Discourse relators and the beginnings of sentences in English and

<TARGET "doh" DOCINFO AUTHOR "Monika Doherty"TITLE "Discourse relators and the beginnings of sentences in English and German"SUBJECT "LIC 3:2"KEYWORDS "word order, adversative connectors, English/German"SIZE HEIGHT "220"WIDTH "150"VOFFSET "4">
Discourse relators and the beginnings
of sentences in English and German
Monika Doherty
Humboldt-Universität zu Berlin
Felicitous translations between English and German show that discourse
relators like adversative connectors interact with the basic strategies of discourse linking superimposed upon the grammatical rules of word order. For
initial position they compete with elements from the propositional meaning
of a sentence, which they precede if the propositional elements are given
information; if the elements are new (or resumed) information, discourse
relators tend to follow them. Attitudinal expressions like probably and possibly
precede or follow propositional elements in the same way, but tend towards a
greater degree of explicitness in initial positions of English sentences.
Keywords: word order, adversative connectors, English/German
1.
Candidates for initial positions
There are many ways of beginning a sentence and even if we concentrate on
texts of a certain type, say, academic written prose, the diversity leaves little
hope for any regularities, all the more so if we compare texts in different
languages. But if we concentrate on the syntactic classes of beginnings, we can
make out certain regularities in the use of grammatical relations constituting
the propositional meaning of sentences and in the use of discourse operators
specifying the discourse relations between sentences. Canonical discourse
operators are connectors, like ‘but’, or adverbs, like ‘however’, but also adverbial phrases like ‘in addition’, ‘on the other hand’ and the like. There is a second
class of operators, like ‘probably’ or ‘unfortunately’, specifying attitudes
towards propositions. Both types of operators compete with the propositional
parts for the initial position in a sentence. We can observe regularities concerning the positional options for both types of operators and regularities concerning
Languages in Contrast 3:2 (2000/2001), 223–251.
issn 1387–6759 / e-issn 1569–9897© John Benjamins Publishing Company
<LINK "doh-r9">
"doh-r5">
"doh-r11">
"doh-r18">
"doh-r13">
"doh-r16">
"doh-r10">
224 Monika Doherty
the structural explicitness of attitudinal expressions. Sections 2 and 3 will take
up discourse connectors and attitudinal expressions succinctly. Regularities
concerning the beginning of sentences in terms of propositional meaning will
be outlined in the remaining parts of this section, 1.1–1.5. Regularities concerning the right-hand side of sentences will only be included if they are related to
the beginning.
1.1 Some general grammatical and pragmatic aspects of word order
If we assume that there is a grammatically determined basic order of constituents and ignore discourse operators for the time being, we can say that the basic
order of most declarative sentences begins with the subject in English and
German. Basic order is controlled by the grammatical properties of the lexical
heads of phrases determining the number, type and hierarchical positions of
possible structural extensions. This applies primarily to the lexically predetermined complements of a head. Their basic order can be considered a direct
reflection of the semantic hierarchy between the arguments of a lexical head
(Uniformity of Theta Assignment Hypothesis, ‘UTAH’, Baker 1988). The major
category which determines basic word order in a sentence is the verb. In most
cases, its highest argument is mapped onto the subject so that the basic order of
most sentences begins with the subject. (For a concise survey of this ‘lexical
projection’, see for example Rosengren 1991, Haider and Rosengren 1998; for
a more traditional, but well-known approach to English verb phrases, see
Hornby 1995 on verb patterns.)
There is also a grammatically determined basic order of adjuncts, that is, of
constituents which are not predetermined by the lexical meaning of a head. In
a simplified way, we could say that adjuncts reflect the order of the constituents
they modify. As a result, for example, event-related adverbials, like time
adverbials, dominate event internal adverbials, like instrumental adverbials, and
event internal adverbials dominate process-related adverbials, like manner
adverbials. (For a more precise description of German and English adverbials
along these lines, compare Pittner 1999 and Frey and Pittner 1998, 1999; for
adjectives see Drubig 1997.)
The grammatically determined basic order of elements is superimposed by
rules of language use securing a discourse appropriate order of elements.
Roughly, we can say that in discourse we do not start with the grammatically
highest element, but with the element that secures an optimal discourse link. To
begin with, we can assume a universal, discourse linking strategy proceeding
<LINK "doh-r19">
Discourse relators and the beginnings of sentences 225
from contextually given elements to new ones. Let me call it the strategy of
‘given-before-new’, GIN.
There are two basic types of constraints on GIN regarding the choice of
initial elements: grammatical constraints limiting the number of elements that
can be moved to the beginning of a sentence, and pragmatic constraints
following from the general goal of easy processing. Processing efforts are
controlled by a principle of economy aiming at maximal gains for least effort
(Sperber and Wilson 1986 call this the ‘Principle of Relevance’, the presumption
of which guides felicitious communication). Economy may be the driving force
behind GIN. But optimizing discourse linking at the beginning of sentences
should not worsen processing conditions elsewhere in the sentence. Securing
optimal processing conditions for the entire sentence requires us to adopt
something like a strategy of even distribution of information. Let me call it the
strategy of ‘balanced information distribution’, BID, and demonstrate it by way
of examples.
It is clear that the grammars of different languages provide different
conditions for GIN and BID. However, comparing the different conditions is a
complex affair even with relatively simple sentences. We may not only reorder
the constituents of sentences in line with GIN and BID, but also choose a
different perspective by syntactic or lexical means. Nevertheless, comparing
English and German on the basis of ‘controlled’ translations (see below), we
find the beginning of sentences to be in line with the pragmatic strategies more
often in German than in English. That is, GIN and BID together determine the
beginning of German sentences to a greater extent than that of English sentences.
Let me postpone the theoretical discussion of the concepts involved in GIN
and BID for a moment and illustrate the difference by relatively simple examples taken from translations between English and German.
Appealing to original sentences and their translations as evidence of
language-specific preferences may sound arbitrary. But the comparison of
systematically varied paraphrases reveals a high degree of agreement among
native speakers as to which of the paraphrases is to be preferred in a certain
discourse. In particular, the discourse adequacy which the paper claims for the
examples with adversative connectors in Section 2 was confirmed by a group of
twelve students specializing in translation. The examples of Section 3 were
confirmed by 8–10 German and English students and members of a research
group. The examples were taken from the Berlin translation corpus, which
comprises English and German popular-scientific texts and their translations
(altogether about 50 000 words). The corpus is very small, but the German
<LINK "doh-r2">
226 Monika Doherty
translations are subjected to the method of control paraphrases, which assesses
paraphrases of the original and its translation in terms of discourse appropriateness and revises the German translation if this is felt to improve discourse
appropriateness.
1.2 Binary information units
The first example contains two arguments, which constitute two informational
segments. The context, a text on supernovae, tells us that the indefinite phrase
is new and the definite phrase given information (see appendix). The original
sentence is English:
(1) Two classes of progenitors seem to fulfil these conditions. (m66)
This has been translated into German as
(2) Diese Bedingungen sind offensichtlich bei zwei Klassen von Ausgangssternen
gegeben.
English uses given information after new information, while German restructures the sentence in line with GIN.1 That is, in German we prefer (2) to the
more analogous
(3) #Zwei Klassen von Ausgangssternen scheinen diese Bedingungen zu
erfüllen.2
#Two classes of progenitors
seem these conditions to
meet/fulfil
On the other hand, English prefers the original version (1) to a back-translated
(4) #These conditions seem to be met/fulfilled by two classes of progenitors.
As there are more cases like these, they suggest a regular difference between
discourse linking strategies in English and German. This is also true of the
second class of cases, sentences consisting of more than two informational
segments. But here the given-new pattern is too simple and has to be replaced
by a more differentiated one. The unifying concept will be ‘discourse relevance’
(see 1.3 below), which replaces ‘given’ and ‘new’, ‘background’ and ‘focus’ by
graded information values. It avoids ‘theme’ or ‘topichood’ altogether. Thus,
questions about the border between theme and rheme or the hierarchy of
multiple themes involved in their order (relative to the finite verb) as raised by
Altenberg (1998) do not arise — despite the considerable conceptual overlap in
Discourse relators and the beginnings of sentences 227
terms of ‘textual continuity’ (high degree of givenness or low relevance) and
‘textual break’ (contrastiveness, novelty or high relevance).
1.3 Information units with more than two segments
To describe the difference between the English and German versions of cases
with more than two segments, we have to distinguish between more than two
information values contributing to progress in discourse. Some informational
elements will be more, others less relevant; one element will be most/least
relevant. The different degrees of discourse relevance cannot be determined in
absolute terms but only relative to each other. We can assume that immediately
given elements are least relevant: they do not contribute to progress in discourse. Elements which are new or which are resumed only after some time are
more relevant. If we were to apply GIN to a tripartite informational unit, we
would have to start with the least relevant element in initial position and
squeeze all of the more relevant information into the verb phrase. The result
would be an unbalanced distribution of information. If we aim at an even
distribution of information in line with BID, we have to place the least relevant
element in between the more relevant ones. This is what we find in German.
Consider the following example from a text on genetics. The original
sentence
(5) Several research groups have recently achieved this […] (g67)
has been restructured in the German translation as
(6) In der letzten Zeit ist das einigen Forschergruppen gelungen […]
with the least relevant element, the immediately given information of the pronominal subject (das), between the more relevant information of the temporal adjunct
and the object. (For the time being, the informational hierarchy between adjunct
and object shall be ignored — but see the remarks on sentence focus below.)
The basic word order:
(7) #Das ist in der letzten Zeit einigen Forschergruppen gelungen […]
or the passive:
(8) #This has recently been achieved by several research groups […]
which would both have secured the simple given-new pattern, have not been
used. Nor has the adjunct been topicalized in English:
228 Monika Doherty
(9) #Recently, several research groups have achieved this […]
or left in its VP internal position in German:
(10) #Einigen Forschergruppen ist das in der letzten Zeit gelungen […]
Obviously, word order variations and sentence perspective, in particular active
or passive, are controlled by language-specific conditions for a discourse
appropriate distribution of information. In examples like the above English
prefers the active perspective and makes no use of topicalization even if the
resulting structure violates GIN; German uses passive(-like) structures and
topicalizes freely in line with GIN and BID. Binary information structures
follow GIN, tripartite structures follow BID. In both cases, the more or most
relevant information follows the less or least relevant information. (If we
identify the former with the focus of a sentence, both patterns secure end focus;
see endnote 11). With this interpretation, GIN and BID explain (2), (3), (6), (7)
and (10), but they do not explain the English examples.
At first sight, the differences between English and German look quite
arbitrary — all the more so as we find passive sentences and sentences with
preposed adjuncts also in English. But if we think of the different grammatical
‘mould’ of German and English sentence structures, the tendencies may not be
arbitrary after all.
1.4 English constraints on preposing
On the one hand, there is the limited variability of English word order: except
for a few cases of prepositional objects, mostly representing benefactive roles,
objects can only be used in initial position under very specific conditions.3 This
applies also to lexically predetermined adjuncts (as for example the local
adjunct after a verb like ‘live’), which can be considered a border line case of
arguments. The greater invariability may be due to the higher degree of
configurationality in English, which can be considered one way of compensating for case syncretism, that is the lack of morphological case, which forces the
language user to rely more heavily on structural configurations for the syntactic
identification of constituents.
On the other hand, the fixed order subject-before-verb pushes any preposed
element at the left periphery of the sentence into a more ‘dislocated’ position
than in German, where any preposed element is fully integrated through the
inversion of subject and verb.
<LINK "doh-r8">
Discourse relators and the beginnings of sentences 229
As a result, verb phrase internal arguments which are preposed in German
in line with GIN will remain in their basic position in English in many cases.
But so will many adjuncts that are preposed in German in line with BID.
However, as the percentage of preposed adjuncts is still quite high in English
texts, BID must to some extent determine the choice of the initial element also
in English. So, what exactly are the conditions for adjunct preposing in English?
Translational evidence suggests that one of the constraints on preposing is
just another variant of BID. Except for special reasons, the adjunct is not
preposed in English if this leaves too little information in the verb phrase. Thus,
(5) Several research groups have recently achieved this […]
does not only avoid the preverbal cluster of phrases of a version with a preposed
adjunct:
(9) #Recently, several research groups have achieved this […]
but distributes its relevant elements more evenly around the (finite) verb.
The constraint does not apply in a sentence like
(11) Early in this century scientists attempted a few ingenious investigations
of growth in electric fields, and of the fields produced by growth, […]
(e12)
where the adjunct is not needed to fill the verb phrase as the object presents
sufficient relevant information after the verb.
It is not always easy to determine the informational value of a segment. Not
only may contextual givenness have to be differentiated further into immediately given and less (immediately) given, but given elements may participate in
new relations and become more relevant this way. Even structural length need
not be identical with discourse relevance. We may find rather heavy adjuncts
before short verb phrases if the discourse relevance of the postverbal information is high enough. Compare for example a sentence like
(12) Between the Pioneer 10/11 flybys of 1973–74 and the Voyager flybys of
1979, it shifted hemispheres. (j39)
where the context tells us that the verb phrase carries the most relevant information.4
If the postverbal condition is met, the adjunct may but need not be fronted.
Next to the avoidance of garden paths as in (12), contrastiveness is the normal
condition for adjunct preposing in English (see Doherty 2003a and b).
230 Monika Doherty
1.5 Passive sentences
Things are yet more complicated as the constraints on preposing in English
could often be avoided by a passive. But it seems that English prefers active
sentences whenever there is a specific agent or even at the cost of indefinite,
‘dummy’ agents. Compare once more (11) with its general subject, deleted in
the translation by passivization:
(13) In den Anfängen dieses Jahrhunderts wurden auch einige raffinierte
Experimente zum Wachstum in elektrischen Feldern und zu den durch
Wachstum erzeugten elektrischen Feldern durchgeführt […]
The insertion of a dummy agent may also have pragmatic reasons. An analogous passive version in English with every information before the verb and
none after it would strongly violate BID:
(14) Early in this century a few ingenious investigations of growth in electric
fields, and of the fields produced by growth were attempted […]
The verb is clearly not sufficiently relevant on its own.
On the other hand, if an active sentence does not contain sufficiently
relevant elements in the verb phrase, English will often use a passive perspective
in line with GIN. Thus, we find
(15) The conductivity is strongly affected by the presence of sulphate particles
or sulphuric acid from volcanic acid rain. (v45)
and not:
(16) #The presence of sulphate particles or sulphuric acid from volcanic acid
rain strongly affects the conductivity.
The object is immediately given and its discourse relevance too low to fill the
postverbal position properly.
While English allows concentration of all the relevant information in the
verb phrase, German prefers an additional preposing of the adjunct:
(17) Durch Schwefelpartikel oder Säure aus vulkanischem Säureregen wird die
Leitfähigkeit stark verändert.5
A passive sentence with basic word order fails to meet BID at the left periphery:
(18) #Die Leitfähigkeit wird durch Schwefelpartikel oder Säure aus vulkanischem
Säureregen stark verändert.6
<LINK "doh-r8">
Discourse relators and the beginnings of sentences
In both languages, GIN and BID require a properly filled verb phrase, but only
German can also apply BID freely in placing new or contrasted elements to the
beginning of a sentence. There are, of course, many more aspects to take into
consideration. For example, copular sentences, which are very frequent in
English, present yet another class of language-specific conditions for the
application of GIN and BID. But it is clear that in a verb-second language like
German, with only one position before the verb, and almost no constraints on
preposing, GIN and BID will produce different results than in English with
more positions before the verb and heavy constraints on preposing (for a more
explicit presentation of the details see Doherty 2003a and b).
So far, we have only looked at sentences without discourse connectors. But
discourse connectors compete with the propositional parts for initial position
in a sentence. This is particularly obvious in the case of adversative connectors
in German.
2. Adversative connectors
The most frequent adversative connectors in English are but and however. If we
ignore the case of parenthetical ‘however’ for the time being (but see 2.4), we
can say that these connectors are used in the initial position independently of
the discourse status of the following parts. German uses many more adversative
connectors: aber, jedoch, allerdings, dennoch, doch, and places them either in
initial position — with or without inversion of the following sentence structure
— or in sentence internal position.
Adversative connectors do not alter the distribution of the propositional
elements along the lines of GIN and BID. Thus, we find
(19) But that’s not all.
(20) Doch das ist nicht alles.
or
(21) But here the resemblance ends.
(22) Doch hiermit hört die Ähnlichkeit auf.
where the adversative connector precedes a binary information structure in the
first case and a tripartite one in the second case. But there are also sentences like
(23) But this is only part of the story. (h27)
231
232 Monika Doherty
(24) Das ist aber noch nicht alles.
It seems that doch is just added to the discourse determined information
structure, its position being invariant, while aber is itself subjected to discourse
criteria as its position is variable.
Other connectors like allerdings and jedoch are variable, too. The position
of variable connectors is controlled by discourse conditions as they compete
with other potential candidates for initial position.
2.1 Initial connector
In most cases initial connectors in Geman are correlated with given information
in the position after the finite verb. Thus, a sentence like
(25) But their analysis does place eruptions like those of Mount St. Helens in
historical perspective. (v25)
is translated as
(26) Dennoch weist ihre Analyse Vulkanausbrüchen wie dem des Mt St Helens
ihren Platz in der Geschichte zu.
The “analysis” which is referred to by the subject was described in some detail
in the preceding paragraph. The initial position of the connector is also due to
scope conditions: dennoch could only be used in medial position with a change
in meaning (rejecting a straight alternative of the meaning of the verb phrase):
(27) #Ihre Analyse weist dennoch Vulkanausbrüchen wie dem des Mt St Helens
ihren Platz in der Geschichte zu.
Or consider a sentence like
(28) But a CAM3 plant finds itself in a physiological catch-22. (f31)
which is translated as
(29) Jedoch befinden sich diese CAM3 Pflanzen physiologisch in der Klemme.
as its subject takes up the referent introduced by the immediately preceding
sentence. And
(30) However, first impressions can be deceptive […] (s5)
is translated as
Discourse relators and the beginnings of sentences 233
(31) Aber der erste Eindruck täuscht oft […]
because a possible first impression is mentioned in the preceding sentence.
Unlike jedoch and dennoch, aber is not associated with a verb-subject
inversion, but it is unclear whether this difference plays any part in discourse
linking (see 2.5. for some regularities, though).
An analogous position of the adversative connector in English and German
may also be due to a different perspective. Thus, an intransitive sentence with
a new subject
(32) But a warning bell has sounded. (h125)
has been reframed in the translation as
(33) Dennoch haben wir einen Warnruf vernommen.
where the indefinite subject of the original has been turned into the object of a
perceptive verb, while the position of the subject has been filled by the
informationally weak element wir. However, if there are other elements
competing for initial position, the connector will have to give way.
2.2 Medial connector
If the subject contains new or contrasted information, it will precede the
variable connector. Thus
(34) However, animal experiments have shown that for a strong effect interferon has to be used at doses at least 10 times greater than were
employed in any of these clinical trials. (i93)
becomes
(35) Tierversuche haben jedoch ergeben, daß …
because the preceding passage talks about (human) patients with cancer
therapy.
Or take a sentence like
(36) However, at that time interferons had never been properly tested […]
(i7)
which has been translated as
(37) Zu diesem Zeitpunkt verfügt man jedoch ganz allgemein […]
<LINK "doh-r8">
234 Monika Doherty
Ignoring any other restructuring in the example, we can say that the subject and
the adverbial present given information, but as the adverbial allows a contrastive interpretation,7 it is chosen for initial position. Consequently, the connector
is used sentence internally.
The competing elements may originate in different positions in English.
Thus,
(38) However, no ill effects have been seen so far in these patients. (i111)
has a propositional part with a negated subject and the temporal adjunct with
new information before the local adverbial with given information. The
German translation
(39) Bisher sind aber solche negativen Effekte noch bei keinem Patienten
beobachtet worden.
presents the temporal adjunct in initial position and uses the connector
internally. The basic word order would have been:
(40) #Solche negativen Effekte sind aber bisher noch bei keinem Patienten
beobachtet worden.
As the subject of the German version is not negated, it presents merely given
information. Placing it into the lowest position of the information structure
yields a more balanced information structure.
In many cases we find the subject restructured even more radically in the
translation. Compare
(41) However, modern theories of the deep interior of Jupiter […] suggest
that […] (j31)
which is translated as
(42) Aus neueren Annahmen über das Jupiterinnere […] folgt aber, dass […]
The subject of the original is restructured as a prepositional object, and is placed
initially in line with its contrastive discourse link; the adversative connector
follows sentence internally.
Another case of restructuring is due to the reduction of cleft(-like) sentences, that is, of sentences with extra copular clauses (see Doherty 2001). Thus, in
(43) But one difference between the Earth and Jupiter is that […] (j67)
the main clause is reduced to an adverbial prepositional phrase:
Discourse relators and the beginnings of sentences 235
(44) Im Unterschied zur Erde scheint jedoch […]
The subject of the main clause is changed into an adverbial, its contrastive
discourse function locates the adversative connector sentence internally.
The restructuring may involve even more elements of the sentence. Thus,
(45) But additional material is injected volcanically from the mantle, beneath
the Earth’s crust, along the active mid-oceanic ridges. (s26)
is translated as
(46) Entlang der aktiven mittelozeanischen Rücken kommt jedoch zusätzlich
vulkanisches Material aus dem Mantel unterhalb der Erdkruste ins Meer.
As ‘volcanically’ is a lexical gap in German, the modifiers of the original have
been exchanged in the translation. The chain of prepositional phrases has been
broken up to place one of the phrases at the beginning of the translation.8 But
as the local adverbial is new information, the adversative connector is assigned
to an internal position.
Another case of restructuring is illustrated by
(47) But we do not know how patients would respond […] (i96)
which is presented impersonally in the translation
(48) Es ist aber nicht bekannt, wie […]
As the initial pronoun itself does not carry any meaning, we might expect the
connector to be used initially. But the placeholder subject is a syntactic trace of
the extraposed subject clause with all its new information, pushing the connector further down in the sentence.9
Preposed material may also originate in a complement clause, as in the
following example
(49) But it is difficult to work out just how much fresh, as opposed to recirculated material, comes out from the vents in the mid-oceanic ridges. (s28)
where the translation renders the interrogative clause by a nominal phrase
(50) Das Verhältnis zwischen dem neuen, aus mittelozeanischen Rücken
stammenden und dem zirkulierenden Material ist allerdings schwer zu
bestimmen.
As the head of the phrase carries new information, the adversative connector is
used internally.
236 Monika Doherty
2.3 Exceptions
There are some cases where the adverbial connector is used initially although
the competing part of speech contains new or contrastive information. Firstly,
if the adversative connector is doch, it is used initially. Thus, we get
(51) However, in man for instance […] (i17)
which becomes
(52) Doch im menschlichen Körper beispielsweise […]
although the local adverbial contrasts with the immediately preceding ‘animal
species’. This also applies to restructured cases like
(53) But a completely new perspective has come from the incredible images
taken during the encounters of the two Voyager spacecraft with Jupiter
in 1979. (j19)
(54) Doch die sensationellen Jupiteraufnahmen, die die beiden Voyager-Sonden
1979 zur Erde übermittelten, haben zu völlig neuen Einsichten geführt.
Secondly, there are cases where the competing contrasted element leaves the
initial position to the adversative connector because it has its own focus
indicator:
(55) However, only in the case of calcium carbonate have studies shown that
[…] (s61)
is translated as
(56) Allerdings konnte nur für Kalziumkarbonat nachgewiesen werden, daß […]
where the focusing element ‘only’/nur secures focus identification for the
prepositional phrase following it. The same is true of the postposed indicator
‘alone’/allein restricting the immediately given subject in
(57) However, rapid reactions alone are not sufficient to establish an equilibrium. (s67)
As the contrastive element has its own focus indicator, the connector is placed
initially:
(58) Dennoch sichern schnell ablaufende Reaktionen allein noch kein Gleichgewicht
Discourse relators and the beginnings of sentences 237
2.4 Postposed ‘however’
If the adversative connector is postposed in English, it indicates a preceding
contrast, which might otherwise be overlooked. Thus, we can expect a sentence
internal position for the connector in German. Example (59/60) uses the
adversative connector after a temporal adverbial:
(59) In the past few decades, however, they have largely ignored […] (e8)
(60) In den vergangenen Jahrzehnten hat man allerdings die Elektrizität weitgehend vernachlässigt […]
Sometimes the contrast is rather subtle — although following information may
confirm it. In
(61) Small voltages, however, can add up to bigger ones […] (e32)
it might easily be overlooked at first as the preceding sentence mentions small
voltages, too. Yet there they are introduced by a generic indefinite noun phrase,
a voltage drop too low to be picked up […], while (61) focuses on the possibility
of multiplying the small voltage, that is, it focuses on the plural. The contrast is
lexicalized in the German translation
(62) Mehrere geringe Spannungen ergeben aber zusammen eine größere
with the adversative connector in the internal position.
In most cases, the need for restructuring is even greater. In the following
example, the adversative connector separates two participle phrases
(63) The tiny currents pumped by cells, however, compounded by the low
resistance of the sea water outside, produce a voltage drop outside the
cell (the only way the current could be measured) too low to be picked
up by conventional microelectrodes. (e31)
The new information is contained in the second phrase. As an analogous
translation would have to squeeze most of the information into the verb phrase,
German reorders and reframes the sentence, using the local adverbial initially
(64) Außerhalb der Zelle aber, wo sich der Strom messen läßt, ergeben die
winzigen Zellströme zusammen mit dem geringen Widerstand des Meerwassers einen zu geringen Spannungsabfall, um mit herkömmlichen Mikroelektroden erfaßt zu werden.
238 Monika Doherty
The adversative connector is here also used as a boundary between two segments
of the initial constituent, separating the nominal head of the local adverbial from
its clausal extension, which the original uses as an apposition in parenthesis.
Restructuring can even result in a cleft-like structure in German. Thus,
(65) One final ionic event in eggs, however, does exert a purely electrical
effect […] (e90)
has been translated as
(66) Es gibt jedoch eine wichtige Ionenreaktion […] die ausschließlich
elektrischer Natur ist.
In the translation, the contrastive interpretation of the subject is promoted by
an extra clause. As the placeholder es introduces the entire focus, the adversative
connector is, again, used in the sentence internal position.
2.5 Choosing the adequate connector
At first glance, adversative connectors seem to be interchangeable except for the
structural differences associated with them: the invariability of the connector
doch, and the inverting effect of jedoch (doch), dennoch and allerdings vs. aber in
initial position. In discourse, however, there seems to be a clear preference for
one or the other connector, depending upon the textual relevance of the
segments which participate in the adversative relation.10
The textual relevance of the elements related to each other may differ, with
the preceding segment being more relevant (allerdings) or less relevant, in
particular less relevant than an original assumption returned to (doch and
dennoch). Let me call allerdings ‘concessive’, doch and dennoch ‘resumptive
adversative’ connectors.
The examples with doch suggest that it returns to an earlier assumption in
relation to a new or contrasted element:
(52) Doch im menschlichen Körper […]
“However, in human bodies […]”
(54) Doch die sensationellen Jupiteraufnahmen […]
“However, the stunning shots of Jupiter […]”
while dennoch returns to an earlier assumption in relation to an immediately
given (if explicitly contrasted) topic:
Discourse relators and the beginnings of sentences 239
(26) Dennoch weist ihre Analyse […]
“However, their analysis shows […]”
(33) Dennoch haben wir (einen Warnruf vernommen)
“However, we (heard a warning signal)”
(58) Dennoch sichern schnell ablaufende Reaktionen allein […]
“However, rapid reactions alone do not guarantee […]”
As dennoch is associated with the inversion of the finite verb, it assigns the given
element to the weakest position, while sentence initial doch without inversion
is followed by the new/contrasted element itself.11
The adversative particles, aber and jedoch, are subject to the same formal
difference: jedoch inverts, aber does not invert.
(29) Jedoch befinden sich diese CAM3 Pflanzen […]
“However, these CAM3 plants find themselves […]”
(31) Aber der erste Eindruck täuscht oft […]
“But the first impression is often misleading […]”
An alternative distribution would be discourse inadequate. Identifiying the
most relevant element of
(67) #Jedoch täuscht der erste Eindruck oft […]
with the finite verb, which has left the canonical focus position,12 requires more
processing effort than in (31). On the other hand,
(68) #Aber diese CAM3 Pflanzen befinden sich physiologisch in der Klemme.
“But these CAM3 plants find themselves physiologically in trouble”
is less discourse appropriate than (29), which places the least relevant element
in the weakest position of the sentence, right after the verb.
The inversion-bound difference is restricted to connectors used in initial
positions, in their medial positions jedoch and aber are freely interchangeable.
If this analysis is correct, the information structures of versions like the following can be considered equivalent:
(20) Doch das ist nicht alles.
(69) Das ist aber nicht alles.
Doch indicates a contrast on the subject no less than initial position of das
before the VP internal aber does. That is, we can associate the proposition of
these sentences with a prosodic hat-pattern, assigning some textual relevance to
240 Monika Doherty
the discourse segment referred to by the pronoun. This would not be the case
with the inverted version
(70) Doch ist das nicht alles.
or with the initial aber:
(71) Aber das ist nicht alles.
But the pronominal nature of the ‘competing’ element and the brevity of the
sentence can blur a difference which stands out more clearly in all those more
complex cases of discourse linking through adversative connectors presented
above.
Except for the few cases of postposed ‘however’, indicating a preceding
contrast, English uses the adversative connectors ‘but’ and ‘however’ in initial
position. Most of the German counterparts are positionally variable and leave the
initial position to a propositional element which is placed initially in line with BID.
3. Attitudinal expressions
Attitudinal expressions like possibly and it is possible are not only variable in
their position — competing with the propositional parts in very much the same
way as adversative connectors — they may be of different degrees of explicitness, varying between adverbs or particles and matrix clauses in a regular way.
In some cases there are semantic differences between adverbs and matrix
clauses. Thus, for example offensichtlich in the matrix clause expresses an
‘objective’ attitude, cf.:
(72) Es ist offensichtlich, daß der Anteil dieser Elemente an den Staubteilchen
wesentlich geringer ist, als ursprünglich angenommen wurde.
while in
(73) Der Anteil dieser Elemente an den Staubteilchen ist offensichtlich wesentlich
geringer, als ursprünglich angenommen wurde.
it expresses the speaker’s opinion. Other cases of contextually equivalent
paraphrases are illustrated in the following examples:
(74) Es ist wahrscheinlich, daß zusätzliche Gaben […] zur Stärkung der Abwehrkräfte des Körpers beitragen würden.
<LINK "doh-r12">
Discourse relators and the beginnings of sentences 241
(75) Zusätzliche Gaben […] würden wahrscheinlich zur Stärkung der Abwehrkräfte des Körpers beitragen.
(76) Es ist möglich …, daß diese Schichten sogar in einem Zusammenhang mit
[…] stehen.
(77) Möglicherweise stehen diese Schichten […] sogar in einem Zusammenhang
mit …
But even if they are equivalent in meaning, paraphrases with different categorial
or structural status (Halliday 1985) are used differently. The positional differences are subject to GIN and BID and their language-specific constraints. The
more explicit versions help to avoid processing difficulties, which are associated
with those constraints. The following sections will concentrate on adverbs and
attitudinal clauses expressing epistemic attitudes of probability and possibility.
This time, German and English sentences show more positional similarities, as
attitudinal expressions in English may also be used in medial position.
3.1 Attitudinal adverbs in German
The strategy of BID determines also the position of attitudinal adverbs, which
compete with propositional candidates for initial position. As the attitudinal
expression always carries new information, it will ‘beat’ any propositional
candidate with given information.
In a text on genetics, in which the subject of the following sentence represents the immediately given discourse topic, we find the attitudinal expression
in initial position:
(78) Höchstwahrscheinlich lassen sich genetische Fehler beim Menschen noch
nicht im einzelligen Stadium der befruchteten Eizelle korrigieren (obwohl
dies theoretisch ideal wäre).
and not in medial position:
(79) #Genetische Fehler beim Menschen lassen sich höchstwahrscheinlich noch
nicht im einzelligen Stadium der befruchteten Eizelle korrigieren (obwohl
dies theoretisch ideal wäre).
However, if the competing propositional element is more relevant, it will come
first. For example, if the subject expresses contrastive information as in the
following example, where zusätzlich modifies what is otherwise a subject with
contextually given information, the attitudinal adverb is used in medial position:
242 Monika Doherty
(80) Zusätzliche Gaben von richtig dosiertem Interferon würden wahrscheinlich
bei der Behandlung vieler Viruserkrankungen zur Stärkung der Abwehrkräfte des Körpers beitragen.
The alternative distribution would be less discourse appropriate as it would be
more difficult to perceive the contrastive part in the subject:
(81) #Wahrscheinlich würden zusätzliche Gaben von richtig dosiertem Interferon
bei der Behandlung vieler Viruserkrankungen zur Stärkung der Abwehrkräfte des Körpers beitragen.
The general assumption that attitudinal adverbs indicate the border between
background and focus in German has to be relativized, at least for sentences with
more than two informational segments. It is the more relevant propositional
beginning which the medial adverb separates from the remaining information.
Thus, it is the background which comes after the initial and the medial adverb.
The distribution is different when a discourse topic is extended by additional information which is of little relevance. Although its local attribute is
introduced only now, the otherwise given subject in the following sentence is
used in medial position:
(82) Möglicherweise stehen diese mehrere hundert Kilometer tiefen Schichten
unterhalb der Wolkendecke sogar in einem Zusammenhang mit dem
Großen Roten Fleck […]
and not initially:
(83) #Diese mehrere hundert Kilometer tiefen Schichten unterhalb der Wolkendecke stehen möglicherweise sogar in einem Zusammenhang mit dem
Großen Roten Fleck […]
3.2 Attitudinal expressions in English
The most striking difference between German and English concerns the degree of
explicitness of attitudinal expressions in initial position. Almost all sentences from
our corpus beginning with adverbs in German begin with attitudinal clauses in
English. While probability is expressed medially by ‘probably’, it turns into ‘it is
likely’ in initial position. Except for the difference in explicitness, the distribution
of competing attitudinal and propositional elements is similar in German and
English. But the different lexical-syntactic conditions will often yield different
sentence structures. The following presents three classes we can distinguish.
Discourse relators and the beginnings of sentences 243
3.2.1 Initial attitudes.
We find the contextually given subject after the attitudinal expression:
(84) It is very unlikely that human genetic defects will be corrected while the
fertilized egg is still in the single cell stage […] (g102)
The alternative distribution is discourse inappropriate as it associates too much
relevance with the subject, which merely continues an immediately given
discourse topic.
(85) #Human genetic defects will probably not be corrected while the fertilized
egg is still in the cell stage […]
An attitudinal adjective with an infinitival complement in medial position
would not be an improvement either:
(86) #Human genetic defects are not very likely to be corrected while the fertilized egg is still in the cell stage […]
Adverbs like ‘possibly’and ‘probably’ do not normally occur in initial position
(for a possible reason see 4.):
(87) #Probably, human genetic defects will not be corrected while the fertilized
egg is still in the cell stage […]
Let us look at another example. The English original of (82) uses the attitudinal
expression in initial position for the same reason as the German translation.
Again, we find a clause in English (one of the rare cases of epistemic ‘possible’
in my material):
(88) It is possible that the deeper layers of the atmosphere beneath the visible
clouds could be associated with the large scale features […] (j90)
The alternative distribution would be less appropriate:
(89) #The deeper layers of the atmosphere beneath the visible clouds could
possibly be associated with the large scale features […]
The positional choice in English is subject to the same strategy as that of the
German translation, localizing the given subject after the attitudinal expression
with its new information.
A striking difference between English and German is the double expression
of possibility by ‘possible’ and ‘could’. It seems to be quite frequent and could
serve to strengthen the epistemic interpretation of the modality. German would
244 Monika Doherty
also need it if it were to use an attitudinal clause, which is more vague in this
respect than the adverb.
3.2.2 Medial adverbs.
If there is a serious propositional candidate for the initial position, the attitudinal expression occurs in medial position in the form of an adverb in English
and German. Thus, for example
(90) Without some such stirring this gas would probably have all condensed
into stars long ago […] (n43)
begins with a contrastive subject. Although the nominal head of the initial
phrase refers to an immediately given eventuality, the preposition without
causes it to participate in a contrastive discourse relation. The attitudinal
expression follows in medial position, where it takes on the form of an adverb
just as in German:
(91) Ohne diese Bewegung hätte sich alles Gas wahrscheinlich schon vor langer
Zeit zu Sternen verdichtet […]
But in English, the medial adverb may also be replaced by clause internal ‘likely’
+ infinitive, if this helps to avoid repetition. Both beginnings in (92) express
new/contrastive information:
(92) The reason, probably, is that the site of carbon fixation is well insulated
in C4 species: any CO2 formed within the vascular bundle cells […] is
likely to be trapped by another PEP molecule on its way through the
cytoplasm. (f59)
The subject of the first clause introduces a new discourse topic, the subject of the
second continues a given topic but presents a summing up of previous paragraphs. The attitudinal expression is used medially, in the first case as an adverb
before the finite verb, in the second in the form of the attitudinal adjective
‘likely’ to avoid a direct repetition of ‘probably’ in will probably be trapped.13
3.3 Restructuring
Except for the different degrees of explicitness, the positioning of attitudinal
expressions follows the same rules in English and German. But as there are also
other factors participating in the distribution of information, discourse appropriate structures may turn out differently. Thus, for example
Discourse relators and the beginnings of sentences 245
(93) The major source of energy for the Jovian plumes is probably released in
layers of water clouds hidden by the other visible clouds (j112)
refers to part of a given discourse topic, which deserves the initial position due
to the narrow focus on ‘major’. The perspective of the copular sentence is
reversed in the translation, which turns the given referent into the object of an
ergative verb14 and adds the focused part in the form of an adverbial vor allem:
(94) Wahrscheinlich stammt die Energie der Stürme vor allem aus Wasserwolken, die sich unterhalb der sichtbaren Wolken befinden
This changes the information of the subject into a pure background, leaving the
initial position to the attitudinal adverb.
A similar, but more complex case is the following sentence, where the
translation reorders the clauses. In the English original, we find the causal
clause at the end, which makes the subject of the main clause a potential
candidate for the beginning. But as it is only given information, the attitudinal
expression comes first and takes the form of a clause:
(95) It is very likely that the Great Red Spot behaves in a similar way as this
feature is now only 24000 km long compared with 46000 km a century
ago. (j100)
In German, the clauses are reordered in line with GIN as the discourse characterizes the information in the main clause as more relevant than that of the
subclause. But the causal clause as a whole contains sufficiently relevant
information to fill the initial position and push the attitudinal particle into the
medial position:
(96) Da der Große Rote Fleck im Vergleich zu 46000 km vor hundert Jahren
heute nur noch 24000 km mißt, verhält er sich wohl ähnlich […]15
4. Summary
The discourse conditions for the beginnings of sentences are the same in
English and German. Both languages use pragmatic strategies like GIN and BID
to achieve a discourse appropriate distribution of information. But the beginnings vary greatly in original sentences and their translations as the constraints
on preposing are much stronger in English than in German. GIN is normally
followed in binary information units in German, but not in English if the result
<LINK "doh-r8">
246 Monika Doherty
were to violate the grammatically determined word order. Information units
with more than two segments of differing discourse relevance require the verb
phrase to be properly filled in both languages; but they allow an additional
propositional element to be preposed more often in German than in English,
where — due to the fixed order of subject before verb — this could easily lead
to too much information before a verb with too little information after it.
Discourse connectors like adversative connectors and attitudinal expressions compete with the propositional elements of a sentence. Provided they are
positionally variable, they are used initially before given information and
medially after new or contrastive information. The fact that attitudinal expressions like ‘possibly’ and ‘probably’ take on a greater degree of explicitness in the
initial position of English sentences can also be seen as a variant of BID as it
diminishes the informational load of the doubly filled preverbal position. Other
adverbials or connectors specifying discourse relations at the beginning of
sentences are normally not extended into clauses. Unlike modalizing expressions,
they do not modify the propositional meaning they refer to, but merely place the
propositions in a special discourse relation, whether adversative, additive,
consecutive, explicating or confirming: ‘however’, ‘moreover’, ‘indeed’ […].
Attitudes about the possibility/probability of a state of affairs modify the truth
condition of their proposition. At the beginning of English sentences, where the
attitudinal expression has to share the preverbal position with the subject, a
more explicit structural encoding will secure an extra focus on the attitude.
Concentrating on discourse relators and attitudinal expressions at the
beginning of sentences in translations of popular-scientific texts from English
into German, the study had to leave many questions open — even if we ignore
its quantitative limits. A major question concerns the various conditions
affecting other registers, especially stylistically marked uses of language. But the
most intriguing question concerns the various factors contributing to the
informational values (discursive relevance) of the propositional elements of a
sentence and the formal reflection of these values. Even if some of the regularities concerning the topicalization of arguments and adjuncts (adverbials) can be
spelled out in more detail (as I have done in papers concentrating on propositional elements at the beginning of sentences, cf. Doherty 2003a and b), the
precise nature of the interaction between the grammatical ‘mould’ — the
language-specific properties of noun phrases and verb phrases, for example —
and the discursive links is far from clear.
The internal make-up of roughly equivalent phrases may provide different
conditions for sentence processing and call for different strategies of discourse
<LINK "doh-r6">
Discourse relators and the beginnings of sentences 247
linking despite equivalent discourse properties. A sentence like
(97) However, more recent work on the interstellar abundances has changed
the picture significantly.
presents ‘however’ before contrastive material and not after it as the German
translation does
(98) Neuere Untersuchungen der interstellaren Häufigkeiten haben das Bild
jedoch entscheidend verändert.
But postposing the connector in English results in an ambiguous focus structure of the subject:
(99) More recent work on the interstellar abundances, however, has changed
the picture significantly.
In (99) the contrast could include the complement of the subject, which would
be wrong as the interstellar abundances have been discourse topic for some time.
On the other hand, the contrast on the inherently focused comparative ‘more
recent’ has its own, grammatical indicator so that deviation from the basic
position of ‘however’ as an additional formal indicator is not necessary.
The postverbal position of jedoch in German seems to have no such impact
on the preceding noun phrase. But examples like (97) and (98) suggest that
concepts like contrastiveness may have to be differentiated further to account
for the internal make-up of their linguistic carriers. Nevertheless, Occam’s razor
requires us to keep the profile of our theoretical concepts low. Whether
informational values assigned to segments of sentence meaning can ‘do the
trick’ remains to be seen.
Notes
1. German uses a statal passive to adhere to the given-before-new strategy and an adverb
instead of seems to indicate the boundary between the two information segments.
2. ‘#’ marks a version which was assessed as not discourse-optimal.
3. Birner and Ward (1998) present a great variety of cases which could be summarized under
GIN. However, most of their examples are subject to special licensing conditions (of register,
emphasis and the like) which are rare in the popular scientific texts we are focusing on.
4. It may be noted in passing that the principle of end-weight (see e.g. Quirk et al 1985),
which is an important word order strategy for the right-hand side of English sentences, does
not apply here as its result would mislead us into a syntactic ‘gardenpath’:
<LINK "doh-r1">
"doh-r14">
"doh-r7">
"doh-r15">
248 Monika Doherty
(i) It shifted hemisphere between the Pioneer 10/11 flybys of 1973–74 and the Voyager flybys of
1979.
5. In English, this would only lead to a heavy cluster before the verb :
(i) ##By the presence of sulphate particles or sulphuric acid from volcanic acid rain, the
conductivity is strongly affected.
6. Although an active version would distribute the information in line with BID, too
(i) Schwefelpartikel oder Säure aus vulkanischem Säureregen verändern die Leitfähigkeit stark,
it assigns a marked (narrow) focus to the degree modifier, which blocks ‘integration’, i.e. a
neutral interpretation, in the sense of Jacobs (1993).
7. In the sense of Büring’s (1998) implicational topic.
8. In line with the basic word order, it is the hierarchically highest adjunct which is
topicalized. Cf.:
(i) Vulkanisches Material kommt jedoch entlang der aktiven mittelozeanischen Rücken aus dem
Mantel unterhalb der Erdkruste zusätzlich ins Meer.
9. Bengt Altenberg suggests that the contrastive predicate attracts the connector, which
explains the verb phrase initial position — in line with what is generally assumed in German
linguistics. But as contrastiveness would not cover all cases (cf. e.g. (35), (39), (42)), I would
like to retain my ‘competition model’ for the time being.
10. For a detailed semantic-pragmatic analysis of aber and similar connectors, which the
present paper largely shares, cf. Lang 1991, 2000.
11. Used with inversion, doch needs an additional focus marking if the postverbal element is
contrasted, as for example in:
(i) Doch könnte gerade dieses Gas in einem Vierteljahrhundert, wenn […], in unseren
Haushalten Verwendung finden.
In this case, doch and dennoch are synonymous.
12. As a rule, the most relevant element of a sentence, its main focus, is characterised by the
main stress of the sentence on the so-called ‘focus exponent’, which is mostly placed in a
verb-adjacent position (for a cross-categorial presentation of this idea cf. Abraham 1992).
Due to the right-peripheral position of German verbs in basic sentences, we expect the main
focus more to the end of a German sentence. This clashes with the verb-second rule in
German main clauses if the verb itself is the focus exponent, as in (67).
13. Only coordinated parallel structures license a repeated use of the same attitudinal
expresssion:
(i) Of the historically-recorded supernovae […] the 1054 supernova was probably Type II, and
the 1572 and 1604 supernovae […] were probably Type Is.
14. The original may be somewhat ‘liberal’ as source and release are not compatible in English
either.
15. It may be interesting to note that the following sentence
(i) Perhaps the Great Red Spot will disappear one day, too.
is subject to the same conditions as the preceding sentence, but replaces the attitudinal
expression at the beginning by an adverb so as to avoid repeating it. In German, it is the
<DEST "doh-r14">
"doh-r10">
"doh-r11">
"doh-r12">
"doh-r13">
"doh-r1">
"doh-r2">
"doh-r3">
"doh-r4">
"doh-r5">
"doh-r6">
"doh-r7">
"doh-r8">
"doh-r9">
Discourse relators and the beginnings of sentences 249
repetition of the subject which is felt to be inappropriate and avoided by a coordinative
conjunction:
(ii) verhält er sich wohl ähnlich und wird vielleicht auch eines Tages nicht mehr da sein.
The attitudinal expressions, which are used medially as they underlie the same discourse
conditions, differ in German anyway.
References
Abraham, W. 1992. “Structural Properties of Information Packaging in German and in
Universal Grammar”. Groninger Arbeiten zur germanistischen Linguistik 35: 37–71.
Altenberg, B. 1998. “Connectors and Sentence Openings in English and Swedish”. In Corpora
and Cross-linguistic Research, S. Johansson and S. Oksefjell (eds), 115–143. Amsterdam/Atlanta: Rodopi.
Asher, N. 1993. Reference to Abstract Objects in Discourse. Dordrecht, Boston, London:
Kluwer Academic Publishers.
Bader, M. 1996. Sprachverstehen: Syntax und Prosodie beim Lesen. Opladen: Westdeutscher
Verlag.
Baker, M. C. 1988. Incorporation: A Theory of Grammatical Function Changing. Chicago:
University of Chicago Press.
Birner, B. J. and Ward, G. 1998. Information Status and Noncanonical Word Order in English.
Amsterdam/Philadelphia: John Benjamins.
Büring, D. 1998. “Topic”. In Focus: Linguistic, Cognitive, and Computational Perspectives, P.
Bosch and R. van der Sandt (eds), 142–165. Cambridge (MA): Cambridge University
Press.
Doherty, M. 2001. “Cleft-like Sentences”. Linguistics 39–3: 607–638.
Doherty, M. 2003a. “Parametrized Beginnings of Sentences in English and German”. Across
Languages and Cultures 4(1), 19–51.
Doherty, M. 2003b. “Topikalisierungsstrategien aus der Perspektive diskursadäquater Übersetzungen”. Linguistische Berichte 194, 183–212.
Drubig, H. B. 1997. Some Cross-Categorial Generalizations on Focus Structure. Stuttgart/
Tübingen: Arbeitspapiere des SFB 340, Bericht 109 [Sprachtheoretische Grundlagen für
die Computerlinguistik].
Frey, W. and Pittner, K. 1998. “Zur Positionierung der Adverbiale im deutschen Mittelfeld”.
Linguistische Berichte 176: 489–534.
Frey, W. and Pittner, K. 1999. “Adverbialpositionen im deutsch-englischen Vergleich”.
Studia Grammatica 47: 14–40.
Haider, H. and Rosengren, I. 1998. Scrambling. Univ. Lund / Germanistisches Inst.: Arbeitsberichte 49 [Sprache und Pragmatik].
Halliday, M. A. K. 1985. An Introduction to Functional Grammar. London: Edward Arnold.
Hornby, A. S. 1995. Oxford Advanced Learner’s Dictionary. Fifth edition. Oxford: Oxford
University Press.
Jacobs, J. 1993. “Integration”. In Wortstellung und Informationsstruktur, M. Reis (ed),
63–116. Tübingen: Niemeyer [Linguistische Arbeiten 306].
<DEST "doh-r19">
"doh-r15">
"doh-r16">
"doh-r17">
"doh-r18">
250 Monika Doherty
Lang, E. 1991. “Koordinierende Konjunktionen”. In Semantik, A. v. Stechow and D.
Wunderlich (eds), 597–622. Berlin/New York: Mouton de Gruyter [Handbücher zur
Sprach- und Kommunikationswissenschaft, Bd. 6].
Lang, E. 2000. “Adversative Connectors on Distinct Levels of Discourse: a Re-examination
of Eve Sweetser’s Three-level Approach”. In Cause — Condition — Concession —
Contrast, E. Couper-Kuhlen and B. Kortmann (eds), 235–256. Berlin/New York:
Mouton de Gruyter.
Pittner, K. 1999. Adverbiale im Deutschen: Untersuchungen zu ihrer Stellung und Interpretation. Tübingen: Stauffenburg [Studien zur deutschen Grammatik 60].
Quirk, R., Greenbaum, S., Leech, G. and Svartvik, J. 1985. A Comprehensive Grammar of the
English Language. London/New York: Longman.
Rosengren, I. 1991. “Zur Fokus-Hintergrund-Gliederung im Deklarativsatz und im
w-Interrogativsatz”. In Fragesätze und Fragen, M. Reis and I. Rosengren (eds), 175–200.
Tübingen: Niemeyer [Linguistische Arbeiten 257].
Sperber, D. and Wilson, D. 1986. Relevance. Oxford: Blackwell.
Appendix
Extracts from the Berlin translation corpus
(1) = m66
(m52) Initially, engineers would have to bring from Earth about 63 tonnes of hydrogen,
fluorine and sodium to provide some of the raw materials. (m53) Thereafter, these materials
could be recovered from the plant’s products.
…
(m63) Finally, the product is formed by applying pressure or temperature or being allowed
to set for a long time. (m64) People on Earth have been using powder metallurgy since 1200
BC. (m65) But two factors make it look promising in Moon factories. (m66) First, in space
the surfaces of powdered materials remain in pristine condition for a long time; there is very
little to contaminate them. (m67) Thus, thanks to the process of cold welding, the particles
will adhere well to each other to give large masses of pure material.
(5) = g67
(g66) Although a lot can be learned about gene control by putting new genes into cultured
cells, the most realistic approach would be to incorporate the genes into living animals. (g67)
Several research groups have recently achieved this, including Franklin Constantini and
Elizabeth Lacy at Oxford University’s Department of Zoology. (g68) They removed newly
fertilized eggs from mice and injected into the nuclei a gene coding for a protein called betahaemoglobin.
(88) = j90
(j86) On Jupiter, the corresponding “meteorological region” must be much deeper for the jet
streams to remain so stable while the visible features are constantly changing. (j87) But how
deep is it? (j88) The gaseous atmosphere is about 1000 kilometres thick, and below that lies
</TARGET "doh">
Discourse relators and the beginnings of sentences
70 000 km or so of liquid hydrogen. (j89) Consequently, we must give some future thought
to the possible involvement of the deeper layers of the Jovian atmosphere, which may cause
the star-like interior of the planet to assist in the production of terrestrial-like weather
systems. (j90) It is possible that the deeper layers of the atmosphere, a few hundred kilometres beneath the visible clouds, could be associated with some of the large-scale features,
such as the Great Red Spot, which has been the centre of debate for many centuries. (j91)
Voyager’s observations have confirmed that the Great Red Spot is not special, but simply the
largest of a family of meteorological features.
Sources
Berlin translation corpus (articles published in New Scientist)
e: electricity
f: photosynthesis
g: genetics
h: hydrogen
i: interferon
j: jupiter
m: moon
n: novae
s: saltsea
v: volcano
“The electricity that shapes our ends”. 28 January 1982: 217–220.
“The varied ways plants tap the sun”. 12 February 1981: 394–397.
“The reconstruction of animals and plants”. 26 August 1982: 562–564.
“The hidden strength of hydrogen”. 30 July 1981: 291–293.
“Interferon: a progress report”. 25 March 1982: 783–785.
“The weather on Jupiter”. 21 May 1981: 485–487.
“Working on the Moon”. 1 Oktober 1981: 30–33.
“Bigger and better bangs in the sky”. 12 August 1982: 431–434.
“The salt sea — accident or design”. 1 April 1982: 14–17.
“Do volcanoes affect the climate?” 21 January 1982: 150–153.
Author’s address
Prof. Dr. Monika Doherty
Humboldt-Universität zu Berlin
Institut für Anglistik und Amerikanistik
Unter den Linden 6
10099 Berlin
Deutschland
[email protected]
251