Local grammar and translation equivalents - Preliminary findings for consider and its German translations

Renate Reichardt
University of Birmingham


The premise of this paper is that the local grammar of words, specifically their syntactic valency complements, guides the choice of translation equivalents. The hypothesis is that the preferred translation equivalent of a word is the one which is closest to the valency pattern of the original word. Valency theory is concerned with the local grammar of words, i.e. with the property of a word to combine with a certain number of elements in forming larger units (Emons 1974). A corpus linguistic approach was chosen for the exploration of the relationship between the valency patterns of a word and its corresponding meaning in another language. For exemplification the languages English and German were investigated, but the approach is equally suitable to explore the grammar-lexis interplay between a wide range of languages. The case study examines the polysemous verb consider and its German equivalents. Frequency analysis showed a preference of the syntactic valency patterns of consider for certain translation equivalents. Valency patterns are thus a useful indicator of likely translations into another language, a finding which can be useful in the second language classroom, in translation training or in the compilation of dictionaries and grammars. On the other hand, the contrastive analysis has also shown that there is great overall flexibility in the choice of translation equivalents, which illustrates that what is often considered a straightforward rule-based construction process is much more flexible and unpredictable.

1. Assumptions and methods

The research is inspired by Firth (1957) and Sinclair (1991), who argued that meaning depends on the environment in which a word occurs. The purpose of this investigation is to explore the lexis-grammar interface from a contrastive perspective. Focusing on the local grammar of words, i.e. their syntactic valency patterns, a corpus investigation into the polysemous verb consider and its German translation equivalents is undertaken.

1.1 Meaning, paraphrase and translation

Any language is based on conventions amongst its users. Words do not in themselves carry an inherent meaning, but the meaning, usually expressed as a paraphrase, is negotiated in the discourse and thus acquired by language users (Teubert 2004). In contrastive linguistics the paraphrase expressing the meaning of a word is the chosen translation equivalent. This study differs from similar studies (e.g. Duffner et al. 2009; Noël 1996) in that translation equivalents are established from a multilingual corpus (EuroParl), rather than being taken from existing dictionaries or being based on monolingual meaning interpretations.

Most words are polysemous, i.e. they have more than one meaning depending on the environment of usage. Thus, when learning the meaning or meanings of a word, users also acquire knowledge about the environment of a word (Hoey 2005), i.e. its usage context, typical occurrences (collocations) and its structural use (grammar and colligations). Valency theory is a suitable method for the contrastive analysis of syntactic and semantic investigations between languages.

1.2 Syntactic and semantic interpretation of language

The traditional distinction between lexis and grammar as separate areas of study has changed towards a theory of a lexis-grammar continuum (e.g. Römer & Schulze 2009; Sinclair 1991) which emphasises the strong interaction between the two. However, as noted by Fischer (1997), it will always be debatable whether syntax influences semantics or vice versa, and, secondly, to what extent syntax and semantics are independently capable of analysis. It is noteworthy that various linguistic theories tend to emphasise the one or the other level, thereby showing the philosophical stance of the linguists concerned. Furthermore, no generalisations can be made regarding congruence of syntactic and semantic properties of words – these relationships are based on the local grammar, i.e. the individual properties, of words.

1.2.1 Valency theory

Valency theory relates to local grammar, i.e. grammar that is specific to the individual word or lexical item and cannot be explained by the general grammar rules of a language (Teubert 2003). Cornell et al (2003: 8; see also Welke 1988; Engel and Schumacher 1976; Tesnière 1959) summarize the valency approach as follows:

  • Lexical items have the power to structure their surroundings syntactically and semantically.
  • Sentences are organised bottom-up, from words to larger units.
  • Lexical items, in particular the verb, demand complements to create phrases that are syntactically and semantically complete. Adjuncts can be added freely, giving additional information.

The basic assumption of valency theory is that the verb occupies a central position in the sentence because the verb determines how many other elements have to occur in order to form a grammatically correct sentence (Homberger 2001). Thus, valency patterns primarily represent syntactic patterning. However, complements also have semantic functions, since valency is not to be seen simply as a ‘slot-and-filler’ theory (Götz-Votteler 2007). Sentence elements expressing syntactic relations are commonly referred to as ‘complements’, while elements expressing semantic relations are called ‘arguments’. Probably because of this dual aspect, Sinclair (2004: 18) predicted that “valency grammar ... is likely to see an upsurge of interest.” The sentence elements which could occur with almost any verb in any sentence are generally referred to as adjuncts. It should also be noted that the current understanding of valency is not restricted to verbs but can be applied to all word-classes, e.g. nouns (‘consideration for others’ but not ‘belief for others’) or adjectives (‘considerate of others’ but not ‘contemplative of others’). [1]

Recently, valency theory has received fresh interest in British and American linguistics as it forms part of the analysis of constructions (Goldberg 2006; Fillmore 2007). However, the approach to valency, or ‘valence’, applied in construction grammar focuses on the semantic aspects, mainly the semantic roles of the sentence elements. In contrast, the focus in this investigation is on the syntactic valency patterns and their function in meaning identification, i.e. the translation choices.

Due to its versatility and the possibility to investigate language largely independently, at various levels, valency theory is a suitable theoretical framework to undertake contrastive studies.

1.3 Corpus investigation

The corpus used for the cross-linguistic English-German investigation is the EuroParl corpus, which consists of European Parliament Proceedings published in the 11 official languages of the European Union and thus represents a specific language domain. In order to validate the identified syntactic valency patterns of the verb consider and those of its translation equivalents (TEs) in the EuroParl investigation the findings are compared with monolingual reference corpora. The reference corpus for English is the Bank of English corpus (BoE; corpus at the University of Birmingham), and for German the Deutsches Referenz Korpus (DeReKo; corpus at the Institute of German Language, Mannheim). Such a comparison with a reference corpus will identify any possible idiosyncrasies of word use, i.e. the syntactic patterning, in EuroParl.

Differences between corpora are not uncommon in corpus linguistics. These may be due to the different compositions of corpora. For example, while EuroParl is a spoken corpus where mental verbs are an important device used to express stance, the reference corpora are mixed corpora including a number of different genres. Table 1 shows a comparison of the frequencies per million words of consider, its word-forms, and two complementation patterns, a non-finite ing-clause (example sentence 1) and a that-clause (example sentence 2), in EuroParl and the BoE.

(1) THINGS have been going so badly for Marks & Spencer, they should consider hiring a female marketing boss. (BoE)
(2) We consider that the expenditure relating to the agreement is compulsory. (EuroParl)
EuroParl BoE

per million


per million





































Table 1. Occurrences consider in EuroParl and BoE

It is notable that, overall, the lemma consider is more frequent in EuroParl than in the BoE, and that there are differences in the distribution of the individual word-forms. Also non-finite ing-clauses are more frequent in the BoE, while that-clauses are more frequent in EuroParl.

Due to the differences between various corpora, the validity of generalisations drawn from corpus findings is often questioned. As noted by Leech (1991: 27) the representativeness of a corpus “must largely be regarded as an act of faith”. This study is concerned with a comparison of the syntactic valency patterns between words and those of their preferred translations, therefore genre differences are of no interest as long as the investigated syntactic patterns occur sufficiently frequently in the reference corpora.

For the present study, translation direction is not seen to be relevant since the interest lies in the syntactic differences between an English verb and those of its German counterpart. However, it should be noted that since source language (original) and target language (translation) are unknown in EuroParl, it is, strictly speaking, not correct to talk of TEs in a comparison of different languages in EuroParl. Nevertheless, for simplicity, this report refers to English as the source language and German as the target language, i.e. the TE.

1.4 Frequency analysis

Starting with the verb consider, 200 randomly chosen concordance lines from EuroParl were analysed for their TEs. Occurrences where a form of consider functions as adjective, as in example sentence 3, were excluded, since this study is concerned with verb valency. Excluded lines were replaced.

(3) I hope that the House will, on reflection, accept the considered view of the Commission.

In the next step, the possible valency sentence patterns for consider were identified. For this, the 200 lines which were used for the identification of the TEs were analysed. Additionally, 200 lines from the BoE were analysed for reference.

Having established the valency sentence patterns and the preferred TEs, it was possible to investigate whether the TEs occur with the same or a different pattern. Two approaches were available for this investigation:

  • Sort concordance lines according to valency complements and compare these with the TEs.
  • Sort concordance lines according to TEs and compare the valency complements of original and TE.

The first option would require that EuroParl is annotated or parsed for the identified valency complements. For this reason the second option was pursued.

In the next step, the previously identified TEs for the verb consider were extracted. For example, the analysis of the 200 lines from EuroParl showed halten as the most frequent TE for consider. A search in EuroParl for consider and halten produced 1,730 concordance lines. It has to be noted that the actual number of occurrences is lower, since ParaConc, the concordance programme used for the investigation, will look for search words at sentence level mis-hits, as demonstrated in example sentence 4, are included. [2]

(4‑E) Two major areas of concern about the proposal were considered by the Committee.
(4‑G) Der Ausschuß befaßte sich mit zwei wichtigen Aspekten des Vorschlags, die er für problematisch hielt.

As can be seen, consider is actually translated as befassen (single underlining), whereas halten refers to ‘of concern’ (bold underlining).

50 concordance lines, every n-th occurrence, of consider with a specific TE were initially investigated. For example, from the 1,730 lines of consider with the TE halten 50 lines were extracted and transformed into active canonical clauses without adjuncts as demonstrated in example sentence 5.

(5‑E) Our airport is very close to housing and, like 20% of Europe’s citizens, we suffer levels of noise from aircraft which health experts consider to be unacceptable.
(5a‑E) Health experts consider the levels of noise from aircraft to be unacceptable.
(5‑G) Unser Flughafen liegt ganz in der Nähe der Wohngebiete, und wie 20% der europäischen Bürger leiden wir unter einer Fluglärmbelastung, die Gesundheitsexperten für untragbar halten.
(5a‑G) Gesundheitsexperten halten die Fluglärmbelästigung für untragbar.

These 50 transformed lines were then analysed for the syntactic patterns of the original and the TE (figure 1).

Figure 1

Figure 1. Extract from valency comparison of consider and the TE halten

The question which arises at this point of the investigation is whether 50 concordance lines are sufficient to produce reliable and viable findings. For this reason, a further three sets of 50 concordance lines were analysed for the two most frequent TEs of consider, viz. halten and betrachten. As can be seen in Table 2, the frequent patterns for a TE of consider are the same in all four data sets, although the rank order changes slightly between the sets. Therefore, it seems sufficient to work with 50 concordance lines, especially given that the remainder of the TEs are less frequent than halten and betrachten. A similar approach was used by Groom (2007) in order to analyze semantic sequences of salient grammatical words.

Table 2

Table 2. Valency pattern distribution in four samples (of 50 concordance lines each) of consider with the most frequent TEs halten and betrachten

In the final step of the investigation the syntactic valency patterns of the German TEs were compared with the patterns of consider.

It is believed that the approach taken is sufficiently reliable to identify trends regarding the interplay of local grammar and word meaning, i.e. TEs. However, this investigation does not claim to be exhaustive, but attempts to provide an overview of the key principles in contrastive linguistic research using valency theory and corpus investigation.

2. Valency complements in contrastive linguistics

Valency does not relate to general properties of words but to specific subclasses of words. For example, in English a distinction is often made between subclasses of verbs typically followed by an -ing or a to-inf-form, while German distinguishes between verbs typically followed by a sentence element in the accusative case and those typically followed by a dative (Table 3, example sentences 6 to 9).

English Example verb Example sentence

-ing form

(function: object)


(6‑E) The European Parliament should consider contributing to the introduction of a fundamental change in attitude in this area.

(6‑G) Das Europäische Parlament muß in Betracht ziehen, an einer solchen prinzipiellen Veränderung mitzuwirken.

to-infinitive form (function: object)


(7‑E) Cohesion policy needs to be strengthened further.

(7‑G) Die Kohäsionspolitik muß weiter gestärkt werden.

German Example verb Example sentence

Accusative complement
(function: object)

halten (consider)

(8‑G) Ich persönlich halte ihn für eine sehr gefährliche Person.

(8‑E) I personally consider him to be a very dangerous politician.

Dative complement

(function: object)


(9‑G) Die damit verbundenen Stabilitäts­vereinbarungen helfen dem Eurosprößling sich klar nach vorne blickend zu entfalten.

(9‑E) The stability pacts will help the young euro to develop with a clear eye on the future.

Table 3. Examples of subclasses of verb valency in English and German

As can be seen in Table 3 different forms fulfil the same function; viz. that of object (Quirk et al. 1985). [3] Sub-classification of realisation forms, as shown in Table 4 example sentence 10, helps to show the link between function and form between languages. The English ing-complementation functioning as object can be labelled <obj-ing>. Its German counterpart is realised in German with a wh-clause which also functions as object and can be labelled <obj-wh>.






raising the matter with the Turkish authorities,

together with other issues concerning regional development policies in the framework of the new pre-accession strategy for Turkey.











inwiefern wir diese Angelegenheit ...

... ansprechen werden.

... sowie weitere Probleme im Zusammenhang mit der regionalen Entwicklungspolitik im Rahmen der neuen Heranführungsstrategie bei den türkischen Behörden ...






Table 4. Example analysis of valency complement types for English and German

Differences in the realisation forms, i.e. valency patterns of words, exist between different languages and often make a contrastive analysis difficult. The parameters for the analysis of different languages may vary as language specific characteristics need to be accounted for (Emons 1974). On the other hand, a contrastive analysis needs to be based on homogeneous criteria for the classification of the valency complements (Bianco 1980). It is therefore important to use the same categories or labels for the classification of valency complements in order to make the comparison between languages as transparent as possible. That valency theory never experienced a breakthrough in the analysis of English might have to do with the morphological properties of the language. English has mostly lost the noun-inflections indicating the cases which are generally seen as parameter for the analysis of valency complements (Polenz 2008).

German English

Case complements


Nominative complement; main function: subject

Subject complement




Accusative complement; main function: direct object

Object complement


Genitive complement



Dative complement; main functions: indirect object or direct object

Indirect object

Prepositional complements

Prepositional complement

Prepositional complement

Adverbial complements

Situational complement

Situational complement


Directional complement

Directional complement


Expansive complement

Expansive complement

Predicative complements

Nominal complement

Nominal complement


Adjectival complement

Adjectival complement



Modificational complement

Verbal complements

Verbal complement

Verbal complement

Table 5. Valency complements in German and English (based on Engel 2009 and Fisher 1997)

Table 5 shows a comparison of the core syntactic valency complement types for English and German. A compromise had to be made with regard to the German case complements as their function and use in German is not fully congruent with the chosen categories of subject complement <sub> and object complement <obj>. However, since in the majority of occurrences the German cases can be matched to a respective syntactic function in English, function categories seem to be the most suitable categorisation to accommodate both languages. Thus, the nominative is generally equivalent to the subject <sub> in English, the accusative mainly corresponds to the direct object <obj>, and the dative often matches the indirect object <ind> in English (Fischer 1997). These eleven English and twelve German valency complement types are sufficient to analyse any clause or sentence regarding its valency sentence patterns. For the various individual verbs there are only a limited number of possible combinations (Schumacher et al. 2004); depending on how many complement types a verb can occur with, its valency number can be established as mono-, di-, tri- or tetravalent.

2.1 The Valency Sentence Patterns of consider

Table 6 shows the 17 valency sentence patterns identified for the verb consider based on 200 randomly chosen concordance lines. As can be seen, the core valency types are further differentiated depending on their realisation form. For example, the object in the di-valent pattern has four realisation forms: with a noun phrase <sub obj>, with a that-clause <sub obj-that>, with a wh-clause <sub obj-wh> and a non-finite ing-clause <sub obj-ing>.



(11) The Commission should hear the sector’s views, consult, listen, consider.


<sub obj>

(12) We / consider / exchange rate mechanisms.

<sub obj-that>
(13) I / do not consider / that the Council tried to answer my question.

<sub obj-wh>
(14) We / consider / how the European Union might be provided with a constitution.

<sub obj-ing>

(15) We / consider / revising the Structural Funds.


<sub obj nom>

(16) We / consider / this agreement / a milestone in future relations with the Latin American countries.

<sub obj adj>

(17) We / consider / the reforms / necessary.

<sub obj nom-as>

(18) The report / considers / labour costs / as the main source of inflation.

<sub obj adj-as>

(19) We / consider / these matters / as tabooed.

<sub obj vb-to-be-nom>

(20) We / considered / building motorways / to be a fundamental complement.

<sub obj vb-to-be-adj>

(21) Health experts / consider / the levels of noise pollution / to be unacceptable.

<sub obj vb-to-inf>

(22) The Presidency / considered / this subject / to fall within the competence of the Committee.

<sub obj prp-for>

(23) The government / considered / him / for a peerage.

With correlate ‘it’ structure

<sub it nom vb-that>

(24) I / consider / it / a scandal / that Europe stands by watching such a thing happen.

<sub it adj vb-that>

(25) We / consider / it / only logical / that funds are made available.

<sub it nom vb-to-inf>

(26) We / consider / it / a bad idea / to take the funding from the farming sector.

<sub it adj vb-to-inf>

(27) We / consider / it / necessary / to discuss this topic.

Table 6. Valency sentence patterns of consider

It should be noted that other researchers may decide on different sub-patterns and different labelling. For example, Noël (1996) uses word-class labels as in ‘consider + NP’, which is equivalent to the pattern <sub obj> in Table 6; he also identifies a category ‘consider + if/whether-clause’ in addition to the category ‘consider + wh-clause’. Both of these structures are combined under the category <sub obj-wh> in Table 6.

An interesting case for discussion is constituted by a probable mono-valent pattern <sub> for the verb consider as shown in Table 6.

(11) The Commission should hear the sector’s views, consult, listen, consider.

The Valency Dictionary of English (Herbst et al. 2004: 175) includes a mono-valent valency pattern which is exemplified with the sentence ‘Cook tilted her head to one side, considering’, while others, such as the Oxford Advanced Learner’s Dictionary (2005: 324), only show it as transitive verb. Francis et al (1996: 1) note that ‘many verbs are used with this [mono-valent] pattern only when something involved in the action, apart from the subject, has already been mentioned.” This seems to clearly imply that most verbs usually occur with an object, i.e. they are transitive, and only occur without the object due to stylistic reasons as the object can be retrieved from the context. For example, sentence 11 could be rewritten as 11a:

(11a) The Commission should hear, consult about, listen to, and consider the sector’s views.

Furthermore, the low frequency of the mono-valent use certainly implies rare usage. For this reason, though shown in Table 6, the mono-valent pattern <sub> is ignored for the verb consider. Occurrences such as sentence 11 are included in the analysis under the bi-valent sentence pattern <sub obj>.

For the tri-valent valency sentence patterns of consider a decision had to be made as to how many sub-categories based on the realisation form should be differentiated. In traditional grammar (e.g. Swan 2005, Biber et al. 2002, Quirk et al. 1985, Allerton 1982) it is often implied that there is apparently no difference in meaning between sentences 16 to 16c and 17 to 17c, thus implying that the different surface structures are merely a stylistic choice.

(16) We consider this agreement a milestone.
(16a) We consider this agreement to be a milestone.
(16b) We consider this agreement as a milestone.
(16c) We consider that this agreement is a milestone.
(17) We consider the reforms necessary.
(17a) We consider the reforms to be necessary.
(17b) We consider the reforms as necessary.
(17c) We consider that the reforms are necessary.

Should one pattern be chosen to represent all the others in these cases? And if so, which one, and based on what rules? An attempt at combining the above example sentences 16 to 16c and 17 to 17c under one valency category would run counter to the inclination of many grammarians of dependency and valency grammar not to manipulate the surface appearance of sentences if at all possible (Fischer 1997), but treat different surface structures in their own right. This approach was chosen in this study.

A variation on the tri-valent valency sentence patterns are occurrences with a correlate or anticipatory ‘it’ structure (example sentences 24–27). Since this structure is relatively frequent for the verb consider it was decided to investigate it separately in the analysis.

2.1.1 Frequency analysis of the valency sentence patterns of consider

consider EuroParl 200 lines
Total %
BoE 200 lines
Total %
Total (%) Total (%)

sub obj

72 (36.0%)

79 (39.5%)

sub obj-that

27 (13.5%)

12 (6.0%)

sub obj-wh

8 (4.0%)

13 (6.5%)

sub obj-ing

9 (4.5%)

28 (14.0%)

sub obj nom

3 (1.5%)

18 (9.0%)

sub obj adj

17 (8.5%)

28 (14.0%)

sub obj nom-as

9 (4.5%)

8 (4.0%)

sub obj adj-as

1 (0.5%)

sub obj vb-to-be-nom

16 (8.0%)

3 (1.5%)

sub obj vb-to-be-adj

20 (10.0%)

5 (2.5%)

sub obj vb-to-inf

2 (1.0%)

3 (1.5%)

sub obj prp-for

1 (0.5%)

1 (0.5%)

sub it nom vb-that

1 (0.5%)

sub it adj vb-that

6 (3.0%)

sub it nom vb-to-inf

sub it adj vb-to-inf

9 (4.5%)

1 (0.5%)


2 (1.0%)

4 (2.0%)


200 (100%)

200 (100%)

Table 7. Frequency of valency sentence patterns of consider

Table 7 shows the valency sentence pattern distribution of consider for 200 randomly chosen concordance lines from the corpora EuroParl and BoE. Due to the different genres of the corpora, variation in the valency pattern distribution was expected. However, despite some differences an overall similar tendency is notable, which indicates that the verb consider occurs frequently with certain syntactic valency sentence patterns irrespective of the context. The most frequent valency sentence pattern for consider in both corpora, highlighted in bold in Table 7, is <sub obj> with the object realised by a noun phrase. This is followed by object realisation with a that-clause (EuroParl) or an ing-clause (BoE). Shortened infinitive clauses representing a nominal or adjectival complement, <sub obj nom> and <sub obj adj>, are frequent in both corpora, but the extended infinitive clauses in the patterns <sub obj vb-to-be-nom> and <sub obj vb-to-be-adj> are relatively more frequent in the EuroParl corpus. In both corpora there were uses of consider as adjective, such as ‘a carefully considered agreement’, which were replaced for the further analysis.

2.2 The translation correspondences of consider in EuroParl

The verb consider has a wide range of TEs in German with 68 different translation possibilities for the 200 concordance lines. Working with the assumption that two or fewer occurrences could be chance occurrences due to the idiosyncracies of the translators, only TEs with more than two occurrences were seen as relevant. This still leaves 20 possible TEs for the verb consider:

halten: 20x
betrachten: 12x
ansehen, nachdenken: 8x each
berücksichtigen, Erachtens: 7x each
bedenken, Betracht, prüfen: 6x each
Auge/Augen, Ansicht, denken, erwägen, finden, Meinung: 5x each
Auffassung, befassen, behandeln, erachten: 4x each
überlegen: 3x

These are either verbs, shown in capital letters representing the verb lemma, or multi-word units such as ‘meines Erachtens’ (Lit.: ‘in my view’) or ‘in Betracht ziehen’ (Lit: ‘take into consideration’). For simplification only the noun (in lower case) of these multi-word units is shown as TE in the above list. The multi-word units either function as support-verb-constructions or as adjuncts. For example, the noun ‘Ansicht’ (Lit.: ‘view’) can occur in the following translations:

(28‑E) As the Committee on Budgets we consider that the establishment of the budget by activities causes us problems.
(28‑G) Als Haushaltsausschuss sind wir der Ansicht, dass uns die Aufstellung des Haushaltsplans nach Tätigkeiten Probleme bereitet .
(29‑E) We consider that this new rule does not guarantee the necessary balance.
(29‑G) Diese neue Regelung garantiert unserer Ansicht nach nicht die erforderliche Ausgewogenheit.

In 28-G ‘Ansicht’ is part of a support-verb-construction with the verb sein ‘be’, the whole TE is ‘der Ansicht sein’ (Lit.: ‘can be considered’), while in 29-G ‘Ansicht’ is part of the phrase ‘pronoun Ansicht nach’ (Lit.: ‘in someone’s view’) which represents an adjunct, functioning as a hedging device that can be added to almost any sentence. [4]

3. Do the syntactic valency sentence patterns of consider relate to TEs?

In the following analysis an investigation into the correspondence of the valency sentence patterns of consider and its most frequent German TEs is undertaken. The key questions to be explored are whether the valency sentence patterns of consider have preferred TEs, and if so, whether these TEs occur with a valency sentence pattern similar to that of consider.

Table 8 shows an extract of the valency sentence patterns of consider and the most frequent TEs of consider. The multi-word TEs are still shown as single words, but a distinction is made as to whether these function as support-verb-constructions or as adjuncts. The occurrence of a TE for a valency sentence pattern of consider is given in percentages, based on 50 randomly chosen concordance lines from the EuroParl corpus for each German TE.

Table 8

Table 8. Valency patterns of consider and translation equivalents

The patterns are colour coded in Table 8 in order to identify whether the TEs show preferences for certain valency sentence patterns of consider. Occurrences of 10% and higher of a TE with a valency sentence pattern of consider are highlighted in the respective colour of the pattern. As can be seen, the TEs seem to cluster around certain valency patterns of consider, indicating a congruence between local grammar and the TEs, i.e. the meaning. Therefore it can be deduced from Table 8 that the TEs are not equally suitable for all the valency sentence patterns the verb consider can occur with. Rather, it seems to be the case that the TEs are dependent on the valency sentence pattern of consider as many TEs show a preference for a valency sentence pattern of consider (occurrences of 50% and higher are in bold). Despite these tendencies it is also notable that all TEs show a spread over several valency sentence patterns of consider, indicating that the choice of a TE is not a rule-based construction process.

The preferred TEs for each of the patterns of consider are summarised in Table 9. The TEs are listed in order of their preference for the pattern.

Valency sentence pattern of consider Preferred TEs

<sub obj>

berücksichtigen, Betracht (in Betracht ziehen), erwägen, prüfen, Auge (ins Auge fassen), nachdenken, Erwägung (in Erwägung ziehen), denken an, bedenken, überlegen, sehen, betrachten

<sub obj-that>

Ansicth (der Ansicht sein), meinen, Auffassung (der Auffassung sein), bedenken, glauben, Meinung (der Meinung sein), Erachtens (pronoun Erachtens)

<sub obj-wh>

überlegen, prüfen, nachdenken

<sub obj-ing>

Erwägung (in Erwägung ziehen), erwägen, Betracht (in Betracht ziehen)

<sub obj nom / adj>

erachten, gelten, ansehen, finden, betrachten, halten

<sub obj nom-as / adj-as>

betrachten, gelten

<sub obj vb>

gelten, halten, betrachten, finden, ansehen, erachten

with correlate ‘it’

halten, erachten, finden

Table 9. Valency sentence patterns of consider and their preferred translation equivalents

In the following sections the valency patterns of consider and those of their preferred TEs will be compared.

3.1 The pattern <sub obj>

The most frequent pattern <sub obj> (Table 7) occurs with the widest variety of TEs (Table 9) and can either be expressed with a verb (30) or a support-verb-construction (31) in German. The German verb berücksichtigen seems to be the preferred correspondence with 78% (39 occurrences) of the 50 concordance lines for consider with this pattern:

(30‑E) In the resolution we have considered all the points which are important for the future. <sub obj>
(30‑G) Wir haben alle Faktoren, die für die Zukunft wichtig sind, in der Entschließung berücksichtigt. (Wir berücksichtigten alle Faktoren.) <sub obj>

The TE ‘Betracht’ in the support-verb-construction ‘in Betracht ziehen’ is the second most frequent option with 74% (37 occurrences) of the analysed concordance lines:

(31‑E) We should at least consider policy changes. <sub obj>
(31‑G) Wir sollten einen Richtungswechsel zumindest in Betracht ziehen.
(Wir ziehen einen Richtungswechsel in Betracht.) <sub obj>

It is notable that the preferred German TEs of consider with the valency sentence pattern <sub obj> are verbs or support-verb-constructions that also occur with the valency sentence pattern <sub obj>. The only exception is the verb nachdenken which generally occurs with a prepositional complement (Institut für Deutsche Sprache):

(32‑E) There are a lot of things people could consider, but ... <sub obj>
(32‑G) Es gibt vieles, über das man nachdenken sollte, aber ... <sub prp>

Thus, the choice of nachdenken requires a change in the valency sentence pattern between the English and the German sentence structure.

3.2 The pattern <sub obj-that>

Generally, the English valency pattern <sub vb-that> for consider is used to express or report someone’s opinion. The most frequent TEs also occur with the corresponding dass-clause in object position in German. Most frequently, the corresponding German expressions are support-verb constructions, such as ‘der Ansicht / Meinung / Auffassung sein’ as shown in example sentence 33, or verbs, for example meinen, bedenken, finden or glauben, exemplified in sentence 34.

(33‑E) On the other hand, the rapporteur considers that neither Amendment No 10 nor Amendment No 11 belong in a regulation about help with marketing. <sub obj-that>
(33‑G) Andererseits ist der Berichterstatter der Ansicht, daß weder Änderungsantrag 10 noch Änderungsantrag 11 in eine Verordnung über Beihilfen zur Vermarktung gehören. <sub obj-that>
(34‑E) We consider that international law must be respected. <sub obj-that>
(34‑G) Wir meinen, dass das Völkerrecht respektiert werden muss. <sub obj-that>

There are two exceptions to the observed regularity. The first is the verb halten, which takes a trivalent valency sentence pattern, either with a nominal complement <sub obj nom> or an adjectival complement <sub obj adj> as shown in example sentences 35 and 36.

(35‑E) We considered that increasing confidence was one of the central tasks.<sub obj-that>
(35‑G) Wir hielten eine Erhöhung des Sicherheitsgefühls für eine zentrale Aufgabe. <sub obj nom>
(36‑E) The Commission considers that a less systematic spot-check system is adequate. <sub obj-that>
(36‑G) Die Kommission hält ein weniger systematisches Stichprobenkontrollsystem für angemessen. <sub obj adj>

In the case of halten the corresponding meaning expression for consider is actually the multi-word verb halten für, which is in this investigation treated as a meaning unit. The second exception is the translation as an adjunct, either ‘pronoun Erachtens’, ‘pronoun Meinung nach’ or ‘pronoun Ansicht nach’ as shown in examples 37 and 38.

(37‑E) I consider that the European Union should pay more attention to these problems. consider <sub obj-that>
(37‑G) Die Europaeische Union sollte meines Erachtens diesen Problemen mehr Aufmerksamkeit schenken. schenken <sub dat acc>
(38‑E) We consider that the Trakatellis report is heading in such a direction. consider <sub obj-that>
(38‑G) Der Bericht Trakatellis geht unserer Meinung nach in eine solche Richtung. gehen <sub prp>

When consider with the sentence pattern <sub obj-that> is translated as an adjunct, the verb of the that-clause becomes the main verb in the corresponding German sentence. Thus, pay in 37 becomes the main verb schenken with the valency sentence pattern <sub dat acc>, and head in 38 with the translation of gehen and its pattern <sub prp>.

3.3 The pattern <sub obj-wh>

The TEs for consider followed by a wh-clause in object position also frequently occur with a wh-clause in German, as in example 39.

(39‑E) We must consider how we are to react to this situation. <sub obj-wh>
(39‑G) Wir werden uns überlegen müssen, wie wir darauf reagieren. <sub obj-w>

It is notable that verbs, very rarely support-verb-constructions and never adjuncts, occur as German correspondences, and that the pattern <sub obj-wh> is retained in the translations.

3.4 The pattern <sub obj-ing>

There is no direct grammatical equivalent in German to the English non-finite ing-clause. The most frequent TEs are the nouns ‘Erwägung’ and ‘Betracht’ in the support-verb-constructions ‘in Erwägung ziehen’ or ‘in Betracht ziehen’ and the verb erwägen (examples 40 and 41).

(40‑E) I think you are considering introducing a safeguard clause. <sub obj-ing>
(40‑G) Sie ziehen die Einführung einer Schutzklausel in Betracht. <sub obj>
(41‑E) The Commission will consider setting up a monitoring unit. <sub obj-ing>
(41‑G) Die Kommission erwägt die Einrichtung einer Forschungsstelle. <sub obj>

As examples 40 and 41 show, the ing-form in object position is generally translated as a noun phrase (introducing = Einführung, setting up = Einrichtung). Therefore it is not surprising that, as seen in tables 8 and 9, the correspondences for consider with the valency patterns <sub obj> and <sub obj-ing> are largely shared.

3.5 The patterns <sub obj nom / adj>

consider followed by an adjectival complement after the direct object is more frequent than by a nominal complement (Table 7). These valency patterns occur most often with the TEs betrachten (als), ansehen (als), halten (für) and finden, as in example sentences 42 to 44.

(42‑E) ... which the Court considers too time-consuming and ...
(The Court considers this too time-consuming.) <sub obj adj>
(42‑G) ... , die der Rechnungshof als zu kompliziert ansieht und ...
(Der Rechnungshof sieht dies als zu kompliziert an.) <sub obj adj-als>
(43‑E) ... allow each of the parties to take measures that it considers essential ...
(The parties consider the measures essential.) <sub obj adj>
(43‑G) ... jedem Vertragspartner gestatten, die Maßnahmen zu ergreifen, die dieser für notwenig hält.
(Die Vertragspartner halten die Massnahmen für notwendig.) <sub obj adj-für>
(44‑E) We considered this insufficient and ... <sub obj adj>
(44‑G) Wir fanden dies nicht ausreichend und ... <sub obj adj>

finden is the only TE that can occur with the same valency structure as consider (44), i.e. finden with the meaning of ‘having an opinion’ can be directly followed by a nominal or adjectival complement in German (Helbig 1992). For the other verbs a particle, either ‘als’ or ‘für’, is needed to express the meaning in German. Since the corresponding verbs, for example ansehen and betrachten, can also occur as translations without the particle, sub-categorisation is needed. However, the particles do not function as prepositions (Heringer 1970) in German. Therefore the German valency sentence patterns are best shown as <sub obj adj-als> or <sub obj adj-für> respectively (42, 43).

3.6 The patterns <sub obj nom-as / adj-as>

Fischer (1997) notes that in English the nominal and adjectival complements occur in two different forms: with ‘as’ and without ‘as’. He argues that the classification for both should therefore be <nom-as> and <adj-as>. The frequent TEs are betrachten, ansehen, gelten and bezeichnen. It is notable that these also occur in German with the particle ‘als’, which is equivalent to the English ‘as’, as shown in example 45:

(45‑E) Parties who consider Professor Vermeersch as a moral beacon, ... <sub obj nom-as>
(45‑G) Parteien, die Professor Vermeersch noch immer als einen moralischen Leuchtturm betrachten, ... <sub obj nom-als>

The local grammar, i.e. the syntactic valency pattern of consider, thus remains consistent in German for the most frequent correspondences.

3.7 The patterns <sub obj vb-to-inf> and <sub obj vb-to-be-nom/adj>

Verbal complements with ‘to-infinitive’ represent non-finite clauses. Fischer (1997: 151) comments: “English features more verbal complements than German, because English can expand nominal and adjectival complements into verb phrases”. This change does not affect the meaning. Therefore it is probably not surprising that the frequent TEs for these valency patterns are the same as for the patterns <sub obj nom> and <sub obj adj> (section 3.4). Extension with an infinitive clause is not possible for German verbs. The German sentence structure will therefore show no difference in the translation between the extended and the non-extended valency sentence patterns of consider, as shown in examples 46 and 47.

(46‑E) The Parliament has always considered the promotion of these rights to be one of its most important tasks. <sub obj vb-to-inf>
(46‑G) Das Parlament hat die Wahrung dieser Rechte stets als eine seiner wichtigsten Aufgaben angesehen. <sub obj nom-als>
(47‑E) The decision can be considered an historically significant fact.<sub obj nom>
(47‑G) Die Entscheidung kann als ein Ergebnis von historischer Bedeutung angesehen werden. <sub obj nom-als>

3.8 Patterns with correlate ‘it’

The most frequent TEs that occur with a correlate ‘it’ in the valency sentence pattern of consider are halten, erachten and finden which also occur with the equivalent correlate ‘es’ in German. As can be seen in example sentences 48 and 49, there is no difference in the valency sentence pattern of consider and the valency sentence pattern of the TE.

(48‑E) I consider it crucial that we have a true internal market. <sub it adj vb-that>
(48‑G) Ich halte es für äusserst wichtig, dass wir über einen echten Binnenmarkt verfügen. <sub es adj-für vb-dass>
(49‑E) The vast majority of Member States did not consider it necessary to amend the article. <sub it adj vb-to-inf>
(49‑G) Die Mehrzahl der Mitgliedstaaten erachtete es nicht als notwendig, den Artikel zu ändern. <sub es adj-als vb-zu-inf>

4. Conclusion

Based on the analysis, it appears that the preferred TEs for a valency sentence pattern of consider are attributable to a syntactic affinity between the patterns of consider and the patterns of the respective TEs. However, this is only partly true. Although all of the patterns investigated show some preference for translation equivalents and their possible valency patterns based on frequency analysis, it is also notable that there is also a degree of freedom in the translations. Suitability for replacement concerns the question of synonymy. It is generally accepted that accredited semantic similarity of certain words does not necessarily include an unrestricted exchangeability of these (Lyons 1981: 50–51; Bußmann 1983: 525–526). This is also true for translations.

Frequency analysis seems to be a suitable criterion for the analysis of patterns and their TEs because it provides information about the usage of words. In addition, it can also be expected that translators and language learners come across the frequent structures regularly. An alternative expression may be a grammatically correct substitute, but it may be less frequently used and therefore be a marked form.

Overall, the preliminary findings presented seem to indicate that words in a bilingual context are more likely to be interpreted or translated within meaning categories than one-to-one equivalents. A chosen synonym or translation equivalent will have its own valency pattern and impose this on the clause structure. Nevertheless, the investigation has also shown that syntactic patterns can be a useful indicator of likely meaning and are therefore valuable for learners of English or in translation training.

The contrastive approach demonstrated, working with parallel corpora, frequency analysis, and presentation of valency sentence patterns, highlights similarities and differences between two languages with regard to the choice of TEs and is a useful aid to ranking key translations based on actual usage. These findings can be utilised in dictionary compilation, language teaching and translation studies.


EuroParl corpus: www.statmt.org/europarl/.

Bank of English corpus: http://www.titania.bham.ac.uk.

Deutsches Referenz Korpus: http://www.ids-mannheim.de/kl/projekte/dereko/.

Institut für Deutsche Sprache: http://www1.ids-mannheim.de/index.php?id=1 and http://hypermedia.ids-mannheim.de/evalbu/index.html.


[1] Examples based on occurrences or non-occurrences in the BoE.

[2] The contrastive examples show the letter ‘E’ for the English sentence and the letter ‘G’ for its corresponding counterpart in EuroParl.

[3] The definition of ‘object’ is based on Quirk et al (1985) who argue that the object has a variety of realisation forms, such as that-, wh- or non-finite clauses.

[4] sein is not the only possible verb, but the most frequent. For example, the verb vertreten as in ‘die Ansicht vertreten’ is also possible.


Allerton, D.J. 1982. Valency and the English Verb. London: Academic Press.

Bianco, M. T. 1988. “Zur Kontrastierung deutscher und italienischer Verben auf Basis der Valenztheorie”. Valenzen im Kontrast, ed. by P. Mrazović & W. Teubert. Heidelberg: Julius Gross Verlag. 40–61.

Biber, D., S. Conrad & G. Leech. 2002. Longman Student Grammar of Spoken and Written English. Harlow: Pearson.

Bußmann, H. 1983. Lexikon der Sprachwissenschaft. Stuttgart: Alfred Kröner.

Cornell, A., K. Fischer & I.F. Roe. 2003. Valency in Practice – Valenz in der Praxis. Bern: Peter Lang.

Duffner, R., A. Kamber & A. Näf. 2009. “Europäisch eingestellt – Valenzforschung und Parallelkorpora”. Linguistik online 39: 45–60. www.linguistik-online.de.

Engel, U. & H. Schumacher. 1976. Kleines Valenzlexikon deutscher Verben. Tübingen: Gunter Narr Verlag.

Emons, R. 1974. Valenzen englischer Prädikatsverben. Tübingen: Max Niemeyer.

Fillmore, Ch. J. 2007. “Valency Issues in FrameNet”. Valency – Theoretical, Descriptive and Cognitive Issues, ed. by Th. Herbst & K. Götz-Votteler. Berlin: de Gruyter. 129–162.

Fischer, K. 1997. German-English Verb Valency – A Contrastive Analysis. Tübingen: Gunter Narr.

Firth, J.R. 1957. Papers in Linguistics 1934–1951. London: Oxford University Press.

Francis, G., S. Hunston & E. Manning, eds. 1996. Collins Cobuild Grammar Patterns 1: Verbs. London: HarperCollins Publishers.

Goldberg, A. 2006. Constructions at Work – The Nature of Generalizations in Language. Oxford: Oxford University Press.

Götz-Votteler, K. 2007. “Describing semantic valency”. Valency – Theoretical, Descriptive and Cognitive Issues, ed. by T. Herbst & K. Götz-Votteler. Berlin: Walter de Gruyter. 37–50.

Groom, N. 2007. Phraseology and Epistemology in Humanities Writing: A Corpus-driven Study. PhD thesis, University of Birmingham.

Helbig, G. 1992. Probleme der Valenz- und Kasustheorie. Tübingen: Max Niemeyer Verlag.

Herbst, T., D. Heath, I.F. Roe & D. Götz. 2004. Valency Dictionary of English. Berlin: Walter de Gruyter.

Heringer, H. J. 1970. Theorie der deutschen Syntax. München: Max Hueber Verlag.

Hoey, M. 2005. Lexical Priming – A New Theory of Words and Language. Oxon: Routledge.

Homberger, D. 2001. Grammatik für den Deutschunterricht. Leipzig: Klett Schulbuchverlag.

Leech, G. 1991. “The state of the art in corpus linguistics”. English Corpus Linguistics: Studies in Honour of Jan Svartvik, ed. by K. Aijmer & B. Altenberg. London: Longman. 8–29.

Lyons, J. 1981. Language, Meaning & Context. London: Fontana Linguistics.

Noël, D. 1996. “Grammar in bilingual dictionaries: contrasting English/French dictionaries and the CVVD”. Aspects of Contrastive Verb Valency, ed. by A.-M. Simon-Vandenbergen, J. Taeldeman & D. Willems. Gent: Studia Germanica Gandensia. 319–338.

Oxford Advanced Learner’s Dictionary. 2005, 7th edition. Oxford: Oxford University Press.

Polenz, v., P. 2008. Deutsche Satzsemantic, 3rd edition. Berlin: Walter de Gruyter.

Quirk, R., S. Greenbaum, G. Leech & J. Svartvik. 1985. A Comprehensive Grammar of the English Language. London: Longman.

Römer, U. & R. Schulze. 2009. Exploring the Lexis-Grammar Interface. Amsterdam: John Benjamins.

Schumacher, H., J. Kubczak, R. Schmidt & V. de Ruiter. 2004. VALBU – Valenzwörterbuch deutscher Verben. Tübingen: Gunter Narr Verlag.

Sinclair, J. 1991. Corpus, Concordance, Collocation. Oxford: Oxford University Press.

Sinclair, J. 2004. Trust the Text – Language, Corpus and Discourse. London: Routledge.

Swan, M. 2005. Practical English Usage. Oxford: Oxford University Press.

Tesnière, L. 1959. Grundzűge der strukturalen Syntax (German translation by Engel, U., 1980). Stuttgart: Klett-Cotta.

Teubert, W. 2003. “Die Valenz nichtverbaler Wortarten: das Substantiv”. Dependenz und Valenz / Dependency and Valency Volume 1, ed. by V. Ágel, H.-W. Eichinger, P.H. Eroms, H.J. Heringer & H. Lobin. Berlin: Walter de Gruyter. 820–835.

Teubert, W. 2004. “Corpus Linguistics and Lexicography: The Beginning of a Beautiful Friendship”. Lexicographica 20: 1–19.

Welke, K. 1988. Einführung in die Valenz- und Kasustheorie. Leipzig: Bibliographisches Institut.