrolliert.
[radio, written text, 1988]
‘This shadow economy is controlled increasingly less by the party.
E67. Das echte Elfenbein wurde Ende des letzten Jahrhunderts zunehmend selten und
teuer. [radio, written text, 1989]
‘The real ivory became increasingly rare and expensive by the end of the last
E68. Die Nachfrage verlagert sich zunehmend ins Ausland. [constructed on a common
‘The demand increasingly shifts to foreign countries.’
E65 - E67 clearly show the degree of desemanticization which this device has reached in
a few years (cf. the semantically parallel E58). I suspect that this is, again, a loan
translation from English. Since immer (mehr) is probably a grammatical device of
continuous intensification, zunehmend, too, may be conceived of as a marginal case of
5. Sentence-level syntax
The following two subsections deal with changes in main constituent order. Such are not
prima facie cases of grammaticalization. However, one of the factors in
grammaticalization is the reduction of syntagmatic variability (see Lehmann 1982, ch.
IV.3.3). This leads to a fixation of the sequential order of grammaticalized elements in
Christian Lehmann
23 V: verb; X: one constituent; Y, Z: remaining clause constituents.
relation to their host elements. A language that used to have postpositions can acquire
prepositions through the grammaticalization of relational nouns or coverbs which precede
their complements. The result would be an apparent change in the word order of
Again, rigid ordering constraints at low grammatical levels cannot be directly loosened
(that would be degrammaticalization). However, they can be replaced by freer ordering
if the construction in question is renewed. For example, if there was a rule that possessive
attributes must be prenominal, there is no possible rule change that allows the same
attributes to alternatively stand in postnominal position. There is, however, the possibility
of acquiring, maybe through the grammaticalization of an anaphoric pronoun to an
attributor, a new kind of possessive attribute, which would then be postnominal.
Word order change can, therefore, be the consequence of grammaticalization. It is from
this perspective that the following two cases should be viewed.
5.1. Verb-second position in subordinate clauses
German is famous for requiring final position of the finite verb in subordinate clauses, an
order which never appears in main clauses. There have been enough analyses considering
one of the orders as basic and the other as derived. This issue becomes superfluous as
soon as one realizes that the orderings in main and subordinate clauses are in a
grammaticalization relationship. According to the generalized Penthouse Principle (Ross
1973), syntactic diversity is reduced at lower grammatical levels. In particular, constituent
order is more rigid in subordinate clauses than in main clauses, and it is more rigid in
non-finite subordinate constructions than in finite subordinate clauses. If we apply this to
the facts of German, the following (simplified) picture of German main constituent
structure emerges:23
F3. Constituent order in clause types
subordination type
sentence type
independent dependent
finite non-finite
declarative #X V Y# #Z V# #Z V#
non-declarative #V Z#
Whereas there are two alternative orderings in main clauses according to sentence type,
there is only one ordering in dependent clauses. In addition to this greater paradigmatic
variability, there is also greater syntagmatic variability in main clauses. There is, to begin
with, a ternary subdivision of positional constituents in declarative main clauses, but only
Grammaticalization and related changes in contemporary German
a binary one in subordinate clauses (cf. Heidolph et al. 1981:705f).
Ex hypothesi, the verb-final rule for subordinate clauses represents an instance of relative
rigidity of sequential ordering. It would therefore be unexpected for non-finite dependent
constructions to again show the ordering possibilities of main clauses. Quite to the
contrary, the locus of the verb-final constraint is in non-finite subordinate constructions.
Finite subordinate constructions display verb-final ordering insofar as they are close to,
or modeled on, non-finite ones.
From this reasoning, the expectation arises that if there is variation of main constituent
order in dependent clauses, it will be in finite rather than in non-finite ones, and it will
show one of the possibilities that obtain for main clauses, rather than yet another one. This
is borne out by the facts. There is considerable variation in both colloquial and written
German with respect to verb order in finite subordinate clauses; but there is almost no
possibility to end a non-finite subordinate construction by anything but the verb.
First of all, there are subordinate clauses which do not need to be introduced by a
conjunction and then regularly show verb-first order. These are conditional clauses and
concessive and comparative subtypes of these (cf. Heidolph et al. 1981:718f). Second, if
a subordinate clause S2 is a constituent of a subordinate clause S1, S2 will normally follow
S1, so that S1 does not have verb-final order. The same tends to be true for other heavy
constituents of a subordinate clause, which more often than not follow the finite verb even
in written German. Thus, E69.b is a perfectly acceptable alternative to a.
E69. a. Als er nun endlich nach all der schweren und langweiligen Arbeit Feierabend
machte, da kam es ihm vor, als ob ...
‘When he then finally called it a day after all that hard and boring work, it
appeared to him as if ...’
b. Als er nun endlich Feierabend machte nach all der schweren und langweiligen
Arbeit, da kam es ihm vor, als ob ...
While these are cases of free variation which have to be conceived as subcases of the
verb-final constraint, there is a third case of non-final position of the finite verb in
subordinate clauses which is different from the first two mentioned. This is verb-second
ordering, as in declarative main clauses. I am not here referring to asyndetic constructions
of the type ich glaube, er kommt ‘I think he comes’, whose subordinate nature is doubtful.
The constructions which are of interest in the present connection are illustrated in E70 and
E70. Wir konnten den Wagen noch nicht reparieren, weil wir mußten auf die Ersatzteile
‘We could not yet repair the car because we had to wait for the spare parts.’
E71. Wir können auch noch wárten, obwohl morgen ist er bestimmt áuch nicht weiter.
‘We can wait a little more, although tomorrow he will surely not be any further
Christian Lehmann
24 E72 and the relevant observation are due to Ekkehard König. Cf. also König & Van der Auwera
Such constructions are not standard, but very common at the colloquial level. They have
been in the language since Old High German; cf. Sandig 1973. So far, they appear to be
restricted to causal and concessive clauses. They are not to be interpreted as variations of
subordinate clause constituent order. Instead, the subordinative conjunctions weil and
obwohl (and some synonymous ones) are reinterpreted as coordinative conjunctions, on
the analogy of denn and the adverb allerdings, respectively (which could be substituted
in E70 and E71, with little or no change in meaning, to produce standard German). After
them, main clause syntax is required. This means that subordinate clause order is renewed,
at least in certain areas, by main clause order.
An independent clause has its own illocutionary force, whereas a subordinate clause
generally has none of its own. Just as main clauses linked by denn and allerdings,
verb-second clauses introduced by weil and obwohl are subduced from the scope of the
would-be main predication and independently express the speaker's point of view. This
is illustrated by the following minimal pair:24
E72. a. Maria verehrt ihren Mann, obwohl er es nicht verdient hat.
‘Mary reveres her husband although he does not deserve it.’
b. Maria verehrt ihren Mann, obwohl verdíent hat er es nícht.
‘Mary reveres her husband, but he does not deserve it.’
The concessive clause in E72.a is within the scope both of the main clause illocutionary
force and of the main clause predication (i.e. it is a part of the circumstances taken into
account by the main clause subject). In E72.b, the concessive clause is outside these two
scopes and gives a separate assessment of the situation by the speaker. It is this main
clause status of the concessive clause in b which also renders possible the expression of
emphasis by non-finite predicate fronting.
The main clause hypothesis also explains why the constituent ordering observed in the
"subordinate" clauses of E70 and E71 never occurs when the subordinate clause precedes
the main clause. Inversion of main and subordinate clause is possible in E72.a and would
be possible in E70 and E71 if the subordinate clauses had verb-final order, but not as they
stand. The explanation is that a coordinative conjunction must go between what it
5.2. Verb-first position in main clauses
The German constraint on verb-second position in declarative main clauses differs
substantially from the situation in English. The English rule may essentially be conceived
as a V-3 constraint which says that adverbial, subject and main verb must come in this
order at the beginning of the clause. The German constraint presupposes that first and
Grammaticalization and related changes in contemporary German
25 Even then the sentence would be substandard (just as E80 below), because preposition stranding is
not standard.
second position are fixed constituent structure slots and that the first can be filled by
anything, and determines that the second must be filled by the finite verb.
There are at least two potentially dysfunctional aspects about such a constraint. First of
all, it leads to a substantial amount of discontinuity in main clauses. The non-finite verb
has to be separated from the finite verb, especially from the auxiliary; an adverb that is
part of the verbal lexeme and appears as a verb prefix in subordinate clauses has to be
separated from the verb in main clauses; and so forth. The phenomenon has been
described by Germanists as Satzklammer (‘[sentential] braces’; cf. Heidolph et al.
1981:707-709). We will see in a moment that this is not the interpretation given to it by
speakers. Second, the constraint forbids clause-initial position of the finite verb, which is
necessary if the clause consists only of a verb, and desirable, inter alia, in presentative
The second of these shortcomings is remedied by expletive es ‘itand da ‘there’, as
illustrated in E73 and E74.
E73. a. Es regnet.
‘It is raining.’
b. Es wird getanzt.
‘There is dancing.’
E74. Da kreuzt plötzlich dieser Holtermann auf.
‘Suddenly old Holtermann shows up.’
To understand the working of main constituent ordering, it is essential to be aware that,
in such constructions, es and da are nothing but dummy slot holders which allow formal
observance of the verb-second constraint. In the colloquial language, constructions with
expletive es are not affected, although they are relatively rare (E73.b would more
commonly be da wird getanzt). However, the expletive da is increasingly dropped. Here
are some recorded examples:
E75. A: Das is kaputt das Auto. - B: Kann man nich mehr mit fahren. [1987]
‘A: It is broken, the car.’ - B [4 years]: ‘You can no longer ride in it.’
E76. Wenn es so weiter geht, können wir pünktlich ankommen. Kommen natürlich
nachher noch Geschwindigkeitsbeschränkungen. [conversation, 1987]
‘If it goes on like this, we can arrive in time. There will be, of course, speed limits
later on.’
From the point of view of the standard language, an initial da is missing in B's utterance
in E75,25 which would anaphorically resumethe car’; and the second sentence in E76
likewise should begin with either of the expletive particles.
This construction is allowable in a variety of registers and text sorts. It is typical of
Christian Lehmann
26 Bunny goes to the pharmacist's and asks: ‘Do you have carrots?’ The pharmacist says: ‘No, bunny,
this is a drugstore; we don't have any carrots.Next day, bunny comes to the drugstore and again asks:
‘Do you have carrots?’, and the pharmacist again explains that he doesn't. Third day, the same story. On
the fourth day, the pharmacist has put a sign on his door saying ‘Today no carrots’. Bunny appears and
reproachfully says to the pharmacist: E77.
narrative expositions, both poetic (Sah ein Knab' ein Röslein stehn [Goethe] ‘a boy saw
a little rose’) and colloquial, especially in joke-telling (Kommt eine Frau zum Arzt ... ‘A
woman comes to the doctor's’). It is also no accident that one of the examples is from
child language, as omission of sentence-initial da is very common there. Ten years ago,
there used to be a genre of jokes called Häschenwitze ‘bunny jokes’, whose main character
was a bunny-rabbit. The punch line was always reserved for him, and it consisted of a
declarative main clause introduced by the finite verb. E77 is an example.
E77. Hattu [hast du] also doch Möhrchen gehabt.
‘So you did have carrots, after all.’ 26
The omission of sentence-initial da has been generalized to some of the forms of the
anaphorical pronoun der. In particular, das is as freely omitted as da.
E78. kenn ich, weiß ich, war ich schon [title of a German movie, 1980]
‘I've seen it, I know it, I've been there’
E78 combines some common sayings with omission of both das and da. In examples such
as E75 and E78, the missing form is not merely expletive, but is the sole legitimate
representative of a sentence constituent. In such cases, initial position of the element in
question is the necessary condition of its omission. Consequently, in the alternative
versions E75' and E78', where another constituent takes the first position, the anaphoric
form has to appear.
E75'. Man kann nich mehr damit fahren.
E78'. ich kenn das, ich weiß das, ich wár da schon.
The insensitivity of this process to grammatical relations shows that we are not faced with
an instance of ‘pro-drop’. Instead, it seems best to interpret the examples as the result of
a tendency towards verb-initial order in main declarative clauses. This is also suggested
by the sayings in E79, which are extremely common in the colloquial language.
E79. a. Is[t] ja noch mal gút gegangen.
‘It has turned out well once again.’
b. Wird schon nicht so schlimm wérden.
‘It won't turn out that badly.’
c. Hast du eben Péch gehabt.
‘You just have had tough luck.’
d. Ham [haben] wir ja noch níe gesehen.
Grammaticalization and related changes in contemporary German
‘That we have never seen before.’
e. Könnte ja jéder kommen.
‘Then anybody could come.’
f. Mußtu [mußt du] halt noch mál hingehen.
‘You just have to go there again.’
The missing initial pronominal form has the syntactic functions of subject in E79.a and
b, direct object in d and place adverbial in the others.
Another point is noteworthy about the syntax of verb-initial sentences. While probably any
verb can start a sentence of this type (cf. E78), the examples in E75, E77 and E79 are
typical in beginning with auxiliaries or modals. It remains to be seen why this should be
In this way, both of the limitations of the verb-second constraint are overcome in the
colloquial language. Not only can the sentence start with the verb; but the particularly
disturbing discontinuities are also removed. The full verb is relegated, in its non-finite
form, to that part of the sentence where the syntagms most closely related to it are
assembled. The only remaining discontinuity is the one between the auxiliary or modal
and the full verb. However, this may well be merely an apparent one. There are so many
languages which separate the expression of such grammatical categories as person,
number, tense, mood and aspect from the expression of the lexical verb that it might be
an Indo-European preconception to think that they should go together.
Incidentally, if verb-first order in declarative main clauses is colloquial and the
replacement of verb-final order in subordinate clauses by main clause order is colloquial,
too, one wonders whether it is not an anachronism to introduce verb-second order into
subordinate clauses. In fact, it is exactly that hybrid between verb-first and verb-second
order which is being used in so-called subordinate clauses in colloquial German. E80
would be an example.
E80. Das sollten wir nicht tun, weil (da) hát ja keiner was von.
‘That we should not do, since nobody gets anything from it.’
If this development succeeds, German will become a verb-initial language. The
constructions which are currently at the center of the movement remind one of the
situation in the Celtic languages, where verb-initiality also means, in the first place, initial
position of the auxiliary. In the present context, we have yet to consider in which respect
the development is one of grammaticalization. Under closer analysis, there are two parallel
ways of paying lip service to the verb-second constraint. One is to put an expletive
element in the first position which then reduces to zero. What is grammaticalized here is
the constraint of filling the preverbal position. The other way is to fill the second position
with a grammatical verb, thereby allowing the lexical verb to come later in the clause.
Here the constraint of filling the second position by a verb is depleted, grammaticalized.
Grammatical rules are changed by obeying them.
Christian Lehmann
6. Grammaticalization in contemporary German
6.1. It is difficult to derive a typologically relevant generalization from the diverse
phenomena reviewed above. There is grammaticalization both in the nominal and in the
verbal sphere, both at the low and at the high grammatical levels. The advance of
prepositions to the detriment of cases, and of modals, auxiliaries and function verbs to the
detriment of synthetic verb forms, bespeaks a general tendency towards analytic
morphological structure. Prepositions are separated from the governed noun by the
determiners and attributes of the latter. Grammatical verbs are separated from their
non-finite or nominalized lexical verb by the rules of main constituent order. There is,
therefore, at present no chance that the analytic constructions will coalesce to yield new
synthetic ones.
6.2. At a universal level, the data conform to the rule that the greatest freedom reigns at
the highest grammatical levels. Therefore, the changes that are potentially most
far-reaching are those initiated at the highest levels. At its origin, a change in clause order
or main constituent order is merely an instance of free variation and, thus, inconspicuous
as regards the typological structure of the language. However, if and when the structures
so created grammaticalize, the erstwhile variant will become the rule, and lower level
grammar will be shaped in ways that are completely incompatible with its former state.
6.3. While many grammatical changes go on and probably originate at the colloquial level,
colloquial speech is far from being the only, or even the principal, source of grammatical
change. The mass media play an extremely important role in the coinage and diffusion of
new collocations, which then go the way of lexicalization or grammaticalization.
Grammatical change is, to some extent, register-bound in the sense that some of the
changes are limited to the colloquial level, others to formal style. For some of the
grammaticalizational phenomena observed, especially complex prepositions and function
verbs, it is not the colloquial, but bureaucratic and journalistic style which promotes them.
Some of these innovations may never reach the colloquial level.
6.4. Grammaticalization manifests itself on the synchronic and diachronic axes. Insofar,
phenomena may be, and have been, classed as cases of grammaticalization even if there
is no clear evidence of historical change. In fact, the set of phenomena reviewed in the
foregoing is quite heterogeneous as regards socio-historical aspects of language change.
Some phenomena have appeared in the language only recently, such as the mehr als and
zunehmend-constructions. Others have been well-established in non-standard or dialectal
varieties of German for a long time, such as the progressive or the pseudo-subordinate
clause with the verb in second position. Yet others have always been possible in the
standard language in limited contexts, but have considerably expanded their domain in
recent times. This is true for the main clause with verb in first position.
Again, some of the changes, esp. those relating to word order, lead to entirely new
grammatical structures. Others, e.g. those producing complex prepositions, are but
Grammaticalization and related changes in contemporary German
contemporary instances of a model that has been productive for a long time. In this paper,
the common denominator of all the phenomena has been their grammatical character,
centering around the concept of grammaticalization, and their recency in colloquial
standard German. In all cases, there has been a shift in the pattern of variation which leads
to greater prominence of a phenomenon in colloquial standard German.
6.5. The study of grammatical change on the basis of data on synchronic (or, at best,
brachychronic) variation in the contemporary stage of a language is subject to a serious
problem of verification. Given presently available methodological means, it is next to
impossible to know which of the changes that speech habits currently exhibit are
synchronic manifestations of ongoing genuine language change, and which of them are
but ephemeral fashions. In this situation, a study such as this can hope to elucidate the
synchronic dynamism of the language; it cannot hope to tell which innovations will result
in changes and which ones will disappear without a trace in future synchronies.
Burger, Harald & Buhofer, A. & Sialm, A. 1982, Handbuch der Phraseologie. Berlin &
New York: W. de Gruyter.
Chafe, Wallace L. 1976, "Givenness, contrastiveness, definiteness, subjects, topics, and
point of view". Li, Charles N. (ed.), Subject and topic. New York etc.: Academic
Press; 25-55.
Heidolph, Karl Erich et al. 1981, Grundzüge einer deutschen Grammatik. Berlin:
Helbig, Gerhard 1979, "Probleme der Beschreibung von Funktionsverben im Deutschen".
DaF 16:273-285.
Hopper, Paul & Thompson, Sandra A. 1980, "Transitivity in grammar and discourse". Lg.
Inghult, Göran 1975, Die semantische Struktur desubstantivischer Bildungen auf -mäßig.
Stockholm: Almqvist & Wiksell.
König, Ekkehard & Van der Auwera, Johan 1988, "Clause integration in German and
Dutch conditionals, concessive conditionals, and concessives". Haiman, J. &
Thompson, S.A. (eds.), Clause combining in grammar and discourse. Amsterdam: J.
Benjamins (TLS, 18); 101-133.
Lehmann, Christian 1982, Thoughts on grammaticalization. A programmatic sketch. Vol.
I. Köln: Institut für Sprachwissenschaft der Universität (akup, 48).
Lehmann, Christian 1985, "Grammaticalization: Synchronic variation and diachronic
change". LeSt 20:303-318.
Lehmann, Christian 1989, "Grammatikalisierung und Lexikalisierung". ZPSK 42:11-19.
Matthews, Peter H. 1981, Syntax. Cambridge etc.: Cambridge UP (Cambridge Textbooks
in Linguistics).
Pilz, Klaus Dieter 1983, "Suche nach einem Oberbegriff der Phraseologie und
Christian Lehmann
Terminologie der Klassifikation". Mate÷siºc, Josip (ed.), Phraseologie und ihre
Aufgaben. Beiträge zum 1. Internationalen Phraseologie-Symposium vom 12. bis 14.
Oktober 1981 in Manheim. Heidelberg: J. Groos (Mannheimer Beiträge zur slavischen
Philologie, 3); 194-213.
Plank, Frans 1980, "Encoding grammatical relations: acceptable and unacceptable
non-distinctness". Fisiak, J. (ed.), Historical morphology. The Hague etc.: Mouton
(Trends in Linguistics; Studies and Monographs, 17); 289-325.
Ross, John R. 1973, "The penthouse principle and the order of constituents". Corum,
Claudia et al. (eds.) 1973, You take the high node and I'll take the low node. Papers
from the Comparative Syntax Festival. The differences between main and subordinate
clauses. 12 April 1973; a paravolume to Papers form the Ninth Regional Meeting.
Chicago: CLS; 397-422.
Sandig, Barbara 1973, "Zur historischen Kontinuität normativ diskriminierter
syntaktischer Muster in spontaner Sprechsprache". Deutsche Sprache 1973/3:37-57.
Seiler, Hansjakob 1973, "On the semantosyntactic configuration ‘possessor of an act’".
Kahane, Henry & Kahane, Renée (Festschrift), Issues in linguistics. Papers in honor
of Henry and Renée Kahane. Ed. by Braj B. Kachru et al. Urbana etc.: University of
Illinois Press; 836-853.
Seiler, Hansjakob 1983, POSSESSION as an operational dimension of language.
Tübingen: Narr (LUS, 2).
Werner, Otmar 1979, "Kongruenz wird zu Diskontinuität im Deutschen". Szemerényi,
Oswald (Festschrift) 1979, Festschrift for Oswald Szemerényi on the occasion of his
65th birthday edited by Béla Brogyanyi. Amsterdam: J. Benjamins (CILT, 11);
Wurzel, Wolfgang Ullrich 1984, Flexionsmorphologie und Natürlichkeit. Ein Beitrag zur
morphologischen Theoriebildung. Berlin: Akademie-Verlag; Darmstadt: WB (Studia
grammatica, 21).
... A third hypothesis is that certain movement verbs that contained a path mark in the root (such as entrar, sair, subir, descer, avançar, voltar etc.,) suffered an impoverishment of their semantic content -a process called bleaching (GIVÓN, 1991), semantic depletion (LEHMAN, 1991) or desemanticization (HEINE; REH, 1984) -which led to the loss of path, leading to the mark of this category in a complementary structure, which we assume here as "satellite". ...
Full-text available
This work analyzes the marking of both path and manner categories in verbal expressions of movement in Latin, Old Portuguese, and vernacular Brazilian Portuguese, based on Talmy’s (2000) typology, which divides languages between two patterns: verb-framed and satellite-framed. The aim is to explore in an essayistic way, in order to surprise new research problems, such as marking in these languages and formulate alternative hypotheses to explain an apparent change in Brazilian Portuguese. Based on data extracted from different corpora, the results indicate that Latin behaved like a typical s-framed language (similar to modern Germanic languages); in the transition to modern romance languages, there was variation between the two patterns (v-framed and s-framed); later, the neo-Latin languages were consolidated as typical v-framed languages; today, however, the Brazilian vernacular presents consistent data that enable us to classify it as a language prone to the s-framed pattern, similar to Latin. KEYWORDS: Movement verbs. Verb-framed languages. Satellite-framed languages. Romance languages. Brazilian Portuguese.
... Im Zusammenhang mit der Beschreibung von Grammatikalisierungserscheinungen in der deutschen Gegenwartssprache hat Lehmann (1991) (Lehmann 1991: 518). Als grammatische Funktionen wird dabei nicht nur das Inventar der grammatischen Kategorien Person, Numerus, Genus verbi, Tempus sowie Aktionsart angesehen, sondern, nach den Beispielen zu schließen, auch Kausativität. ...
... Consequently, and as opposed to what generally happens in the process of grammatical change (cf. Lehmann 1991Lehmann , 2002a, one cannot claim that, in Serbian, the rise of prepositions -either simple or complex ones -has ousted case marking, nor that the possible attrition of case inflection stimulated the prepositions to fill the gap. What seems to have happened is a certain loss of the capacity for some cases to express certain relationships autonomously (cf. ...
This paper takes a data-driven perspective on the grammaticalization of German light verb constructions (LVCs) with kommen ‘come’. LVCs are complex predicates consisting of a semantically light verb and an eventive noun realized within a phrasal complement, e.g. German zur Vollendung kommen , lit. ‘come into completion’. We assume that (at least) two different processes interact in the emergence of LVCs: the desemanticization of the verb on the one hand and the realization of eventive nouns in the complement-PP of the verb on the other. In order to check whether these processes take place in parallel or if one precedes the other, we conduct a corpus study based on samples from the Reference Corpus of Middle High German (REM) and the Bonn Early New High German Corpus (FnhdC), focusing on the animacy and concreteness of the subject NPs and the PP-internal nouns. Our results indicate that we can first observe an increase in the use of abstract nouns in subject position and that only later – from Middle High German to Early New High German – eventive nouns in PP-internal position become more frequent.
Change is an inherent feature of all aspects of language, and syntax is no exception. While the synchronic study of syntax allows us to make discoveries about the nature of syntactic structure, the study of historical syntax offers even greater possibilities. Over recent decades, the study of historical syntax has proven to be a powerful scientific tool of enquiry with which to challenge and reassess hypotheses and ideas about the nature of syntactic structure which go beyond the observed limits of the study of the synchronic syntax of individual languages or language families. In this timely Handbook, the editors bring together the best of recent international scholarship on historical syntax. Each chapter is focused on a theme rather than an individual language, allowing readers to discover how systematic descriptions of historical data can profitably inform and challenge highly diverse sets of theoretical assumptions.
Full-text available
Transitivity involves a number of components, only one of which is the presence of an object of the verb. These components are all concerned with the effectiveness with which an action takes place, e.g., the punctuality and telicity of the verb, the conscious activity of the agent, and the referentiality and degree of affectedness of the object. These components co-vary with one another in language after language, which suggests that Transitivity is a central property of language use. The grammatical and semantic prominence of Transitivity is shown to derive from its characteristic discourse function: high Transitivity is correlated with foregrounding, and low Transitivity with backgrounding.
Full-text available
Grammaticalization is a continuous process happening in the synchrony and diachrony of language which affects a construction or sign, reducing its autonomy and subjecting it to the rules of grammar. The degree to which a construction or sign is grammaticalized may be assessed by six parameters, ordered on the paradigmatic and syntagmatic axes along which linguistic entities are manipulated.
Renée (Festschrift), Issues in linguistics Papers in honor of Henry and Renée Kahane
  • Kahane
  • Henry
  • Kahane
Kahane, Henry & Kahane, Renée (Festschrift), Issues in linguistics. Papers in honor of Henry and Renée Kahane. Ed. by Braj B. Kachru et al. Urbana etc.: University of Illinois Press; 836-853.