Semantic Roles in English and German: a Contrastive Study

Share Embed


Semantic Roles in English and German: a Contrastive Study Module QXL 4450 – “Issues in Cognitive Linguistics”

Table of Contents INTRODUCTION AND OVERVIEW ................................................................................................ 1 SEMANTIC PARTICIPANT ROLES .................................................................................................. 2 ACTION CHAINS AND PARTICIPANT ROLES IN COGNITIVE GRAMMAR ........................................ 4 THE GERMAN CASE SYSTEM ........................................................................................................ 7 ANALYSING ENGLISH AND GERMAN ........................................................................................... 7 THE STANDARD CASE .......................................................................................................................... 8 DIATHESES ....................................................................................................................................... 9 SPECIAL CASES ................................................................................................................................ 12 SUMMARY AND FINAL THOUGHTS ............................................................................................ 13 GLOSSING ABBREVIATIONS ....................................................................................................... 14 REFERENCES .............................................................................................................................. 14

Introduction and Overview English and German are both members of the West-Germanic branch of the IndoEuropean family of languages. As such, they are closely related and one would expect them to have much in common. And in fact, on the whole this expectation is confirmed when one takes a look at many linguistic features, be they lexical, grammatical, or even phonological. However, over the many centuries of their geographical separation, which began when the Anglo-Saxon invaders left their previous homes in what is today northern Germany and Denmark to find a new home in the British Isles, the two languages have also developed several distinctive differences. Many of these differences can be found in the lexicon, a fact that is due in large part to the many foreign influences that affected the English language but left German for the most part unaffected. The differences are far from purely lexical, however. Grammatically,


the two languages also differ greatly. While English, in the course of its history, has gradually lost almost all of its inflections, for example, German still boasts quite a complex inflectional system, which covers both verbs, e.g. in the form of tense, and nouns, largely due to its case system. The particular difference between English and German that shall be examined in more detail in this essay mainly concerns the latter of the two mentioned word classes: the nouns and their case inflection. Specifically, what shall be analysed here are semantic participant roles in a sentence, as well as the ways there are of mapping these onto syntactic structure. This issue has been thoroughly researched for English and to a lesser extent also for German. What this essay would like to achieve, however, is a contrastive analysis of the relationship between semantic and syntactic participant roles in English and German, from a Cognitive Linguistic, more precisely a Cognitive Grammatical, perspective. The essay is divided into several sections. Section 1 will introduce the notion of participant roles in general, giving a general overview of the topic. Section 2 will then go on to describe the same issue from the point of view of Ronald Langacker’s (1987; 1991) Cognitive Grammar, in the context of the so-called action chain model. In section 3, then, a very brief overview of the German case system will be given, which is necessary in order to properly understand section 4, which will apply the Cognitive Grammatical model of participant roles to the German language, while contrasting it throughout with English. Finally, section 5 will summarise the findings of the previous sections and conclude with some final thoughts on the issue.

Semantic Participant Roles Verbs, when used in a sentence, almost always require at least one noun phrase in order to make sense. The word ‘swims’, for example, cannot usually be a sentence by itself, since in English it is grammatically required that one states additionally who is doing the swimming. So one might say ‘John swims’, which would be a legitimate sentence. These noun phrases which are required by the verb are known as complements. This term contrasts with the term ‘modifier’, which describes an optional noun phrase not required by the verb, as in the sentence ‘John swims in the lake’.


Several verbs require more than one complement. These are transitive and ditransitive verbs. The verb ‘to hit’ for example requires two complements: one that describes the entity (usually a person) doing the hitting and one that describes the entity being hit. Since it obviously makes a big difference whether one hits someone else or gets hit oneself, one can tell that there are several different types of complement. In this case we have and AGENT (the entity performing the action) and a PATIENT (the entity being affected by the action). These categories describe nuances of meaning attributed to the phrases in question and are thus semantic categories. It is these semantic categories that are meant when one talks of semantic participant roles or, abbreviated, simply semantic roles. It is important to note that semantic participant roles are not just types of verbal complement. Verbal modifiers also always have a semantic role within the sentence. The major semantic roles such as the aforementioned AGENT and PATIENT are generally performed by complements, however, while modifiers are usually used to take on minor, often optional roles. Defining a definitive list of semantic roles is an endeavour which many linguists have attempted and which is highly problematic and controversial. As Dowty (1991: 547) says: “There is perhaps no concept in modern syntactic and semantic theory which is so often involved in so wide a range of contexts, but on which there is so little agreement as to its nature and definition, as


[…]”. The reason for this is that meaning is a continuum. The semantic nuances of complements and modifiers vary from verb to verb, so it is hard to it is very hard to generalise. This has led some (Cognitive) linguists to define semantic roles in terms of prototypicality effects (Lakoff 1977), which may well be the solution but would mean that it is impossible to define the semantic roles. For the purposes of this text, however, it is unnecessary to get involved in this particular debate. It is sufficient here to use the broad terminology most commonly used. The following four semantic roles (taken from Evans and Green (2006: 675)) should suffice: 


volitional initiator of action


undergoes effect of action; change of state


for whose ‘benefit’ action is performed


means by which action is performed


Other common semantic roles include EXPERIENCER, THEME (sometimes also equated with PATIENT (Engels & Vikner 2006: 20)) and LOCATION.

Action Chains and Participant Roles in Cognitive Grammar Several of the major semantic roles can be accounted for in Langacker’s Cognitive Grammar by the so-called action chain model. Langacker’s (1991: 283) theory, which is based upon earlier work done by Fillmore (1968), is that a prototypical action event consists of an energy transfer initiated by an AGENT (the energy source), via an INSTRUMENT, to a PATIENT (the energy sink), which absorbs the energy and in doing so undergoes a change of state. The resulting action chain can thus be represented in schematic form, as in (1). This pattern can be illustrated by the following sentence (2) (Taylor 2002: 421): (1)



John opened the door with the key.

This sentence contains all three parts of the action chain. ‘John’ is the AGENT who sets the action in motion. He does this by means of ‘the key’, which represents the INSTRUMENT. Finally, ‘the door’ absorbs the energy, thus opening, which constitutes a

change of state. It therefore represents the PATIENT. These three semantic roles are conceptual categories that have become entrenched in the mind through repeated use. In order for these semantic elements to be codified through language they need to take on phonetic form in the form of syntax. Due to their conceptual nature and the fact that the semantic properties of a verb’s complements vary from verb to verb, the semantic roles can be (and most likely are) infinite in number. There can however only be a limited number of syntactic roles, since phonological entities need to be learned by speakers. So these infinite nuances of semantic roles have to somehow be mapped onto the set number of syntactic categories. Sentence (2) represents the most basic way of doing so in English for a prototypical action chain: The AGENT is put in subject role, the PATIENT in direct object role, and the INSTRUMENT is realised through a prepositional modifier phrase. Since the aim of this essay is to view the subject matter from the perspective of Cognitive Grammar, the terms ‘subject’ and ‘object’ will not be used for the 4

remainder of the text. Instead, we shall speak of ‘trajector’ (TR) and ‘landmark’ (LM), terms coined by Langacker. When an element in a scene gains linguistic realisation, this is called profiling. A profiled element is termed a ‘profile’, and Langacker distinguishes between ‘nominal profiles’, which designate entities (mostly nouns and pronouns in traditional terminology) and ‘relational profiles’, which designate relationships between elements, i.e. nominal profiles and other relational profiles (verbs, adjectives and prepositions, among others). The primary element linked by a relational profile is called the trajector, while the one or more other elements linked are called landmarks. In the case of (2), ‘John’ is the trajector, ‘the door’ is the primary landmark, and ‘the key’ is the secondary landmark. There are, however, also various alternative ways of linguistically encoding the action chain scene described by (2). In sentence (2) all three members of the action chain are profiled, that is, explicitly linguistically mentioned. In the following examples, only some of the participants are mentioned: (3)

John opened the door.


The key opened the door.


The door opened.

In sentence (3),




are profiled, while the


is not. The

mapping consequently stays largely the same as in (2), since the complements stay the same and it is only a modifier that needs to be eliminated. In (4) and (5), however, things start to change. In (4) it is the AGENT that is not profiled. This leads to the INSTRUMENT becoming the trajector of the verb in the sentence instead. In (5) neither the


nor the


is profiled, so the


moves into the

vacated position of trajector. The important fact here is that although the INSTRUMENT or PATIENT moves into trajector position, it nevertheless remains the





The AGENT in sentence (4) for example is still John, even though he is not profiled. Semantic roles, in other words, are completely independent from syntactic roles, though some semantic roles are more likely than others to be encoded by certain syntactic roles. This is a consequence of the fact that the linguistic realisation of a concept does not change the concept itself. Whether John is profiled or not, any


hearer of sentence (4) will still know that there is an


involved in the event,

since keys, being inanimate, are by nature unlikely to be AGENTS themselves. Attentive readers may notice that this could be understood to contradict a statement made earlier, that a transitive verb such as ‘hit’ or ‘open’ necessarily requires an AGENT and a PATIENT. Some linguists actually do deny that INSTRUMENT and PATIENT retain their identities when put in the position of trajector, as in (4) and (5)

(Welke 2002: 96). In actual fact verbs usually do not strictly require certain semantic roles as complements. As we have now seen, often a similar semantic role can be mapped onto a syntactic role. For example an agent might be replaced by an EXPERIENCER



When this happens, however, the alternative semantic

role often gains certain properties usually only displayed by the semantic role usually occupying the spot. For example, the


in (4) has gained some

vaguely animate qualities. Taylor (2002: 421) illustrates this fact with example sentences such as ‘The key easily opened the door’ or ‘The key just won’t open the door’. There is one common factor that influences the way semantic roles are mapped onto syntactic structure. That factor is grammatical voice, such as the passive. In this context, rather than ‘voice’ the term ‘diathesis’ (from Ancient Greek διάθεσις, meaning ‘arrangement, composition’) is perhaps more appropriate. A diathesis, in broad terms, is a means of rearranging the complements in a sentence in order to change the construal of the scene. It is a kind of template or pattern of profiling and syntactic mapping. The passive diathesis has the effect of eliminating the previous trajector from the linguistic realisation of a scene (i.e. it is no longer profiled) and moving the element hitherto in the role of primary landmark into trajector position. The verb ‘to open’ does not need the passive in order to do this, as previously shown in sentence (5). But other verbs, such as ‘to stroke’, do: (6)

John stroked the dog.


*The dog stroked.


The dog was stroked (by John).

Sentence (7) is ungrammatical. Instead, the passive diathesis (8) is employed in order to take emphasis off the AGENT and instead emphasise the PATIENT. If one still wishes to profile the agent, it can be reintroduced by means of a ‘by’ construction.


The German Case System Before we finally go on to analyse the methods German has of mapping semantic to syntactic structure, a short description of the German case system is useful. German has four grammatical cases, three of which are of major importance for this text. The four cases are the nominative, accusative, dative and genitive case. While case is sometimes marked directly on the noun itself, it is more clearly marked on the definite article. Let us briefly take a look at each in turn. 

The primary function of the nominative case is to mark the trajector of a verb. The nominative forms of the definite article for masculine, feminine and neuter gender are der, die and das.

The accusative designates what is traditionally called the direct object, which usually (but by no means always) corresponds to the primary landmark of the finite verb. The accusative definite articles are den, die and das.

The dative case most often marks a secondary landmark. It can in some cases however also mark a primary landmark. The definite articles are dem, der and dem.

Finally, the genitive case is almost exclusively used to mark a possessor. As such it is not of great relevance to this text. In some rare cases however it can also designate a landmark. The definite articles are des, der and des.

The following sentence (9) contains all four cases, each in its most typical role: (9)


Sohn des


the.M.NOM son the.M.GEN teacher-GEN Mädchen den girl






the.M.ACC bouquet.

“The teacher’s son gave the girl the bouquet.”

Analysing English and German We can now go on to the main part of this text and take a close look at how the German language syntactically encodes the most important semantic roles.


The standard case We will first examine the most usual way of doing this, using as an example the translation of the standard transitive sentence (2) discussed above: (10) John öffn-et-e John open-3SG-PST


Tür mit dem



door with the.M.DAT key

“John opened the door with the key.” The major difference between the sentence in German and in English is the use of case. Apart from this, the coding is identical. The AGENT is the trajector of the verb, which is evident in English through its position at the beginning of the sentence, in German through its nominative case marking (it needn’t necessarily be in sentenceinitial position). The


is the primary landmark, which is marked by its

position as final verbal complement in English and by the accusative case in German. The instrument, finally, is represented in both languages by a prepositional modifier. The same observations can be made when examining the ditransitive sentence (9) above. The AGENT and PATIENT are joined by a new semantic role, the BENEFICIARY. This role is a secondary landmark in both languages, which is coded in English by being inserted between the verb and the primary landmark and in German by the dative case. There is one difference, however. In English, the scenario could equally well be described by sentence (11): (11) The teacher’s son gave the bouquet to the girl. (12) ?The teacher’s son gave the bouquet. In this case, the


retains its identity as a complement, since it is still

required by the verb (The grammaticality of (12) is questionable at best.), but now it is encoded by means of a prepositional phrase. This kind of structure is not possible in German. Let us next take a look at the German equivalents (13), (14) and (15) of sentences (3), (4) and (5) respectively above, which are alternative construals of the scene described by (2), in which various participants are not profiled.


(13) John öffn-et-e John open-3SG-PST





“John opened the door.” (14) Der

Schlüssel öffn-et-e

the.M.NOM key






“The key opened the door.” (15) Die

Tür öffn-et-e

the.F.NOM door open-3SG-PST

sich. REFL.ACC

“The door opened.” Sentences (13) and (14) show no great differences in structure compared to their English counterparts. Sentence (15), however, is interesting. Here we notice that the German verb ‘öffnen’ cannot be used intransitively, as the English verb ‘to open’ can. So, when the


is put in trajector position, the verb nonetheless requires a

landmark, which is represented by a reflexive pronoun with accusative case marking. A literal English translation would be “The door opened itself”. This pronoun functions as a kind of dummy landmark in the absence of an alternative. We see, therefore, that it is not as easy in German to put a


in trajector

position as it is in English. Diatheses As discussed above, Diatheses are common and regular patterns for changing the standard syntactic coding of semantic roles. By far the most widespread diatheses in both English and German are the active and the passive. Arguably, however, there is another diathesis in both languages: the causative. And German even has another: the applicative. First, however, let us examine how the passive works in German, by using as an example the translation (16) of sentence (8): (16) Der


the.M.NOM dog


von John ge-streichel-t. by



“The dog was stroked by John.” We see that the passive in German works in exactly the same way as in English. Reintroduction of the agent by means of a prepositional modifier is equally optional 9

in German. There is however one interesting, if very rare, phenomenon concerning the passive which does not exist in English (adapted from Sommerfeldt et al. 1991: 46): (17) Das

Gebäude ist

the.N.NOM building






“The building is barred by the police.” In this construction the PATIENT, being the trajector, is marked by the nominative, as expected. The AGENT however is not coded by a prepositional modifier but rather by an adverb derived from the original noun. From a Cognitive Grammar point of view, this makes perfect sense, since adverbs are viewed as relational profiles that inherently contain their landmark. This is exactly the same as a prepositional phrase, which as a whole is also a relational profile, since the preposition is a relational profile and has profile determinacy, that is it “determines the profile of the composite structure as a whole” (Evans & Green 2006: 585). The next diathesis to be found in German also exists in English, though to a far lesser extent. That is the causative. While the passive causes one entity (the trajector) in an active sentence no longer to be profiled, thus causing the landmark to ‘move up’ to trajector position, the causative does the opposite: It profiles a new entity which takes the place of trajector. This causes the previous trajector to ‘move down’ into primary landmark position and the previous landmark (if present) to become a secondary landmark. The semantic role of the new entity can be called CAUSER. An English example of a causative would be sentence (19), derived from the

active sentence (18): (18) The boy wiped the blackboard. (19) The teacher made the boy wipe the blackboard. In (19), the teacher is profiled as


of the action. The same is possible in

German, where the causative auxiliary is usually ‘lassen’ (to let):


(20) Der the.M.NOM






the.M.ACC boy-ACC








“The teacher made the boy wipe the blackboard.” Sentence (20) shows that in German causative constructions the nominative case marking, while the




has accusative case, the same as the


As a final case study, let us examine a diathesis in German that is completely alien to English: the applicative. The applicative does not introduce or eliminate participants in a sentence. What it does instead is to switch around the two landmarks of a verb, so that primary becomes secondary and vice versa. The following two sentences exemplify the phenomenon: (21) Der


the.M.NOM man die







throw.3SG the.ACC.PL stone-PL

gegen against

“The man throws stones against the wall.” (22) Der


the.M.NOM man den










the.DAT.PL stone-PL-DAT *“The man throws X at the wall with the stones.” We see that the applicative diathesis, formed by the prefix ‘be-’, causes the PATIENT and primary landmark (the stones) to be coded by a prepositional modifier, while the secondary landmark, in this case a LOCATION (the wall), becomes a complement with accusative case marking. What is more, the modifier containing the PATIENT in the applicative construction is optional, so that it is possible to profile only and LOCATION, while leaving the PATIENT implicit.



Special cases After talking about the most common and typical constructions and their variations discussed in the introductory sections, as well as the effects of the various diatheses, we shall now go on to examining some more unusual cases. We will do this by examining the different semantic roles one by one. When a German sentence contains an AGENT, that AGENT is almost always the trajector, unless a change in diathesis has occurred. The one exception to this concerns certain unusual verbs such as ‘gelingen’ (to succeed), which has the AGENT as its landmark in dative case, while the PATIENT is the trajector in nominative case: (23) Der


the.M.NOM cake





the.M.DAT baker

“The baker successfully made the cake.” The PATIENT in a sentence is usually the primary landmark, unless this is changed by a diathesis. As landmark, it is usually encoded by the accusative case. However, examples can be found where the PATIENT, in spite of being primary landmark, has dative or even genitive case marking, or is encoded by a prepositional phrase. The reason for this is usually the peculiarity of certain verbs. Finally, the


when it is profiled at all, is generally encoded by an

optional prepositional phrase. The main exceptions to this are certain verbs such as ‘dienen’ (to serve), which require an


as a complement, due to their

semantic properties. The complement can either be the trajector, as in (24), or the landmark, as in (25). (24) The sword served him well as a weapon. (25) He used the sword to kill his enemies. As we see, this phenomenon exists in English too. The only difference is that the instrument can be marked by different cases in German, when it is in landmark position.


Summary and Final Thoughts We have now extensively examined several different methods the German language has of mapping semantic roles, an elusive and often indefinable mental category, onto the highly finite syntactic structure of language. In many regards the two languages do this in very similar ways. Both treat the AGENT as the most important semantic role by very consistently assigning it the role of verbal trajector in a sentence. Both languages also tend to put the PATIENT in primary landmark position, though this is not done as consistently. By not profiling certain semantic roles, a scene can be construed in different ways, and both languages do this very similarly, too. In the area of diathesis, the passive and causative exist in both languages and work in very similar ways. The two languages do also differ in various ways, however. Most of these differences are due in large part to German’s linguistic feature of grammatical case, which is almost completely absent in English. The three cases relevant here (nominative, accusative and dative) appear to have certain semantic roles typically assigned to them mentally: the nominative case usually marks the



accusative marks the PATIENT and the dative marks the BENEFICIARY. This has certain advantages. For example, it has the well-known effect that the word order within a sentence is far looser, so that the participants can be moved around relatively freely, thus adding an additional way of emphasising the individual participants. However, it also means that it is cognitively not as easy to assign alternative roles to grammatical cases, because the standard correlation described above is so strongly entrenched in the minds of German speakers. This might explain certain phenomena we have observed in the course of this essay, such as the fact that it is not as easy in German to put the PATIENT in trajector position, which would mark it with the nominative. It may also be the reason why German has more formal ways of changing the semantic-syntactic mapping in a sentence, such as diatheses, as well as more verbs with unusual complements, such as the word ‘gelingen’ in (23): Since speakers find it hard to think of a certain case as corresponding to any other than one particular role, more formulaic methods of assigning alternative roles are required instead.


Glossing Abbreviations 3

3rd person









































References Dowty, David. 1991. Thematic Proto-Roles and Argument Selection. Language 67(3). 547– 619. Engels, Eva & Sten Vikner. 2006. Satzglieder, Kasus und semantische Rollen: Eine Einführung. Tidsskrift for Sprogforskning 4(1-2). 17–37. Evans, Vyvyan & Melanie Green. 2006. Cognitive Linguistics: An introduction. Edinburgh: Edinburgh University Press. Fillmore, Charles J. 1968. The Case for Case. In E. Bach & R. T. Harms (eds.), Universals in Linguistic Theory, 1–88. New York: Holt, Rinehart & Winston. Lakoff, George. 1977. Linguistic Gestalts. In, Papers from the 13th regional meeting Chicago Linguistic Society, 236–287. Chicago. Langacker, Ronald W. 1987. Foundations of cognitive grammar, vol. I: Theoretical Prerequisites. Stanford CA: Stanford University Press. Langacker, Ronald W. 1991. Foundations of cognitive grammar, vol. II: Descriptive Application. Stanford CA: Stanford University Press. Sommerfeldt, Karl-Ernst, Herbert Schreiber & Günter Starke. 1991. Grammatischsemantische Felder: Einführungen und Übungen. Berlin, München, Leipzig: Langenscheidt. Taylor, John R. 2002. Cognitive grammar (Oxford textbooks in linguistics). Oxford, New York: Oxford University Press. Welke, Klaus. 2002. Deutsche Syntax funktional: Perspektiviertheit syntaktischer Strukturen (Stauffenburg Linguistik). Tübingen: Stauffenburg.


Lihat lebih banyak...


Copyright © 2017 DATOSPDF Inc.